Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🍿 Netflix: The Battle for Attention

• 🔎 Google: The $99 Billion Asterisk

Subscribe now

Premium members get:

• 📊 Monthly reports: 200+ companies visualized.

• 📩 Tuesday articles: Exclusive deep dives and insights.

• 📚 Access to our archive: Hundreds of business breakdowns.

PRO members get everything PLUS:

• 📩 Saturday PRO reports: Timely insights on the latest earnings.

Today at a glance:

1. 🏭 Intel: Fifteen-Year High

2. ⚡ GE Vernova: Capacity Bet

3. ⚙️ Texas Instruments: Pricing Lever

4. 💳 Amex: Platinum Premium

5. 📶 T-Mobile US: Momentum Cools

6. 🌐 IBM: Mainframe Air Pocket

7. 📡 Verizon: Subsidy Pullback

8. 🏦 Schwab: The Volatility Dividend

9. 📞 AT&T: Volume Over Price

10. 🛰️ Lockheed Martin: The Missile Surge

11. 🧑‍💻 ServiceNow: Growth Without Credit

12. 💼 Moody’s: Volume Over Mix

13. 🚗 GM: Pricing Over Volume

14. 🍕 Domino’s: Ticket Miss

15. 🦅 American Airlines: Fuel Eats The Record

16. 🍿 AMC: The Odyssey Delivers

1. 🏭 Intel: Fifteen-Year High

Intel’s Q2 revenue rose 25% Y/Y to $16.1 billion ($1.7 billion beat), and non-GAAP EPS was $0.42 ($0.20 beat), against a $0.10 loss a year ago. CEO Lip-Bu Tan called it the strongest revenue growth in more than fifteen years.

Intel lost $10.8 billion on paper because its own stock price went up so fast that it made the free shares it promised the US government way more expensive to give away. INTC has nearly tripled this year but sits close to 30% below its June 22 high. The stock was caught in a sector-wide rotation out of chip stocks as Wall Street questions whether AI hardware spending is sustainable.

• Data Center and AI revenue climbed 59% Y/Y to $6.3 billion, more than double Intel’s overall growth rate, as agentic workloads pull the stack back toward CPUs.

• Client Computing rose 13% Y/Y to $8.9 billion against an $8.0 billion consensus, with AI PCs now two-thirds of the client mix.

• Foundry grew 31% Y/Y to $5.8 billion, accelerating from 16% last quarter, and adjusted gross margin hit 41.8%, roughly 280 basis points above guidance.

CFO Dave Zinsner said rising memory prices will hit the PC business in the coming quarters, and management already expects sub-seasonal PC consumption in the second half. Intel is spending into that anyway, raising 2026 CapEx from $18 billion to more than $20 billion. Management expects 2027 CapEx to be significantly higher. Foundry's growth is still mostly internal. The segment sells almost entirely to Intel's own product groups, and CEO Lip-Bu Tan wouldn't name external customers, pointing to early next year for visible progress.

Q3 guidance of $15.8–16.8 billion implies a midpoint roughly $1.2 billion above consensus, with EPS of $0.38 (vs. ~$0.27 expected) and gross margin of 42% (vs. ~40% consensus). Zinsner said Intel can sell every data center chip it makes, 18A yields are running ahead of plan, and 14A will move to volume production in 2028. Client Computing is the segment to watch in Q3, when higher memory costs hit PC pricing.

2. ⚡ GE Vernova: Capacity Bet

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🏦 Banking on Volatility

• 📊 PRO: TSMC, ASML, J&J…

• 🍿 Netflix: The Battle for Attention

• 🔎 Google: The $99 Billion Asterisk

Subscribe now

🗓️ Earnings season is picking up speed

Tomorrow, we’ll break down 16 key reports for PRO members, including IBM's worst single-session drop ever and ServiceNow fighting the SaaSpocalypse narrative.

Today, we look at Tesla’s earnings and Uber’s biggest acquisition yet.

• 🚖 Tesla: Cash Burn Begins

• ■ Uber: The $15 Billion Order

🚖 Tesla: Cash Burn Begins

Tesla got its demand mojo back. Deliveries hit a Q2 record of 480K vehicles, and trailing-twelve-month revenue crossed $100 billion for the first time.

Yet shares plunged nearly 14%, their worst post-earnings decline since 2013.

Operating profit fell 57% to just $398 million, while free cash flow turned negative. Last quarter, we flagged that Q1’s margin beat leaned on one-time items and that Tesla’s $25 billion CapEx cycle would push free cash flow negative. Both arrived in Q2.

Income statement:

• Revenue grew 26% Y/Y to $28.2 billion ($1.7 billion beat).

• Gross margin was 17% (-0pp Y/Y and -4pp Q/Q).

• Operating margin fell to 1% (-3pp Y/Y).

• Non-GAAP EPS was $0.33 ($0.21 miss).

Cash flow:

• Operating cash flow grew 85% Y/Y to $4.7 billion.

• Free cash flow swung from positive $0.1 billion last year to negative $1.1 billion as CapEx jumped 142% to $5.8 billion

FY26 outlook:

Tesla again withheld full-year guidance. CapEx will exceed $25 billion this year (outlook unchanged), and keep growing for another two to three years as Tesla expands Robotaxi, Optimus, AI compute, semiconductor, solar, and manufacturing capacity.

Management is also securing debt facilities that could provide up to $30 billion of borrowing capacity. Tesla still has $43.5 billion of cash and investments, but debt is becoming part of the funding plan.

So, what to make of all this?

• 🚘 Demand came back: Automotive revenue grew 23% Y/Y to $20.5 billion as deliveries increased 25% to 480K, roughly 74K above the company-compiled consensus. Tesla also said it exited Q2 with its largest order backlog since 2023. Management called FSD a major demand driver, with nearly 1.5 million paid customers and attach rates above 55% on new North American deliveries. The constraint has now shifted toward supply, particularly batteries and electronic components.

• ⚡️ Energy delivered volume, but not margin: Storage deployments reached 13.5 GWh, up 41% Y/Y and 53% sequentially, making Q2 Tesla’s second-largest quarter ever. But energy gross margin fell from 40% to 20% because of a $240 million warranty adjustment, the absence of Q1’s tariff benefit, and falling industrial-storage prices. Management expects margins to settle in the low-to-mid 20s.

• 🔌 Services quietly became the bright spot: Services revenue jumped 50% Y/Y to $4.6 billion, while gross margin climbed from 9% to a record 14%. Growth came from used vehicles, Supercharging, service centers, and insurance. These are no longer negligible side businesses.

• 📉 Q1’s margin support disappeared: Automotive gross margin excluding credits fell from 19% to 16%. Q1 included a $230 million warranty benefit and tariff relief that did not repeat, and management said underlying margins were roughly flat after adjusting for them. Meanwhile, operating expenses rose 47%, primarily due to AI, pre-production R&D, and stock-based compensation. It left operating profit at roughly one-quarter of consensus. Net profit of $1.1 billion looked healthier, but included a $1.0 billion unrealized gain on Tesla’s SpaceX stake.

• 💸 Cash burn begins, and debt enters the plan: CapEx more than doubled sequentially to $5.8 billion, pushing free cash flow to negative $1.1 billion. Tesla spent just $8.3 billion in the first half, so its greater-than-$25-billion outlook implies at least $16.7 billion in H2—more than double H1. Tesla has ample liquidity, but the funding model is changing. The company is preparing to add leverage.

• 🚖 Robotaxi keeps a limited scale: Robotaxi now operates in seven metros. Tesla says its unsupervised fleet has driven more than 380K miles across six cities, with weekly mileage growing at a double-digit rate. But Tesla remains early: Waymo has already accumulated roughly 220 million autonomous miles. Cybercab has entered production and is accumulating the calibration miles required to scale.

• 🤖 Optimus remains the hardest ramp: Tesla is replacing the Model S and X lines at Fremont with its first Optimus production lines. Musk called it the hardest manufacturing ramp Tesla has attempted, warning that the early portion of the production curve will be “flat and long.”

Bottom Line: Tesla delivered a strong quarter for demand, but investors rejected the economics. The nearly 14% sell-off shows the market is becoming less willing to let future AI opportunities offset weak current profits.

Tesla may be right to spend aggressively if Robotaxi, Optimus, and its AI infrastructure produce the returns Musk expects. But more than two-thirds of this year’s CapEx is still ahead, while the businesses meant to justify it have yet to contribute meaningfully to earnings. Tesla has the balance sheet to finance the wait, but the market remains skeptical.

■ Uber: The $15 Billion Order

Uber is making its biggest acquisition yet: $15 billion for a global delivery footprint it could not build market by market.

On July 16, Uber agreed to acquire Delivery Hero, the Berlin-based group behind foodpanda, Glovo, talabat, and Korea’s Baemin. The company will pay €41.50 per share in cash, up from the €33 it floated in May.

Uber already owned 25% outright and held another 12% of economic exposure through financial instruments. Prosus has irrevocably committed its remaining 17%, taking Uber’s economic interest to roughly 53% before other shareholders tender.

Why now? Delivery Hero gives Uber leading local brands across Asia, Latin America, the Middle East, and parts of Europe. Uber will acquire businesses in 50 markets that generated $42 billion in gross bookings last year once you carve out 14 overlapping markets.

Uber wants to connect Delivery Hero’s customers with its mobility network. Uber’s platform will expand from 79 to 99 markets, with combined 2025 gross bookings of $236 billion. The number of markets where it runs both rides and delivery will nearly double, from 34 to 58.

Uber says cross-platform users generate ~3x the gross bookings and profits of customers using only one service, while acquiring them through an existing platform costs more than 50% less than paid marketing. A food customer can become a rider, a rider can become a food customer, and Uber One becomes more useful to both.

The price looks steep at ~14x EBITDA before synergies for a sprawling, relatively low-margin delivery business. But Uber CFO Balaji Krishnamurthy is targeting $1.2 billion in run-rate synergies within 18 months, mostly from moving Delivery Hero onto Uber’s tech stack. If Uber captures the full $1.2 billion, management says its effective multiple falls to ~8x 2027 adjusted EBITDA. Uber calls the deal accretive to non-GAAP EPS from close.

Why it matters

• The scale game: DoorDash bought Deliveroo. Prosus took Just Eat Takeaway. Food delivery rewards density, and the last independents are running out of room to stay independent.

• The regulatory tax: Uber pre-sold 14 overlapping markets (including Türkiye, Spain, Poland, and others) to SSW Partners for ~$1.6 billion to ease antitrust review. A second-half-2027 target close tells you Brussels won’t wave this through.

• The AV hedge: There is also a defensive angle. A much larger delivery network gives Uber another source of demand and customer engagement if autonomous competitors eventually pressure the economics of ride-hailing.

Bottom Line: Uber paid up for a target whose largest shareholders wanted an exit, then structured the deal to improve its odds with regulators. If the promised $1.2 billion of synergies arrives, the effective 8x multiple could look compelling. But it does come with some execution risks The integration will not begin until the deal closes, likely in late 2027.

That’s it for today!

Stay healthy and invest on.

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Start an account for free and save 15% on paid plans with this link.

Author’s Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization’s views.

Disclosure: I am long TSLA, UBER, GOOG, and NVDA in the App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🏦 Banking on Volatility

• 📊 PRO: TSMC, ASML, J&J…

• 🍿 Netflix: The Battle for Attention

• 🧠 SK Hynix Bets AI Broke the Cycle

Subscribe now

Alphabet just reported $112 billion in net profit on $120 billion in revenue.

That’s not a typo. The explanation is a $99 billion paper gain on equity investments, largely reflecting revaluations of Alphabet’s stakes in Anthropic and SpaceX.

Strip that out, and the operating quarter was still exceptional. AI demand is outrunning Google’s capacity, but the cost of catching up is rising just as fast.

Here’s what stood out this quarter.

Today at a glance:

1. Alphabet Q2 FY26.

2. Funding the machine.

3. Key Insights from the call.

4. What to watch moving forward.

1. Alphabet Q2 FY26

Income statement:

Alphabet disclosed a $99 billion paper gain from its equity investments along with small items under other income.

Anthropic’s post-money valuation rose from $380 billion to $965 billion in three months. Alphabet reportedly owns roughly 14%, making Anthropic an obvious contributor to the gain. But Alphabet did not disclose the contribution from each investment, and SpaceX’s June IPO also created a public-market price for a stake Alphabet has held since 2015.

Revenue grew 24% Y/Y to $119.8 billion ($2.8 billion beat).

• 🔎 Advertising: $81.6 billion (+14%).

  • Search: $63.3 billion (+17%).

  • YouTube ads: $11.1 billion (+13%).

  • Network: $7.3 billion (-1%).

• 📱 Subscriptions, platforms, and devices: $12.9 billion (+15%).

• ☁️ Cloud: $24.8 billion (+82%).

Margin trends:

• Gross margin: 62% (+2pp Y/Y).

• Operating margin: 34% (+2pp Y/Y).

  • Services (Advertising & Other): 42% (+2pp Y/Y).

  • Cloud: 36% (+15pp Y/Y).

Cash flow:

• Operating cash flow was $39.1 billion (+39% Y/Y).

• Free cash flow flipped negative to a $5.9 billion outflow (compared to $24.5 billion generated last year).

With CapEx now expected to reach ~$200 billion in FY26 (up from $185 billion previously), most of Alphabet's operating cash flow will be plowed back into AI spending.

Balance sheet:

• Cash, cash equivalents, and marketable securities: $242.5 billion.

• Long-term debt: $98.2 billion.

So, what to make of all this?

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🏦 Banking on Volatility

• 🍿 Netflix: The Battle for Attention

Subscribe now

Premium members get:

• 📊 Monthly reports: 200+ companies visualized.

• 📩 Tuesday articles: Exclusive deep dives and insights.

• 📚 Access to our archive: Hundreds of business breakdowns.

PRO members get everything PLUS:

• 📩 Saturday PRO reports: Timely insights on the latest earnings.

Today at a glance:

1. ⚡️ TSMC: AI Megatrend Deepens

2. 🔬 ASML: Second Guidance Hike

3. 💊 J&J: Oncology & Icotyde Deliver

4. 💼 UnitedHealth: Turnaround Print

5. 👔 Morgan Stanley: $10 Trillion Milestone

6. 🏛️ Goldman Sachs: SpaceX Leads Deal Run

7. 📈 BlackRock: Assets Cross $15 Trillion

8. 🧬 Abbott: Diagnostics Carry The Quarter

9. 🦾 Intuitive Surgical: Growth Deceleration

10. 🛩️ United Airlines: Premium Absorbs Fuel Spike

⚡️ TSMC: AI Megatrend Deepens

TSMC’s Q2 revenue rose 34% Y/Y to $40.2 billion ($900 million beat), while EPS per ADR surged 74% Y/Y to $4.31 ($0.37 beat). Gross margin expanded to a stunning 68%, operating margin hit 60%, and net profit margin reached 56%. It was the fifth straight quarter of record earnings. After a near-40% run this year, the stock is priced for blowouts, and even this one triggered classic “sell-the-news” profit-taking.

CEO C.C. Wei maintained his bullish tone, saying that “our conviction in the multi-year AI megatrend remains very high.” Advanced nodes (7nm and below) now generate 77% of wafer revenue (up 3pp Q/Q), with 3nm at 30% and 2nm debuting at 3% ahead of a steep ramp in the back half. Management flagged agentic AI as a new demand driver, reviving CPU orders in data centers alongside accelerators. Mature nodes (45/40nm, 28nm, 16nm) all declined sequentially, an early sign that higher memory prices may be pinching mainstream semiconductor demand.

The two big strategic announcements were another $100 billion US investment and a massive CapEx raise:

• Total US commitment now reaches $265 billion, funding 12 leading-edge and packaging facilities in Arizona.

• FY26 CapEx was raised to $60–$64 billion (from $52–$56 billion).

• CFO Wendell Huang said CapEx over the next three years will be “significantly higher” than the past three, with 70–80% of the 2026 budget going to advanced nodes.

TSMC raised FY26 revenue growth guidance to slightly above 40% in USD (vs. ~35% consensus), up from >30% prior. It was the second hike this year. Asked about Samsung and Intel, Wei was dismissive. He explained that foundry partnerships aren't like "buying milk from the 7-11," and customers locked in for years won't switch on price.

The real risk for TSMC is cost. Overseas fab ramps and the 2nm expansion will temporarily weigh on gross margin, just as CapEx surges. Watch whether gross margin can hold near 67% in Q3 while record spending and the steepest phase of the 2nm ramp hit the P&L simultaneously. The widening of capital intensity could start to compress industry-leading margins even as the AI demand cycle extends into 2028.

2. 🔬 ASML: Second Guidance Hike

ASML's Q2 revenue rose 21% Y/Y to €9.3 billion (€400 million beat), with GAAP EPS up 29% Y/Y to €7.59 (€0.60 beat). Gross margin came in at 54%, above guidance, on high-margin Installed Base Management sales. Free cash flow reached €1.3 billion, and ASML repurchased €1.1 billion of shares. Shares are up nearly 70% YTD.

The big news was the second guidance raise of the year, along with concrete multi-year capacity plans:

• FY26 revenue guidance was lifted to €43–€45 billion (from €36–€40 billion, and vs. €39 billion consensus).

• FY26 gross margin was raised to 54–56% (from 51–53%).

• Q3 revenue is now €11–€12 billion (vs. €10.4 billion consensus).

• Low-NA EUV capacity is projected to grow ~30% in 2027 to ~85 units, with another ~30% increase in 2028 (~110 units).

• DUV immersion capacity is also expanding 30% in each of the next two years.

The AI infrastructure buildout has been a flywheel for ASML. Advanced logic revenue is expected to grow ~25% this year and memory ~75%, with customers “aggressively adding capacity” across 5nm, 4nm, 3nm, and 2nm nodes and already planning for 1.4nm.

Intel Foundry began production on the Intel 18A node using ASML’s most advanced High-NA EUV tool. Meanwhile, TSMC has said it will hold off on High-NA through 2029. CFO Roger Dassen also flagged potential pricing power on Low-NA tools given “current environment” dynamics. It was a nod to the reported pricing dispute with TSMC. China’s share of system sales dropped further to 14% (from 19% in Q1), tracking below the ~20% full-year target.

ASML confirmed its Capital Markets Day for June 10, 2027, when it will update long-term outlook. The next question is whether the pricing conversations with customers translate into meaningful ASP gains in 2027, or whether TSMC's pushback caps how much of the AI demand tailwind flows through to ASML's margin structure.

3. 💊 J&J: Oncology & Icotyde Deliver

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🏦 Banking on Volatility

• 🏴‍☠️ Can Ubisoft Turn it Around?

• 🦚 Comcast: The Great Unbundling

• 🧠 SK Hynix Bets AI Broke the Cycle

Subscribe now

Netflix no longer needs to prove that streaming can be profitable. It needs to prove that a mature streaming business can continue to grow. The market has expressed skepticism, with shares down more than 40% in the past year. The Q2 FY26 print offered no relief, with the stock falling roughly 8% after hours.

The company walked away from Warner Bros. Discovery last quarter, avoiding an expensive bidding war and collecting a $2.8 billion breakup fee. But the decision also removed a potential shortcut to the next phase of growth. Netflix must now generate it internally through higher prices, advertising, live programming, and a wider variety of content.

That task is becoming more urgent as competition for attention intensifies. YouTube continues to gain television share, free services like Tubi are growing, and traditional media companies are using sports to pull audiences into their platforms. Netflix’s response is to give viewers more reasons to open the app without weakening the economics that made the business so attractive.

Today at a glance:

• 🍿 Netflix Q2 FY26

• 📺 Becoming the bundle

• 📈 The battle for attention

• 💳 Stripe wants to buy PayPal

🍿 Netflix Q2 FY26

Income statement:

• Revenue +13% Y/Y to $12.6 billion ($20 million miss).

• Operating margin 33% (-1pp Y/Y).

• EPS +11% Y/Y to $0.80 ($0.01 beat).

Balance sheet:

• Cash and short-term investments: $9.1 billion.

• Debt: $14.4 billion.

FY26 guidance:

• Revenue +13%-14% to ~$51.2 billion with a narrower range.

• Operating margin 31.5% (+2pp Y/Y, unchanged from last quarter).

So, what to make of all this?

• 📊 Solid, but no raise: The numbers came mostly in line with consensus. The operating margin exceeded Netflix’s own forecast, but the full-year outlook was only narrowed to the same midpoint. The slight Y/Y margin compression reflected front-loaded content amortization, with expansion set to resume in the second half. Management expects Q3 operating margin of 33%, up from 28% a year ago. In a world where outlooks are meant to be beaten and raised, it wasn’t quite what Wall Street was hoping for. The expected 12% Q3 revenue growth would be the slowest since 2023.

• 📢 Ads doubling on schedule: Netflix reiterated its expectation for roughly $3 billion in 2026 ad revenue, with US upfront commitments expected to close within weeks. Programmatic access is expanding to Pause Ads and live inventory this summer, reducing the manual friction that kept smaller advertisers out. How far the business scales will depend heavily on how automated it can become.

• 📺 Live punches above its weight: Members watched 97 billion hours in H1, up 2% Y/Y and accelerating from 1.5% growth in 2025. Live programming will consume just over 5% of 2026 content spending and drive only about 1% of view hours, yet it accounts for six of Netflix’s 10 largest new-member sign-up days over the past five years. With the NFL and Fury-Joshua, Netflix is buying sign-ups more than hours.

• 🤖 GenAI at scale: GenAI workflows touched ~300 titles in 2026, concentrated in post-production. Management framed it as shots and sequences that otherwise wouldn’t exist (enhanced crowds, historical battles, worldbuilding), delivered faster and cheaper than traditional methods. Netflix framed the technology as a capability expansion rather than a cost-cutting exercise.

• 💰 Record buybacks: Netflix repurchased $4.7 billion of stock, its largest quarter ever, with $27.1 billion of capacity remaining after April’s fresh $25 billion authorization. Free cash flow fell to $1.5 billion from $2.3 billion, but the drag was largely cash taxes tied to the Warner Bros. termination fee. The full-year ~$12.5 billion FCF target is unchanged.

• 📉 Retiring the engagement scoreboard: The What We Watched report moves from biannual to annual starting in 2027, deliberately decoupled from earnings. Management wants the quarterly conversation anchored on revenue and operating profit, not view hours. It’s the same signal as removing subscriber counts. Netflix wants to be judged as a scaled compounder, and it is steadily removing the metrics that invite a different debate. The change is more controversial because management has repeatedly presented engagement as its North Star. Reducing disclosure makes it harder to independently judge what is driving performance. Subscriber churn also remains undisclosed, an unusual omission for a subscription business.

📺 Becoming the bundle

Netflix spent years replacing scheduled television with an ad-free, on-demand library. Now it is selectively rebuilding many of the features it disrupted.

As of its last disclosure in May, the ad tier reached more than 250 million monthly active viewers, up from 190 million in November. Netflix expects advertising revenue to approximately double to around $3 billion this year, but audience growth alone is not enough. The next challenge is converting that reach into better targeting, higher fill rates, and more revenue per viewer.

That helps explain Netflix’s selective push into live sports. The company is not trying to replicate ESPN or Prime Video with full-season packages. Instead, it is focusing on tentpole events that attract large audiences, generate premium advertising inventory, and feel culturally significant.

Its new MLB deal includes Opening Night, the Home Run Derby, and the Field of Dreams game. Netflix will also carry five NFL games this season, including Thanksgiving Eve, Christmas Day, and a potentially decisive Week 18 matchup.

The approach creates appointment viewing without paying to fill hundreds of hours with lower-profile games. It may be a more capital-efficient way to use sports as an acquisition and advertising tool, although Netflix still has to prove audiences will follow isolated events to a platform that is not their year-round sports destination.

The strategy extends beyond sports. Netflix recently added TF1’s live channels and on-demand catalog directly to its service in France. Rather than acquiring a traditional media company, Netflix can integrate third-party programming and serve as the interface through which viewers access it..

The company is also expanding into video podcasts, short-form clips, and licensed videos from publishers. Together, these moves suggest Netflix wants to become more than the place viewers visit for a new season of Stranger Things. It wants to become an app they open every day.

That may increase engagement at a lower cost than premium scripted content. But it also creates strategic tension. Netflix is adding ads, scheduled programming, third-party channels, sports, podcasts, games, and vertical video. If those pieces do not feel cohesive, it risks recreating the bloated television bundle it once replaced.

📈 The battle for attention

Netflix’s engagement challenge is visible in Nielsen’s latest data. Streaming represented 47.6% of US TV time in April, up from 44.3% a year ago and near record levels, while cable fell from 24.5% to 21.6%.

Netflix captured 7.8% of US TV time in April, up slightly from 7.5% a year ago but below its 9.0% December peak, when Stranger Things and Christmas Day football lifted viewing. Monthly figures are heavily influenced by release schedules, but the broader competitive trend is harder to dismiss.

YouTube increased its share to a record 13.4% in April. Amazon Prime Video reached 4.2%, helped by its new NBA package, while Tubi hit a platform record of 2.3%. The competitors gaining the most attention are not relying exclusively on expensive scripted series. They combine creator content, sports, free programming, or multiple forms of entertainment.

Netflix argues that retention and customer satisfaction matter more than raw hours watched. That is reasonable for a subscription business. But engagement still supports nearly every part of its strategy. More viewing creates more advertising inventory, improves the return on content spending, reduces churn, and makes future price increases easier to absorb.

Bottom line: Netflix’s ad tier, pricing power, and operating leverage can support low double-digit growth. But at a premium valuation, investors need evidence that its broader entertainment strategy is producing more engagement, stronger monetization, and durable returns on spending. Otherwise, Netflix risks becoming more complex without becoming more valuable.

💳 Stripe wants to buy PayPal

PayPal may be about to cash out. Reuters first reported that Stripe and private equity firm Advent International have offered $60.50 per share to acquire the company, valuing it at more than $53 billion. The proposal is reportedly backed by approximately $50 billion of committed financing.

This is still an unsolicited offer rather than an agreed deal. PayPal has not formally responded, and the companies involved have declined to comment. However, PayPal has reportedly been working with Goldman Sachs and Evercore to evaluate strategic options, including a potential sale or breakup. The company is not necessarily looking for a buyer, but it appears willing to consider what one might pay.

🛒 Stripe Gets the Consumer

The strategic logic is straightforward. Stripe built its business by helping merchants accept payments online, becoming one of the most important infrastructure providers in digital commerce. Consumers use Stripe constantly, but most barely know it exists. PayPal brings the missing half of the network: 439 million active accounts, approximately $1.8 trillion in annual payment volume, and Venmo, one of the strongest consumer payment brands in the United States.

A combination would give Stripe a much larger presence on both sides of a transaction. Stripe could continue powering payments for merchants while PayPal, Venmo and Stripe’s Link wallet deepen its relationship with consumers. Braintree would add another large merchant-processing platform, while PayPal’s stablecoin and crypto assets would complement Stripe’s recent acquisitions of Bridge and Privy.

In theory, that creates one of the most comprehensive payment ecosystems in the world. In practice, Stripe would also inherit overlapping products, aging technology, and a collection of businesses PayPal has struggled to integrate. Buying PayPal could accelerate Stripe’s expansion into consumer payments, but combining the two companies without distracting the faster-growing business would be a major undertaking.

💵 Private Equity Sees the Cash Flow

For Advent, the attraction is less about strategic fit and more about cash flow. PayPal generated approximately $5.6 billion of reported free cash flow last year, or $6.4 billion on an adjusted basis. A $53 billion purchase price values the company at roughly 8 times adjusted free cash flow, before considering debt, transaction costs, and the cost of financing.

That is the kind of setup private equity looks for. It’s a durable but unloved business, with strong cash generation, and a cost structure that can still be cut. Away from the public market, PayPal could restructure more aggressively without every investment or layoff being judged against the next quarter. Stripe and Advent could eventually relist the company, sell individual assets, or separate the consumer and merchant businesses.

🤔 Is $60.50 Enough?

The offer looks generous relative to PayPal’s depressed share price, but less so relative to the cash flow and strategic assets the buyers would receive. PayPal traded above $78 only a year ago, and its network of consumers and merchants would be extremely difficult to recreate from scratch.

Michael Burry, who owns PayPal shares, has already called the proposal too low. His valuation may prove optimistic, but the broader point is reasonable: Stripe and Advent would not be offering $53 billion unless they believed PayPal could ultimately be worth substantially more.

Bottom Line: PayPal became vulnerable because investors stopped believing in its turnaround. Stripe and Advent are betting that the assets are worth more than the public company built around them. At $60.50 per share, shareholders receive a clean exit from years of underperformance. But the first offer may establish only that PayPal is in play, not the price that ultimately gets it sold.

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I own AMZN, META, and NFLX in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 💡 WTF is an ETF?

• 🏴‍☠️ Can Ubisoft Turn it Around?

• 🧠 SK Hynix Bets AI Broke the Cycle

• 📈 Stock Buybacks & Dividends Explained

Subscribe now

A new earnings season is here with the big banks kicking us off.

Later this week, we’ll have a look at Netflix and the picks and shovels of the AI era, TSMC and ASML.

Four of the largest US banks just booked a combined $43 billion in profit, smashing records and beating estimates. They did it despite the Iran war, sticky inflation, and mounting doubts about the staying power of the AI boom.

Even as profits hit new records, the bank chiefs kept pointing to risks shifting beneath the surface: geopolitics, sticky inflation, and elevated asset prices.

High rates and market volatility are a windfall for the banks, while the gap keeps widening between households riding record markets and those squeezed by the cost of living.

Let’s break down the results.

Today at a glance:

• JPMorganChase: As Good As It Gets

• Bank of America: A Blowout With Caveats

• Citigroup: Four Out of Five

• Wells Fargo: Fees Do the Heavy Lifting

The Big Picture

As a reminder, banks make money through two main revenue streams:

1. 💵 Net Interest Income (NII): The difference between interest earned on loans (like mortgages) and interest paid to depositors (like savings accounts). It’s the primary source of income for many banks and depends on interest rates.

2. 👔 Noninterest Income: The revenue from services unrelated to interest. It includes fees (like ATM charges), advisory services, and trading revenue. Banks relying more on noninterest income are less affected by interest rate changes.

Here are the significant developments shaping Q2 FY26:

• 💰 Records fall across the board: The four biggest lenders beat expectations despite the Iran war and sticky inflation. The common engine was the trading floor. Every desk cashed in on a volatile quarter, and the results were strong enough that Jamie Dimon told analysts conditions are “getting close to as good as it gets.”

• 🎰 Trading and dealmaking do the work: Equities desks set records nearly everywhere, led by JPMorgan (+86% Y/Y). A dealmaking revival capped by the SpaceX IPO pushed investment banking to its best showing since 2021. Volatility that could have been a threat became the quarter’s biggest profit source.

• 🏦 Margin divergence: Beneath the capital-markets boom, the lending picture split. JPMorgan raised its full-year net interest income guide, but net interest margins compressed at Bank of America, Citi, and Wells Fargo as deposit costs stayed stubbornly high. The easy-money era of NII growth is uneven, and the banks leaning hardest on rates are feeling it most.

• 💵 Capital floods back to shareholders: Confidence showed up in the payouts. JPMorgan authorized a fresh $50 billion buyback and a 10% dividend hike, Wells Fargo repurchased ~$7 billion in the first half and lifted its dividend 11%, and Citi and BofA kept returning capital aggressively. With balance sheets flush, the big banks are signaling they see more room to run.

• 🛢️ Energy relief, already reversing: June CPI cooled to +3.5% Y/Y, driven almost entirely by a 9.7% drop in gasoline after the Iran ceasefire reopened the Strait of Hormuz. That relief is unwinding fast. The ceasefire collapsed on July 8, oil is climbing again, and the July inflation print will look very different.

• 📉 The K-shaped consumer holds: Card spending rose 9% Y/Y at both BofA and Wells Fargo, and provisions came in below expectations across the group, signs the consumer is still spending and paying on time. But management remained cautious about affordability for lower-income households, and the divide between asset-rich households and everyone else remains.

• 🌋 Risks below the surface: Even while celebrating records, Dimon warned of “risks shifting below the surface like tectonic plates,” naming geopolitical conflict, sticky inflation, large fiscal deficits, and elevated asset prices. Charlie Scharf echoed him, cautioning that favorable conditions “do not go on forever.”

• 🔑 Takeaway: The big banks are posting record profits amid a trading and dealmaking boom, returning capital freely, and leaning on a consumer that’s still holding. But margins are thinning where the boom isn’t reaching, and management keeps pointing to the same tensions that leave little room for error.

Let’s visualize them one by one and highlight the key points.

JPMorganChase: As Good As It Gets

• Net revenue grew 28% Y/Y to $57.3 billion ($6.7 billion beat):

  • Net interest income (NII): $25.5 billion (+10% Y/Y).

  • Noninterest income: $31.8 billion (+47% Y/Y), lifted by a $4.6 billion one-time Visa gain.

• Net income: $21.2 billion (+41% Y/Y).

• Adjusted EPS: $6.14 ($0.34 beat).

• Key developments:

  • 📈 Equities smash the record: JPMorgan’s stock traders pulled in $6.0 billion (+86% Y/Y), beating even the highest analyst estimate and pushing total trading revenue to a fresh record of $12.1 billion (+35% Y/Y). The volatility that started with the Iran war and rippled through global equity markets, including a chaotic stretch in Korean stocks, handed the desks their best quarter ever.

  • 💳 The Visa windfall: A long-held Visa stake paid off to the tune of $4.6 billion, plus another $1.0 billion in equity investment gains. Reported EPS of $7.70 crushed consensus, but ~$1.56 of that was one-time. Even excluding the windfall, the bank still delivered a 23% return on tangible equity.

  • 🚀 Dealmaking roars back: Investment banking fees jumped 30% Y/Y to $3.3 billion, riding the record SpaceX IPO, heavy index rebalancing, and a wave of AI-related financing. M&A advisory rose 20%, though that fell short of the 27% expected by analysts. CFO Jeremy Barnum called the environment “dynamic and interesting across a whole variety of dimensions.”

  • 🏦 NII guidance raised: Management lifted the NII outlook to ~$105.5 billion (from $103 billion), as higher-for-longer rates extend the tailwind. The card net charge-off forecast also improved, to ~3.2% from 3.4%, a sign that consumer credit is holding better than feared.

  • 💸 Expenses are the catch: Full-year expense guidance climbed to ~$107.5 billion (from $105 billion). Management framed the increase as the cost of doing more business, tied to the elevated activity levels that drove the revenue outperformance.

  • 💰 Capital floods back to shareholders: JPMorgan raised its quarterly dividend 10% to $1.65 and authorized a fresh $50 billion buyback, with CET1 at a comfortable 14.1%. With Dimon having pegged excess capital near $40 billion, the firm has ample room to keep returning cash even as Basel III capital rules loom.

  • 🔑 Takeaways: Every business line set records, and the NII guide flipped from headwind to tailwind versus Q1. Underneath a one-time Visa boost, JPM showed it can monetize volatility, dealmaking, and higher rates all at once.

• Key quote:

  • CEO Jamie Dimon: “These results were the product of a particularly favorable environment with an elevated level of market activity, as well as rigorous execution, years of consistent investment, and thoughtful capital deployment.”

Bank of America: A Blowout With Caveats

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🏴‍☠️ Can Ubisoft Turn it Around?

• 🧠 SK Hynix Bets AI Broke the Cycle

Subscribe now

Premium members get:

• 📊 Monthly reports: 200+ companies visualized.

• 📩 Tuesday articles: Exclusive deep dives and insights.

• 📚 Access to our archive: Hundreds of business breakdowns.

PRO members get everything PLUS:

• 📩 Saturday PRO reports: Timely insights on the latest earnings.

Today at a glance:

1. 🥤 PepsiCo: Gas Prices Bite

2. 🛩️ Delta: Premium Absorbs Fuel Shock

3. 🍪 General Mills: Reinvestment Phase Ends

1. 🥤 PepsiCo: Gas Prices Bite

PepsiCo’s Q2 revenue rose 6% Y/Y to $24.2 billion ($230 million beat), with non-GAAP EPS of $2.20 ($0.01 miss). Organic revenue growth of 2.4% included effective net pricing plus modest volume gains, with FX adding 2.2 points and M&A 1.8 points. Shares fell by more than 3% anyway.

The North American segment disappointed after Q1’s tentative snack rebound:

• Frito-Lay volume went flat, and revenue declined 2%, reversing the momentum from Q1’s 2% volume gain.

• North American Beverages volume slid 4%, with operating margin down 90 bps.

• International remained the engine, projected to top $40 billion in revenue this year, with Asia Pacific Foods delivering double-digit volume growth.

CEO Ramon Laguarta said, “The consumer is worse than what we had anticipated, and it’s driven mainly by gas prices.” US gas prices surged above $4/gallon during Q2 due to the Iran conflict, and Laguarta said the pullback was concentrated in convenience stores and other impulse-purchase channels. PepsiCo is now tweaking its 15% price cuts by segment and has noted delays in regaining the shelf space retailers had promised.

The healthier “permissible portfolio” (protein-fortified snacks, portion-controlled multipacks) hit $3 billion in value and is growing double digits, which Laguarta flagged as a bright spot. Activist Elliott’s pressure to accelerate the turnaround continues in the background.

PepsiCo reaffirmed full-year FY26 guidance, with organic revenue growth of 2–4% and core constant-currency EPS growth of 4–6%, though management flagged that results are likely to land at the low end of the EPS range. Tariff refunds will contribute roughly a full point of EPS growth to offset commodity inflation. The next question is whether the impulse channel recovers as gas prices ease, or whether Frito-Lay needs another pricing reset.

2. 🛩️ Delta: Premium Absorbs Fuel Shock

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• ⚽️ How FIFA Makes Money

• 🔒 Micron: Locking In the Boom

• ⚖️ How to Analyze a Balance Sheet

• 🧠 SK Hynix Bets AI Broke the Cycle

Subscribe now

💿 Sony just killed the disc

Starting January 2028, no new PlayStation game will ship on a disc, third-party publishers included. That timeline caught some fans off guard, but the disc gave up its usefulness years ago. Critical day-one patches are now standard, so the plastic in the box is already just a glorified install key with better packaging. Digital accounts for ~80% of Sony's full-game sales, and many smaller titles skip the box altogether.

Take-Two Interactive got a head start on Sony. Its upcoming game, GTA VI, will ship this November with a download code in a box and no disc at all. It’s an admission that a disc has become more of a liability than an asset. A physical disc means locking the code weeks before launch and hoping it doesn’t leak in the meantime.

Killing the disc kills the resale market that’s quietly taxed every publisher for decades. A used copy undercuts the full-price version by a few dollars within days of launch, and the publisher sees nothing from the resale. Publishers tolerated that leak while the market was growing. But Console was gaming's slowest-growing segment in 2025, up just 3% Y/Y to $45 billion, or 22% of the ~$200 billion global games market now dominated by mobile.

AAA budgets keep climbing, and studio layoffs have become routine at big publishers. Microsoft just proved how bad that math has gotten. Xbox is cutting 3,200 jobs, 20% of its workforce, and divesting four studios. CEO Asha Sharma told staff the business is "not healthy," operating at margins three to ten times below those of comparable platform and publishing peers. A business model that stops leaking resale value to GameStop is exactly the fix a stagnant industry needs.

Today at a glance:

• 💰 Pricing the pure plays

• 🏴‍☠️ Can Ubisoft turn it around?

💰 Pricing the pure plays

While consoles have trailed mobile gaming for years, GTA VI is shaping up to be the biggest single entertainment release in history, game or otherwise. Some analysts see it selling over 40 million units at launch, with the lowest-priced edition running at $80.

The market is already pricing that in. Take-Two now trades at 8.0x its average net bookings over the past three years, the highest multiple in this peer group.

Here's how the net bookings multiples of the leading console publishers stack up:

• 🇺🇸 Take-Two (8.0x): With an enterprise value near $47 billion, Take-Two is the largest publicly traded standalone gaming publisher. FY27 guidance has net bookings jumping to ~$8.1 billion from $6.7 billion in FY26, which would pull the forward multiple closer to 6x. The market is already pricing a big GTA VI launch.

• 🇺🇸 EA (7.2x): The Saudi PIF-led consortium took the company private at a $55 billion enterprise value. Live-service sports franchises give EA the most predictable revenue base in the group, which is exactly what a private buyer pays up for.

• 🇯🇵 Capcom (7.2x): The Japanese publisher is valued at nearly $7 billion and trades at a premium multiple, supported by a standout margin profile (~39% operating margin in FY26). A successful remake strategy turned Resident Evil from nostalgia into a repeatable growth engine, while Monster Hunter and Street Fighter continue to prove that Capcom can refresh mature franchises with strong commercial results. Capcom’s efficiency and consistent execution are what earn a multiple this rich.

• 🇯🇵 Square Enix (2.0x): Square Enix trades at the low end of the group, with a ~$4 billion valuation. Its RPG franchises, Final Fantasy and Dragon Quest, carry the portfolio. Operating margin sits around 13%, roughly half the peer average, dragged down by uneven execution, bloated budgets, and several high-profile misses. An activist investor is now pushing management to close that gap.

• 🇫🇷 Ubisoft (0.6x): The French publisher has fallen from a €12 billion enterprise value in 2018 to just over €1 billion today, a collapse driven by escalating budgets on major titles, repeated project cancellations, and franchise fatigue across Assassin's Creed and Far Cry.

Bottom Line: Every publisher in this cohort benefits from Sony killing the disc. But the multiple investors are willing to pay comes down to trust. Great IP only matters if it ships on time and turns into cash.

🏴‍☠️ Can Ubisoft turn it around?

Assassin’s Creed Black Flag Resynced, a full remake of the studio’s 2013 pirate adventure, just launched on July 9. The original game has been played by over 34 million unique players. Ubisoft wants to follow in Capcom’s footsteps by modernizing its biggest hits. Commercial success of remakes is more predictable, and upfront production costs are usually lower, making it a potent combo. Ubisoft’s CEO, Yves Guillemot, officially confirmed that the company is developing multiple remakes of older Assassin’s Creed games.

The remake is arriving while the company is priced as if bankruptcy is on the horizon. Ubisoft has an enterprise value of ~€1.2 billion, with average annual net bookings of nearly ~€2 billion. Even multiplying this valuation by 3 would put it merely in line with struggling publisher Square Enix.

That price implies a company in slow-motion liquidation rather than an underperforming publisher. As of July 2026, nearly 14% of Ubisoft stock on Euronext Paris is short (people betting the stock will go down).

So what’s happening here? Let’s look at the underlying business.

New Structure, New Capital

Ubisoft spent January reorganizing into five Creative Houses, decentralizing decisions that used to run through the center. Vantage Studios ended up holding the crown jewels: Assassin’s Creed, Far Cry, and Rainbow Six. The other four Houses split the rest of the portfolio across multiplayer shooters (Ghost Recon, The Division), live services (The Crew), and casual and family titles (Rayman).

Tencent closed a €1.2 billion investment in Vantage Studios in November 2025, taking a 26% economic interest while Ubisoft retains exclusive control and keeps consolidating the studio’s results. The deal valued Vantage Studios at a pre-money enterprise value of €3.8 billion. That structure suggests that both sides see more value in Vantage than the market is currently assigning to Ubisoft as a whole (€1.2 billion). But the Guillemot family's control and Tencent's right of first refusal make an EA-style buyout unlikely.

Tencent’s cash injection did real work on the balance sheet. Ubisoft closed FY26 (ending in March) with an adjusted net debt of €0.2 billion and cash reserves of €1.3 billion, a very different position from the one that had investors bracing for a covenant breach a year ago.

The relief could be temporary, though. Bondholders can demand ~€0.5 billion back in November. Another ~€0.7 billion follows in late 2027. That is nearly the entire cash pile due before the FY28/FY29 slate can fully repair free cash flow. Add a year of guided burn, and the math stops working. Ubisoft's near-term future hinges on refinancing.

The Back Catalog is already carrying the business

In FY26 (ending in March), Ubisoft’s back catalog generated €1.3 billion in net bookings, compared with just €0.2 billion from new releases and other revenue. That’s 84% of the entire business running on games released in previous years. It illustrates the durability of a digital catalog library that gives the company a revenue floor, even without new tentpole releases. The disc-free future acts as a tailwind for this revenue line, with no second-hand market eating away at the long tail of legacy franchises.

A pipeline with execution risks

• FY24: Net bookings peaked at €2.3 billion, driven by Assassin’s Creed Mirage and The Crew Motorfest, with strong back-catalog performance.

• FY26: Net bookings fell to €1.5 billion, with no notable new releases.

• FY27: Guidance for the fiscal year that started in April points to a high-single-digit decline, implying roughly €1.4 billion in net bookings. Black Flag Resynced is the main new release, and Ubisoft’s guidance appears to leave room for upside. The game does not need to be a massive blockbuster to improve the narrative. That is the advantage of low expectations.

Ubisoft plans for a rebound in FY28 and FY29, when management expects to ship Assassin’s Creed Hexe, Far Cry 7, and a new Ghost Recon, lifting bookings closer to the ~€2 billion of previous years. A success for Black Flag Resynced could remind the market that the IP still works. It would certainly be a good idea ahead of a Netflix tie-up with an Assassin’s Creed show expected in the coming months. The main risk ahead is that the upcoming games could be postponed if the quality is not there, leaving the company in a cash flow hole.

The cash burn is real

Free cash flow has been negative in four of the last five fiscal years, including -€0.5 billion in FY24 and -€0.4 billion in FY26. Management already stated that free cash flow consumption in FY27 will be no more than -€0.5 billion, but that’s a low bar.

On the bright side, Ubisoft expects cumulative free cash flow to turn positive from FY27 through FY29. That would be good news for liquidity and solvency. But that inflection point still depends on a set of games to actually launch over that period.

To achieve such a rebound, Ubisoft has been cutting fixed costs aggressively, reducing them from €1.75 billion in FY23 to €1.44 billion in FY26. The target is €1.25 billion by the end of FY28. Whatever happens to bookings, Ubisoft is guiding to a structurally smaller cost base than the one that burned through FY23 and FY24. To do so, they have reduced their workforce from ~20,000 to roughly ~16,000.

The cost reset lowers the hurdle, but it does not solve the problem on its own. Ubisoft still needs enough successful releases to spread that smaller cost base over the same bookings pool.

Why it matters:

• The market factors a solvency risk: It also assumes the company may face significant dilution if it needs to raise capital to meet near-term obligations. The catalog is de-risked, but there is no margin of safety for the next few high-profile titles. Ubisoft still has to navigate near-term debt maturities while proving that FY27 is the trough and FY28/FY29 can restore positive cash flow.

• There is an asymmetric upside if FCF rebounds: Sub-1x EV/bookings against a back catalog this large only makes sense if the market expects the catalog itself to erode. But the back-catalog net bookings trend suggests the portfolio has been very resilient and that only one high-profile game per year is needed to keep the company in the black moving forward, which seems like a very low bar for ~16,000 employees.

• Tencent put a real price on the crown jewels: The minority stake in Vantage Studios signals that Tencent sees value in the main IPs and studios. That value was €3.8 billion just a few months ago.

• A challenging history: Ubisoft has been plagued with toxic workplace allegations, game controversies, repeated delays, and the founding Guillemot family keeps tight control, with family members placed in key leadership positions.

Bottom Line: Ubisoft doesn’t need a miracle. FY27 guidance looks conservative, the catalog still produces more than €1 billion a year, and fixed costs are finally moving lower. But the company still needs to ship the next major slate by March 2029 and prove the new cost base can turn bookings into cash. At 0.6x average net bookings, the market is not paying for that yet. After years of delays, cancellations, and uneven launches, execution remains the hard part. If cash burn persists, the balance sheet becomes the only story that matters.

That's it for today.

Happy investing!

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I own AAPL, GOOG, META, and TCEHY in App Economy Portfolio. I also personally own UBSFY. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🤖 NVIDIA: Gone Parabolic

• ⚽️ How FIFA Makes Money

• 🔒 Micron: Locking In the Boom

• 📊 PRO: AMD, Shopify, Uber, Airbnb…

Subscribe now

SK Hynix is trying to pull off the largest ADR listing in history.

The planned Nasdaq listing targets over $28 billion (45 trillion won), edging past Alibaba’s ~$22 billion New York debut in 2014. Trading is expected to begin around July 10. Every dollar is earmarked for fabs, packaging, and lithography machines. The funny part is that the company doesn’t really need the money.

SK Hynix is the world’s #1 HBM maker (High Bandwidth Memory) at ~57% share. The company sold out of capacity three years in advance, and it just surpassed Samsung to become South Korea’s most valuable company. It also already trades right in line with Micron, near 7x forward earnings. The re-rating a Nasdaq listing is supposed to deliver already happened.

We broke down Micron’s blowout quarter last week, noting that the company benefited from the AI memory shortage.

This week, the leader that ships the majority of NVIDIA's HBM is asking US investors to price it like the AI infrastructure play it has become.

Today at a glance:

• 📈 The US Listing

• 🧠 How SK Hynix Makes Money

• 📊 Q1 FY26 in Numbers

• 🇰🇷 What You're Paying

• ⚔️ Competition & Risks

• 🔭 What to Watch

• 🧭 Personal Take

📈 The US Listing

SK Hynix shares trade in Seoul, in won, on the Korea Exchange. That makes them awkward and expensive for US institutions to touch directly. An American Depositary Receipt (ADR) solves that. The company issues new shares, parks them with a custodian bank, and a US depositary issues receipts that trade on Nasdaq in dollars like any American stock.

SK Hynix plans to issue up to 178 million ADRs, the equivalent of 2.5% of the company. Each Seoul share trades near $1,580 (₩2,425,000) and splits into ten ADRs, so a single ADR is referenced at around $158 (₩242,500), with the final price set on July 10. The ADRs list on the Nasdaq Global Select Market under SKHY.

Every dollar is going into capacity.

The proceeds are split four ways:

• 🏭 Yongin Y1: the first fab at SK Hynix’s sprawling new Yongin cluster, its next major DRAM and HBM base.

• 📦 Cheongju P&T7: an advanced packaging plant built for HBM, where the die-stacking that sets SK Hynix apart actually happens.

• 🔬 ASML EUV scanners: the extreme-ultraviolet lithography tools that leading-edge DRAM can’t be made without.

• 🇺🇸 Indiana: a $4 billion packaging plant, SK Hynix’s first US fab.

Compare that to the SpaceX IPO. SpaceX is raising funds for its CapEx ramp, with $10 billion in negative free cash flow in Q1 alone. SK Hynix is raising money from a position of strength, with $24 billion in net cash and an order book sold out through 2028.

SK Hynix first floated a raise of around $10 billion, and the board eventually settled on $28 bllion. You don’t do that unless internal demand forecasts have moved well past the old memory-cycle playbook.

There’s a control wrinkle worth mentioning. SK Square, the holding company that owns ~20% of SK Hynix, must keep its stake above 20% under Korea’s holding-company rules. That constraint is why the deal issues new shares sized to protect that floor, rather than selling treasury stock. Some Korean shareholders are unhappy about the dilution. For US investors, the dilution is the price of admission to a stock they couldn’t easily buy before.

🧠 How SK Hynix Makes Money

The comeback

SK Hynix started in 1983 as Hyundai Electronics, the chip arm of the Korean industrial group. The Asian financial crisis reshaped it. Seoul forced a 1999 merger with rival LG Semicon, creating a top-tier DRAM maker overnight and burying it in debt. When memory prices collapsed ~80% in 2001, the renamed Hynix lost billions, fell under creditor control, and became a national symbol of corporate failure.

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 📊 Earnings Visuals (6/2026)

• 🚀 How SpaceX Makes Money

• 🔒 Micron: Locking In the Boom

• 📈 Stock Buybacks & Dividends Explained

Subscribe now

☁️ Meta is pulling an Amazon move

Meta has been shunned by the market this year for overspending on AI. But something good might come out of it.

Bloomberg reported Wednesday that Meta is building a cloud business to rent out its excess AI compute, muscling onto turf held by AWS, Azure, and Google Cloud.

Zuck poured hundreds of billions into data centers for his superintelligence dream, but he might have bought more than he needs. The stock surged nearly 10% on the news, as the new business could offset some of that rapidly growing CapEx line.

It wasn’t good news for everyone. CoreWeave fell 13% and Nebius dropped 15%. Meta spent the past year as one of their biggest customers. Now it could turn into a massive new rival.

It also sharpens the question of where we are in the cycle: did Meta overspend, or is Zuck embracing a new strategy?

Starting a cloud business also confirms that Meta expects to have capacity worth selling. The generous read is that Meta is betting demand keeps coming, the same bet Amazon made when it launched AWS nearly two decades ago.

Zuck called it “definitely on the table” back in May, noting outsiders ask to buy Meta’s compute almost every week. The company, spending up to ~$145 billion this year on its own moonshot, now has a Plan B: if superintelligence takes its time, collect rent.

Today at a glance:

• 🦚 Comcast: The Great Unbundling

• 👟 Nike: Not a Recovery Yet

🦚 Comcast: The Great Unbundling

Comcast is unwinding the ~$30 billion bet it spent a decade defending.

The company is splitting in two, separating NBCUniversal and Sky into a standalone media business, while the parent keeps broadband, wireless, and Comcast Business.

The stock jumped as much as 17% intraday, its biggest move since 2008, before settling up ~5%. Not bad for a name that was down ~22% on the year going in.

This is the second breakup step in less than a year. In January, Comcast spun off its cable networks (CNBC, USA, MS NOW, Golf Channel) into Versant, which fell 13% on its Nasdaq debut and has traded sideways since. The channels were the appetizer. Now the studios, parks, NBC, Peacock, and Sky are leaving too.

Who keeps what

• Comcast retains the cash machine: the largest converged network across 65M+ homes, the largest WiFi network, and a fast-growing wireless business in a $200B+ market. Michael Angelakis, the former CFO who helped architect the 2011 NBCUniversal purchase, returns as CEO to run the company now shedding it (you can’t make this up.)

• NBCUniversal walks away with the IP: Universal Studios, the theme parks (including Epic Universe), NBC, Telemundo, Bravo, Peacock, and Sky. Co-CEO Mike Cavanagh takes over. Brian Roberts, whose family controls both through a dual-class structure, stays “actively involved” in each.

The structure is a tax-free separation in roughly 12 months, with Comcast holding up to a 19.9% stake in NBCUniversal to sell down and cut debt. The buyback is suspended until then.

Why now

Q1 made the logic obvious:

• Connectivity delclined 1% Y/Y to $20.0 billion, a slow, dependable annuity now under siege from fixed-wireless home internet and Starlink.

• Content & Experiences grew +41% Y/Y to $13.0 billion, but only because NBC aired both Super Bowl LX and the Winter Olympics in February. Media revenue spiked +61% Y/Y on events that don't recur next quarter.

One business is a flat-growth utility throwing off free cash flow. The other is a lumpy, hit-driven content engine. Bundling them blurred the story for two very different investor bases, which is exactly the rationale Comcast now cites. After years of insisting the businesses belonged together, Cavanagh told analysts, “(we) simply changed our mind.”

The real story is M&A

Management swears it isn’t about future M&A. Roberts called the split “absolutely not” a prelude to deals. Wall Street seems unconvinced. Some analysts argued the separation matters less than what it enables: two focused companies, each with its own equity currency and board, free to pursue transactions that were unthinkable inside the conglomerate. It’s a way to unlock value and a clear ramp to dealmaking.

Wells Fargo pegs the sum-of-the-parts valuation at ~$25 per share, barely above today's price, and says the upside requires a transaction. A standalone NBCUniversal could use its own stock to buy content and reach the scale to fight Netflix and Disney. Or it might become a target itself.

Bottom Line: Comcast just conceded the conglomerate logic of the last decade is dead. The split should sharpen both businesses, but the market is already pricing the sequel. A leaner NBCUniversal, Sky in tow, becomes either the sector's next buyer or its next target the moment the ink dries.

👟 Nike: Not a Recovery Yet

Nike’s fourth quarter initially looked like a breakout. Reported EPS jumped more than 5x to $0.72. But the quarter included a $986 million tariff recovery tied to the International Emergency Economic Powers Act, which added roughly 900 basis points to gross margin. Adjusted for that one-time boost, EPS was closer to $0.20, a modest beat against the $0.13 consensus rather than the blowout the headline suggested.

The business underneath kept shrinking. Revenue fell 1% to $11.0 billion in the May quarter ($120 million beat) and dropped 4% in constant currency, nearly two years into Nike’s turnaround.

The margin mirage

Reported gross margin hit 49% (+9pp Y/Y), one of the highest in Nike's history, but ~9pp of that was the tariff refund. Underlying gross margin was ~40%, essentially flat and still near the lows of the reset.

• Wholesale rose 4% to $6.6 billion.

• NIKE Direct fell 7% to $4.1 billion, with Nike Brand Digital down 12%.

The trade kept margins compressed and shrank Nike's own channels, but it did help push North America back to growth.

What’s actually working

North America grew 3% to $4.8 billion, with wholesale up 10% as Nike rebuilt the retail relationships it spent years walking away from. Nike Running notched its fifth straight quarter of double-digit growth, adding roughly $1 billion to the category and five points of market share across North America and Western Europe. That was the clearest proof that Hill’s “Sport Offense” (the reorg that moved 8,000 employees into sport-specific teams) could create growth. The progress was real but uneven. A Boston Marathon ad got pulled after backlash, and some World Cup stock missed its retail delivery window.

International was still the drag:

• Greater China: Sales fell 12% Y/Y to $1.3 billion (down 17% in constant currency), with segment profit down 20%. Results beat lowered expectations, but local brands like Anta and Li-Ning continued to take share. The "comprehensive reset" Hill is running is expected to weigh on results through FY27.

• EMEA: Revenue slipped 1% Y/Y to $3.0 billion, but down 6% in constant currency, so the underlying business is losing even more momentum.

• APLA: Roughly flat at $1.6 billion, up 1% Y/Y reported and down 1% in constant currency.

Converse: time to call it?

Revenue plunged 32% Y/Y to just $244 million, and full-year sales were the lowest since 2011. Neil Saunders of GlobalData suggested that if Nike couldn't or wouldn't fix the brand, it should weigh an exit before Converse became a drain on management's attention. Hard to argue.

Soft outlook with a parting warning

Nike expected Q1 FY27 revenue to be down low-to-mid single digits and earnings roughly flat through Q2 FY27 (excluding the tariff benefit), though gross margin should finally start expanding as early as Q1.

Outgoing CFO Matt Friend was blunt: "We are not expecting the environment to improve meaningfully over the next six months." He warned that Nike's customers were under pressure worldwide. Friend will hand the reins to David Denton, Pfizer's CFO, on August 17. A fresh set of eyes, arriving at an awkward moment.

Bottom Line: Nike's profit surge was a one-time tariff windfall. The stock has fallen more than 30% this year, on track for a fifth straight annual decline. That said, North America and Running are the first hard evidence that the reset can produce real growth after months of inventory cleanup. At ~28x forward earnings, this is still a "show me" story, and the real test comes at November's investor day.

That's it for today.

Happy investing!

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I own AAPL, GOOG, META, NVDA, and TSLA in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members. 

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

🔥 The June report is here!

All the key earnings visuals from the past month in one report.

• ✔️ Cut through the noise with clear, concise financial snapshots.

• ✔️ See revenue trends, profit margins, and key takeaways instantly.

We visualized 200+ companies this season:

In case you missed it:

• 📊 Earnings Visuals (4/2026)

• 📊 Earnings Visuals (5/2026)

Download the full report below or log in to your account.

Your voice matters! Help us shape future reports. Got a company or sector you're curious about? Hit 'Reply' and let us know!

Here’s a sneak peek. 👀'

What to expect in this monthly report?

• ⚙️ Semis: Broadcom, Micron.

• 🛒 Retail: Home Depot, Lowe’s.

• 😎 Tourism: Carnival, Vail Resorts.

• 🤖 AI infrastructure: Oracle, Cerebras.

• 👟 Consumer brands: Nike, Lululemon.

• ☁️ Workflow: Adobe, DocuSign, GitLab, UiPath.

• 📊 Data: C3.ai, HPE, MongoDB, Rubrik, Samsara.

• 🛡️ Cybersecurity: CrowdStrike, Palo Alto Networks, SentinelOne.

• And more, like Accenture, Darden, Didi, Meituan, FedEx, Fox, and GameStop.

Download the full report below. 👇

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• ⚽️ How FIFA Makes Money

• 🚀 How SpaceX Makes Money

• 🔒 Micron: Locking In the Boom

• 📈 Stock Buybacks & Dividends Explained

Subscribe now

Cerebras pulled off the biggest semiconductor IPO

The May 2026 IPO priced at $185, surged to $386 on opening day, and closed at $311, a 68% pop.

That was the high-water mark. The stock spent the next six weeks giving back most of its gains, and its first earnings report triggered a record two-day plunge that briefly pushed it below the IPO price.

If that round-trip feels familiar, it’s the pattern we walked through in How To Invest in IPOs. The day-one pop is often short-lived, and the first earnings call is where the hype meets reality.

The quarter itself was impressive. Revenue nearly doubled, Cerebras beat expectations, and management raised full-year guidance above consensus.

The problem was not demand. It was the cost of keeping up with it.

Management argues the market misunderstood the softer margin outlook, which is partly due to Cerebras renting back its own chips from a customer while it waits for new data centers to come online.

Let’s review what we learned.

Today at a glance:

• 🧠 How Cerebras makes money

• 📊 Q1 FY26 by the numbers

• 🎙️ Key insights from the call

• 🔭 What to watch next

FROM OUR PARTNERS

The $7.1 Trillion Flowing Into AI is Not What You Think

When McKinsey & Company projects $7.1 trillion in AI investment by 2030, it’s not meant just for names like OpenAI, Anthropic, and Google.

Investors want to know the new movers in this industry. That’s why they’re looking at companies like BluSky AI.

These companies are building the infrastructure layer of the AI future. And their tech is only getting more important with time as demand for AI compute power outruns supply.

While typical data centers take several years to build, BluSky AI deploys modular, AI-ready data centers in months. BluSky AI utilizes less power, little to no water, and minimizes the impact on local communities.

Startups, enterprises, and governments will depend on tech like this. BluSky AI has a robust portfolio of unique sites with negotiated power ready for deployment. Become an early-stage investor in BlueSky AI today.

This is a paid advertisement for BluSky AI Regulation A offering. Please read the offering circular at https://invest.bluskyaidatacenters.com/

🧠 How Cerebras makes money

We’ve not covered Cerebras before in this newsletter, so let’s go back to basics.

A typical AI processor is roughly the size of a postage stamp. Cerebras builds one the size of a dinner plate: an entire silicon wafer turned into a single chip. Its Wafer-Scale Engine packs ~4 trillion transistors, against ~200 billion on NVIDIA's flagship Blackwell. Cerebras sells it inside a single machine, the CS-3, rather than as a bare chip.

The point of all that silicon is speed. Keeping a model on one giant chip avoids the slow shuffle of data between thousands of smaller ones, which lets Cerebras generate tokens roughly an order of magnitude faster than a GPU cluster. On frontier models, it clears 1,000 tokens per second. The pitch is simple: fast AI is worth a premium because it gets more done.

Cerebras turns that into two revenue streams.

• Hardware: Ceberas sells CS-3 systems and clusters to customers who run them in their own data centers. This segment was $111 million in Q1 FY26 (+59% Y/Y), and it’s the legacy side of the business.

• Cloud & other services: Customers can rent that same inference horsepower by the token from Cerebras-operated data centers. This segment hit $83 million (+178% Y/Y), growing three times faster than hardware and the strategic center of gravity going forward.

The moat is where it gets interesting. Cerebras sidesteps the exact chokepoints throttling everyone else. It uses no high-bandwidth memory (HBM), no CoWoS advanced packaging, and no bleeding-edge 3nm process, leaning instead on plentiful SRAM and mature 5nm. While the industry fights over the same scarce TSMC and HBM allocation, Cerebras isn’t standing in that line, and it’s the only AI accelerator maker building its systems entirely in the US.

The catch is who’s been paying. Cerebras was, until recently, almost entirely dependent on G42, an Abu Dhabi AI group. Its risk disclosures now name four significant customers: OpenAI, G42, MBZUAI University, and AWS. The bull case rests on the first and last names on that list, turning a UAE-heavy customer base into broader Western enterprise demand.

One quirk of the OpenAI deal shapes the numbers throughout, so it’s worth understanding up front:

• The deal: A multi-year agreement for more than $20 billion of Cerebras compute.

• The twist: To win it, Cerebras also handed OpenAI warrants (the right to buy its stock at a discount).

• The accounting: Those warrants count as a discount, so part of what OpenAI pays never shows up as revenue.

• The upshot: The reported figure understates the true size of the OpenAI business, and that gap widens as the deal ramps.

📊 Q1 FY26 by the numbers

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 📈 How to Invest in IPOs

• 🔒 Micron: Locking In the Boom

Subscribe now

Premium members get:

• 📊 Monthly reports: 200+ companies visualized.

• 📩 Tuesday articles: Exclusive deep dives and insights.

• 📚 Access to our archive: Hundreds of business breakdowns.

PRO members get everything PLUS:

• 📩 Saturday PRO reports: Timely insights on the latest earnings.

Today at a glance:

1. 🌐 Accenture: AI And Iran Hit Demand

2. 🚚 FedEx: Premium B2B Pays Off

3. 🛳️ Carnival: Europe Yields Soften

4. 🫒 Darden: LongHorn Carries The Year

1. 🌐 Accenture: AI And Iran Hit Demand

Accenture's Q3 revenue rose 6% Y/Y (3% in local currency) to $18.7 billion ($50 million miss), with GAAP EPS rising 9% Y/Y to $3.80 ($0.11 beat).

New bookings of $19.3 billion fell 2% Y/Y, the first year-over-year bookings decline since Q3 FY25. Shares plunged 18% post-earnings, the worst one-day drop on record, extending the stock's roughly 50% YTD decline.

CEO Julie Sweet flagged two distinct headwinds:

• Middle East impact: $100 million Q3 revenue hit plus ~$400 million in sales impact as the Iran conflict slowed decision-making across EMEA. Sweet expects “more impact” in Q4.

• Managed services deal slippage: A couple of large opportunities pushed into FY27 for company-specific reasons, creating timing-driven softness.

Accenture announced a $4.2 billion cybersecurity acquisition package: a majority stake in Dragos plus 100% of runZero and NetRise. Sweet called the combination "a first-of-its-kind OT Security platform." The three businesses generate ~$208 million in ARR with 48–53% Y/Y growth. Accenture also launched Accenture Edge, a new mid-market business targeting what it sizes as a $240 billion addressable market, and signaled it will tap the long-term debt market to fund elevated M&A. Total FY26 capital return was raised to at least $9.5 billion. Demand for “large-scale reinvention” continued: 104 quarterly bookings over $100 million year-to-date, up 13%.

Accenture cut the top end of FY26 revenue growth guidance to 3–4% (from 3–5%) and narrowed adjusted EPS to $13.78–$13.90 (from $13.65–$13.90). Q4 revenue forecast of $17.8–$18.4 billion fell short of the ~$18.5 billion consensus. CFO Angie Park said "more of the guided range" is in play, given macro uncertainty.

The stock now trades at its lowest multiple ever, roughly 6x EV/free cash flow. The combined dividend and buyback yield is near 10%. The big question is whether the AI displacement thesis continues to show up in slowing bookings, or whether the capital return profile is starting to make the bear case harder to justify at current levels.

2. 🚚 FedEx: Efficiency Over Expansion

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 💰 The M&A Land Grab

• 📈 How to Invest in IPOs

• 🚀 How SpaceX Makes Money

• 🔌 The Trillion-Dollar Off Switch

Subscribe now

Apple is raising iPhone prices, and the culprit is memory.

For decades, memory was the invisible commodity inside every device: cheap, abundant, and easy to ignore. Not anymore. A brutal memory crunch is quietly pushing up the price of nearly every device you own, as the world’s top tech giants fight for supply and leave smartphones, PCs, and automotive lines competing for what remains.

That squeeze has a winner on the other side.

Micron just reported the most extraordinary quarter in its history.

Here’s what stood out.

Today at a glance:

• ☁️ Micron’s blowout quarter

• 🔒 Trying to break the cycle

• 🌐 The squeeze everyone else feels

• 🔭 What to watch next

FROM OUR PARTNERS

Cloud Created AWS. AI is Building the Next Empire.

AWS built an infrastructure for the cloud and became one of the most valuable businesses in history. The segment made $128 billion in sales for 2025.

Now, another company is building what could be an equivalent infrastructure: a new highway for AI.

Traditional AI data centers are massive, take 3 to 7 years to build, and disrupt communities. But BluSky AI has designed a prefabricated SkyMod system: small footprints, powerful, and purpose-built for AI workloads. These data centers take months to stand up, not years.

Hundreds of thousands of enterprises use AWS to power their business. BluSky AI’s robust portfolio of sites will create new onramps for AI’s urgent needs. Invest in BluSky AI as they unlock a power source for 417M+ companies using AI.

This is a paid advertisement for BluSky AI Regulation A offering. Please read the offering circular at https://invest.bluskyaidatacenters.com/

☁️ Micron’s blowout quarter

Micron’s fiscal Q3 (ended May 28) was extraordinary by any standard.

• Revenue: The top line surged +346% Y/Y to $41.5 billion ($5.6 billion beat).

  • ☁️ Cloud Memory: $13.8 billion (+307% Y/Y).

  • 🖥️ Core Data Center: $11.5 billion (+653% Y/Y).

  • 📱 Mobile & Client: $11.5 billion (+254% Y/Y).

  • 🚗 Automotive & Embedded: $4.6 billion (+311% Y/Y).

• Margin trends: Gross margin was 85% (more than doubled Y/Y). Operating margin was 80% (+57pp Y/Y). These NVIDIA-esque margins drove earnings per share up 106% sequentially to $25.11 ($4.25 beat).

• Cash flow: Operating cash flow was $25.4 billion (+451% Y/Y). Adjusted free cash flow was $18.3 billion (vs. just $1.9 billion a year ago).

• Balance sheet: Cash, marketable investments, and restricted cash reached $30.2 billion. Long-term debt was $5.1 billion (down from $14.0 billion a year ago).

• Q4 FY26 guidance: Revenue ~$50.0 billion (~$6 billion above consensus). Gross margin ~86%. EPS ~$31 (vs ~$25.72 expected).

Almost none of the outperformance came from selling more chips. DRAM bit shipments rose only in the low single digits, yet DRAM prices jumped more than 60% in a single quarter, and NAND prices rose in the mid-80s%. Micron is already capacity-constrained, so the shortage isn’t showing up in units. It’s showing up in price.

Now step back and look at the scale.

Micron booked more revenue in three months ($41.5 billion) than in any full fiscal year through 2024, and its data center business alone cleared a $100 billion annualized run rate. A year ago, total quarterly revenue was $9.3 billion.

The market has noticed. Shares are up more than 8X over the past 12 months and jumped as much as 16% after this print. A company worth roughly $150 billion a year ago is now worth close to $1.3 trillion. It’s the kind of move investors usually associate with platform companies, not memory cycles.

🔒 Trying to break the cycle

The numbers are staggering. The strategy behind them is the actual story.

Memory has always been boom-bust: demand surges, the survivors overbuild, prices crater, everyone bleeds. The cycles culled the field to three main players: Micron, Samsung, and SK Hynix. Their shared discipline on supply is what keeps it tight. For 40 years, the industry’s curse was that every boom financed the next bust. This quarter, Micron tried to turn customer panic into a new floor.

The mechanism is the Strategic Customer Agreement (SCA). It’s a multi-year, take-or-pay contract locking in both volume and price, unlike the one-year handshakes the industry has always run on. Micron has now signed 16, typically five-year deals running through 2030. They already cover ~20% of its DRAM and a third of its NAND, with management targeting 40%+ of total revenue. Customers are backing them with ~$22 billion in deposits and guarantees, including roughly $18 billion of it in cash.

The largest deals set a ceiling at today’s prices, and a floor beneath them. Micron says that even at the floor, gross margins would sit well above its prior cyclical peak of ~60%. For the covered volumes, the old ceiling is effectively the new floor. The minimum value of these contracts is ~$100 billion, which management calls a conservative number, not a forecast.

The most telling example is Anthropic, with a memory and storage supply agreement paired with Micron’s investment in the AI lab’s Series H round. A memory maker buying equity in the very demand it’s trying to lock down.

The pricing power makes those floors believable. Micron’s Mobile and Client unit — the part most exposed to squeezed phone and PC makers — ran an 87% gross margin, up from 24% a year ago.

Why it matters

• Business model: Take-or-pay volumes and price floors smooth the industry’s most violent variable. The balance sheet already reflects the boom, and Micron now holds $24.4 billion in net cash. All three agencies upgraded Micron’s credit rating to BBB+.

• Competitive moat: The counterparties are the hyperscalers and NVIDIA — the same names that decide HBM allocation. Locking them in deepens an engagement already hard to displace.

• Investor angle: The bear case is always about what happens when the cycle turns. A floor margin above the old peak is the most direct rebuttal Micron has ever offered. The caveat is that upfront customer deposits are not free money. They come with future supply obligations.

For decades, Micron's curse was that good times invited the bust. By getting customers to commit volume, price, and cash years out, it's betting it can keep the boom without the hangover.

🌐 The squeeze everyone else feels

Micron’s 85% gross margin is someone else’s cost line.

The same shortage minting records in Boise is rippling through the rest of tech. Apple is raising prices, and HP’s memory bill doubled. Micron CEO Sanjay Mehrotra said on the call that there’s “no line of sight” to when supply catches up with demand. He pushed the crunch past 2027, a sharp reversal from earlier guidance that had it easing by next year.

It won’t ease quickly, because adding new supply is genuinely hard. Greenfield fabs take years and run into construction lead times, scarce skilled labor, permitting, and power. Meanwhile, each new node yields fewer incremental bits, and HBM’s heavier wafer appetite eats into everything else. Data-center memory is on track to top half the industry’s total bit demand for the first time this year. Everything non-AI is the residual.

But the edge isn’t only a victim of the crunch. Agentic platforms like OpenClaw and NVIDIA’s NemoClaw are pushing more AI workloads onto devices themselves, where better economics, privacy, and latency all matter. Micron expects that shift, paired with pent-up upgrade demand, to lift the memory inside each phone and PC over time. Today’s squeezed buyers could become the next source of memory demand.

The competition is sprinting after the same wave. SK Hynix, which leads the HBM market, just filed for a ~$29 billion US listing to fund its expansion, and Jensen Huang confirmed NVIDIA will source HBM4 for its next-gen Vera Rubin platform from all three memory makers. The allocation fight among Micron, Samsung, and SK Hynix is far from settled.

The real long-run wildcard sits in China. CXMT and Yangtze Memory are scaling fast, and standard DRAM is fungible enough that OEMs can qualify different suppliers across regions, product tiers, and configurations. Chinese suppliers still trail at the high end, especially in HBM, so this is not the 2026 story. But volume is how that gap closes.

🔭 What to watch next

• 🪜 The margin ceiling: Management itself flagged “a meaningful moderation in the rate of price increases” behind the ~86% gross margin guide for Q4 FY26. The step-up shrank to roughly +1.4pp from the ~+6pp it guided a quarter ago. Some read that as early price softening in 2H26 as SK Hynix adds supply and PC and phone volumes keep contracting. The counter is the SCA floor, which Micron says still clears its old peak.

• 🔒 How far above the floor: The contract terms are now largely public. Five years, take-or-pay, ~$100 billion minimum. The open questions are concentration (four large customers carry most of it) and how far actual revenue runs above that floor.

• 🌐 China and policy: Watch whether US export controls on China’s semiconductor equipment access tighten or loosen, and whether CXMT lands design wins at Western OEMs for Asia-bound products.

• 🏗️ The cost of staying ahead CapEx will jump to ~$10 billion in Q4 (~$27 billion for full-year FY26) and will step up again in FY27, with operating expenses set to rise ~$1 billion as R&D expands. The test will be whether Micron can fund the build without surrendering the balance-sheet discipline it just earned.

Bottom Line: Micron is trying to answer the oldest objection to memory stocks: the boom always invites the bust. This time, customers are locking in supply years in advance because AI has turned memory into a strategic bottleneck. For consumers, that means higher device prices. For Micron, it means margins even NVIDIA would envy.

That's it for today.

Happy investing!

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I own AAPL, AMD, GOOG, and NVDA in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members. 

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 📈 How to Invest in IPOs

• ⚽️ How FIFA Makes Money

• 🚀 How SpaceX Makes Money

• 🔎 Why Google is Selling $85B of Itself

Subscribe now

Three companies announced roughly $86 billion in acquisitions in a single week. None of these deals is really about buying revenue.

SpaceX, Salesforce, and Fox are buying control points (the layers customers touch every day) before the market structure hardens around them.

When technology shifts, those layers are the most valuable real estate there is:

• The workflow: where work gets done.

• The interface: where choices are made.

• The distribution layer: where attention is monetized.

• The data loop: that improves the product over time.

A scarce asset can justify a premium, but a rushed deal can torch billions. The whole difference comes down to one question: is the buyer acquiring a durable control point, or simply paying up for growth it could not build internally?

Are these companies buying the future, or buying time?

Let’s review.

Today at a glance:

• 🤖 SpaceX + Cursor

• ☁️ Salesforce + Fin

• 📺 Fox + Roku

🤖 SpaceX + Cursor: Buying the AI Workflow

SpaceX is acquiring Cursor parent Anysphere in a $60 billion all-stock deal, struck four days after its Nasdaq debut.

The headline number is enormous. Cursor was valued at $29 billion in November 2025, then explored funding above a $50 billion valuation in April 2026, before SpaceX exercised its right to buy the company outright at $60 billion.

But the structure matters: SpaceX is paying with stock, not cash.

Read more

Over 300,000 subscribers turn to us for business and investment insights

In case you missed it:

• 🚀 How SpaceX Makes Money

• 💰 Wall Street’s Top Stocks in Q1

• ❄️ Snowflake: AI Consumption Wins

• 🔎 Why Google is Selling $85B of Itself

Subscribe now

So, you want to invest in a hot IPO?

SpaceX just pulled off the largest IPO in history. The rocket maker was priced at a fixed $135 per share on June 12 and surged more than 50% in its first three days of trading, briefly touching an astounding $2.6 trillion valuation.

More are on the way. Anthropic and OpenAI filed confidentially earlier this month. The most valuable private companies on earth are racing toward public markets.

The temptation to invest is obvious.

These are the companies investors have been waiting years to own: rockets, AI models, frontier labs, category leaders. The kind of names that can make every other stock in your portfolio feel boring.

But there’s an old market joke that IPO stands for “It’s Probably Overpriced.”

The hotter the IPO, the more erratic the early trading can become.

Should you take the plunge?

Four decades of IPO history point to a boring answer: probably not yet.

Today at a glance:

• 📉 The IPO pop mirage

• 📊 What 40 years of data say

• 🧭 The five-rule playbook for IPOs

• 🚀 SpaceX, OpenAI, or Anthropic?

📉 The IPO pop mirage

The IPO pop you read about in the headlines is almost never yours.

The offering price and the opening price are two different numbers. In most hot IPOs, individual investors pay the second one.

That gap is often celebrated as a successful IPO. But it also means the company could likely have sold shares at a higher price. The first-day upside went to whoever received an allocation, not to the business or most public investors.

In traditional IPOs, banks often underprice the deal, allocated investors capture the first-day pop, and the company effectively eats the difference. Most individual investors buy in the open market after the pop.

That difference can be extremely short-lived. Rivian priced its IPO at $78, closed its first day at $101, and kept climbing. A week later, it hit $172, up 70%, making it briefly the third-most-valuable automaker on the planet, ahead of Ford and GM. In retrospect, it was the peak. The stock is down more than 80% from that first-day close, near $16 today.

Cerebras priced its AI listing at $185 and closed its first day at $311, up 68%. Spectacular for whoever got the allocation. Much less helpful for anyone buying after the move, when the valuation already reflected the excitement. A month later? The stock is nearly 30% off its peak.

A monster pop is not necessarily a triumph. Sometimes it is simply a temporary mispricing, and the buyer at the open is on the wrong side of it.

SpaceX tried to make the process fairer. It skipped the usual book-building theater and reserved an unusual 30% of the deal for individual investors.

The stock rose more than 50% in its first three days of trading anyway. The valuation now reflects even more future success than the IPO price did, leaving less room for error for new investors.

Takeaway: The first-day spike is usually a transfer of wealth to people who are not you. The smaller the pop, the fairer the deal. But a fairer price and a good investment are not the same thing.

📊 What 40 years of data say

IPOs tend to underperform. And by a lot.

Why? IPO investing has three built-in disadvantages:

• Limited operating history: The S-1 shows the past, not how the company behaves under public-market pressure.

• Insider selling: Lock-ups often expire after ~6 months, when employees and early investors can finally sell.

• Information edge: The best investors often saw the company years earlier, at much lower valuations, with better access.

Jay Ritter has tracked new listings for decades. The University of Florida finance professor built one of the definitive IPO datasets.

Across thousands of operating-company IPOs with at least $100 million in sales, measured from the first-day close, the five-year record is sobering:

• 50% posted negative returns over the five years after their first-day close.

• 30% lost half their value or more.

• Only 24% became multi-baggers, meaning they gained 100% or more.

This data excludes the smallest IPOs, those with less than $100 million in trailing sales. Including those would make the record look even worse, but also less relevant to the kind of large, category-defining companies most investors care about.

Finishing above the first-day close is not the same as beating the market. A stock that rises 20% over five years may look fine in isolation, but it still failed as an allocation if the S&P 500 did much better over the same period.

That’s why IPO researchers also look at market-adjusted returns. On that basis, the IPO record is even less forgiving. Larger IPOs do much better than tiny speculative listings, but on average, they still underperform comparable public companies after the first-day pop.

The damage clusters in a specific window: IPOs lag comparable companies most from 6 to 24 months after listing, around the time the lock-up expires and insiders are finally allowed to sell.

As the novelty fades and the selling pressure lands, expectations come back to earth, and so does the price.

If we expand to a 3-year window, several patterns show up again and again in Ritter’s 40-year data set:

• Size matters: IPOs with under $100M in sales trailed the market by ~34%, while those with over $500M in sales trailed by just ~4%.

• Profitability helps: Profitable issuers underperformed by ~13%, while unprofitable ones underperformed by ~31%. None of the megacaps heading for IPOs are profitable.

• VC backing boosts the odds: Venture-backed names trailed by ~14% — better than the ~25% for everyone else, but still behind.

Takeaway: A new listing is usually a stock priced for the long-term growth story, at the time of maximum execution risk. Time separates the winners from the wreckage. Buying a great business can be sound, but the IPO is often the worst moment to do it.

🧭 The Five-Rule playbook for IPOs

Warren Buffett has long been skeptical of IPOs. His view is simple: IPOs come to market when sellers choose the timing, not when buyers are likely to get the best deal. He once put it bluntly:

“It isn’t worth spending five seconds thinking about IPOs.”

Of course, an IPO is merely a moment in time, not an asset class. What looks uninvestable at the IPO price can eventually become a compelling idea.

That’s why having a clear process matters.

Here are a few rules I’ve created for myself:

1. Avoid the IPO hype: The first-day pop mostly belongs to investors who received the allocation, not to open-market buyers. Rushing in on Day 1 forces you to pay a hype-driven premium at the moment of maximum risk. The rule is simple: let the initial open-market volatility pass without touching it.

2. Wait for the second earnings call: The S-1 is a snapshot. Public markets are the test. The first earnings call shows how management talks to investors. The second shows whether they can forecast appropriately. Are the right KPIs improving? Is growth accelerating or decelerating? Did guidance hold? Did the lock-up create selling pressure? Two quarters help draw a trendline. Anything earlier is closer to buying blind.

3. Nibble in year one: The first year is usually the noisiest. The stock is absorbing new shareholders, lock-up expirations, analyst coverage, guidance resets, and the first real tests of public-market expectations. A small starter position lets you follow the company closely without making the IPO price your entire cost basis. If a 50% drawdown would damage your portfolio or your sleep, the position is too big.

4. Anchor to valuation: Missing a 20% move is not a disaster. Buying a great company at a price that already assumes perfection can be. The question is not whether the company is exceptional. It is whether the valuation leaves room for error. Don’t sweat daily movements; instead, focus on valuation relative to fundamentals.

5. Give it time: Many great stocks go nowhere in their first few years as public companies. That doesn’t mean the business is broken. It means the market is still figuring out the right multiple, the right expectations, and the right shareholder base. Multi-baggers rarely require buying on day one. They require surviving the early volatility.

Takeaway: The playbook is boring on purpose. Wait for public evidence, size small, and let the lock-up cycle pass. Boring is how you beat the IPO base rates.

🚀 SpaceX, OpenAI, or Anthropic?

Neither of the new-frontier IPOs has demonstrated how the business behaves under public-market pressure. Their margins at maturity are still uncertain, and their future growth is far from guaranteed.

None of this means the companies are bad. It means the setup is hard.

These may become some of the defining franchises of the next decade. The case for owning the leaders of AI and space is real. But the case for owning them on the first day, at peak-euphoria valuations, before a single public quarter, is much weaker.

The better move is boring: put them on your watch list.

Let them report. Let the first wave of excitement pass. Let insiders sell. Let expectations move from story to numbers. Then decide whether the business is getting stronger and whether the valuation leaves room for error.

Some of these stocks may compound for a decade. The playbook is not to ignore them entirely. It’s to invest on your own terms, at a price set with a clearer head.

Bottom Line: The 2026 IPO wave is a readout of how hot the market has become, not a buy signal. SpaceX, OpenAI, and Anthropic belong on the watch list, not the impulse-buy list. The best businesses give you years to get in. The IPO only sells you the urge to rush.

Once Anthropic and OpenAI’s S-1s are available, you can expect our breakdowns and signature visuals. Stay tuned for that!

That’s it for today!

Happy investing!

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I own ABNB, BABA, CPNG, DASH, META, SNOW, and UBER in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• ⚽️ How FIFA Makes Money

• 🚀 How SpaceX Makes Money

• 🔎 Why Google is Selling $85B of Itself

• 📈 Stock Buybacks & Dividends Explained

Subscribe now

A record IPO, then a shutdown

SpaceX just pulled off the largest IPO in history, raising about $75 billion at a valuation near $1.8 trillion. The stock has already surged more than 40% in its first two days, pushing its market cap past $2.5 trillion, nearly as much as Amazon.

But it was only the opening act.

Anthropic and OpenAI have already filed confidentially to go public, setting the stage for massive IPOs in the coming months. Both could list above $1 trillion.

Anthropic’s private valuation has rocketed past OpenAI’s to $965 billion, making the Claude maker, on paper, the most valuable standalone AI lab on earth.

But the company just got a reminder of how fragile that value can be. On Friday, Anthropic received a US export-control directive restricting access to its two most powerful models. Because it could not screen users in real time, the company disabled both models worldwide.

Every major AI company is choosing where to sit on the build-vs-rent spectrum:

• Build a frontier model like Anthropic, OpenAI, and Google.

• Rent one from a company that already has.

Until last week, that looked mostly like a trade-off between cost and control. Now there’s another variable: the risk of regulatory shutdown.

Today at a glance:

1. The frontier’s very bad week

2. Why building just got riskier

3. Apple’s accidental hedge

1. The frontier's very bad week

Read more

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🚀 How SpaceX Makes Money

• 🔎 Why Google is Selling $85B of Itself

Subscribe now

Premium members get:

• 📊 Monthly reports: 200+ companies visualized.

• 📩 Tuesday articles: Exclusive deep dives and insights.

• 📚 Access to our archive: Hundreds of business breakdowns.

PRO members get everything PLUS:

• 📩 Saturday PRO reports: Timely insights on the latest earnings.

Today at a glance:

1. ☁️ Oracle: Backlog vs. Balance Sheet

2. 🎨 Adobe: Freemium Bet

3. ✍️ DocuSign: IAM Crosses 13% of ARR

4. 🎿 Vail Resorts: Worst Season On Record

5. 🐶 Chewy: Stretched But Steady

1. ☁️ Oracle: Backlog vs. Balance Sheet

Oracle beat on revenue, beat on earnings, and grew its backlog by $85 billion during the quarter to $638 billion. The stock fell 12% anyway, its worst day since December.

The quarter wasn’t the problem. The incoming bill was. New CFO Hilary Maxson guided FY27 gross margins to “step down” and laid out ~$70 billion in net cash outlay, partly funded by a fresh ~$40 billion raise. Investors naturally scrutinize the cost of converting that backlog.

Q4 FY26 in numbers

• Revenue grew 21% Y/Y to $19.2 billion ($0.1 billion beat). Oracle Cloud Infrastructure (OCI) remains the main driver:

  • ☁️ Cloud +47% Y/Y to $9.9 billion (OCI +93% to $5.8 billion).

  • 🌐 Software -2% Y/Y to $6.8 billion.

  • 🖥️ Hardware +9% Y/Y to $0.9 billion.

  • 💼 Services +13% Y/Y to $1.5 billion.

• Margin trends: The operating margin remained flat at 32%, while gross margin compressed by 5pp to 65% as lower-margin IaaS scaled.

• Cash flow: For the full year FY26, operating cash flow was $32 billion (+54% Y/Y), but that’s not enough to fund the CapEx ramp. Free cash flow stayed deeply negative (-$24 billion) as CapEx hit $56 billion, above the $50 billion plan.

• Balance sheet: Oracle raised $43 billion in debt and $5 billion in equity in FY26 and expects another ~$40 billion in debt and equity financing in FY27. Oracle now has a net debt of $124 billion, including operating lease liabilities.

• FY27 guidance: Management reaffirmed FY27 revenue guidance of ~$90 billion, implying ~34% growth. Capital spending is expected to reach roughly $70 billion of Oracle’s own cash outlay. Oracle expects to raise about $40 billion through debt and equity, with any new debt likely pushed to calendar 2027.

So what to make of all this?

• 📈 The beat the market ignored: RPO jumped 363% Y/Y and 15% Q/Q to $638 billion, well above the ~$590 billion analysts expected. But the market has stopped paying for backlog alone and started pricing the cost of delivering it.

• ⚖️ Less funding pressure, still massive funding needs: Roughly $75 billion of Oracle’s large AI contracts is prepaid or customer-supplied hardware. That helps. But FY27 capital outlays are still enormous, and Oracle still needs external financing. The funding story improved at the margin. The absolute numbers got bigger.

• 📉 Margins are the new fault line: Oracle is building capacity before it fully bills. That creates a timing gap: CapEx now, revenue later. Management says infrastructure margins recover quickly once utilization ramps, but investors want proof.

• 🧱 The old Oracle is now shrinking: Software revenue fell 2% Y/Y. Oracle still needs that legacy cash engine to help fund the AI infrastructure buildout, and a contracting base makes the math harder.

Takeaway: Oracle delivered 1.2 gigawatts of incremental data center capacity in FY26 and expects Q1 FY27 delivery to approach 1 gigawatt. The bull case is becoming more concrete, but so are the capital intensity, margin pressure, and financing needs. Oracle is now a bet on execution.

2. 🎨 Adobe: Freemium Bet

Read more

Over 300,000 subscribers turn to us for business and investment insights

In case you missed it:

• 🚀 How SpaceX Makes Money

• 💰 Wall Street’s Top Stocks in Q1

• ❄️ Snowflake: AI Consumption Wins

• 🔎 Why Google is Selling $85B of Itself

Subscribe now

⚽️ The biggest World Cup ever is here!

Hosted by Canada, Mexico, and the US, the competition kicked off Thursday. It’s the first one expanded to 48 teams. FIFA president Gianni Infantino calls it "104 Super Bowls in a month" (it's actually 39 days, but who's counting).

This year's teams will share a record $871 million in prize money, while some fans pay as much as $33,000 for a single seat.

Across this World Cup cycle, FIFA expects to take in roughly $13 billion. As a non-profit, most of that money is supposed to flow back into football: staging tournaments, funding member federations, and running the global game.

Today at a glance:

• 💰 The Four-Year Heartbeat

• 🌍 Where The Money Goes

• 🎟️ Pricing the People's Game

💰 The Four-Year Heartbeat

Because FIFA’s business revolves around the World Cup, held once every four years, the organization uses a four-year budget cycle.

The visual below shows the latest budget for the 2023-2026 cycle.

FIFA expects to earn $13 billion in this cycle, up 72% from $7.6 billion for the 2019-2022 Qatar cycle. That’s a big upside from the $11 billion it had previously budgeted.

The jump came from three places:

• The field expanded to 48 teams (more inventory to sell).

• The tournament landed in the US (the richest media market on earth).

• FIFA bolted on a second large event in the US, the revamped Club World Cup.

Here’s where the revenue comes from:

• 📺 TV Broadcasting ($5.3 billion, ~40%): FIFA’s largest stream by far. It sells the right to air the matches to networks region by region, and the bidding runs hot because few events deliver a global audience watching live, all at the same time.

• 🎟️ Hospitality & Ticketing ($3.6 billion, ~28%): Match tickets plus premium corporate packages, run through a FIFA-owned subsidiary so every dollar flows straight back to headquarters. 2026 is the first World Cup with dynamic, demand-based pricing, which sent the top end soaring (more on this in a minute).

• 🤝 Marketing & Sponsorship ($3.3 billion, ~25%): The deals brands like Coca-Cola, Visa, and Adidas pay to attach their name to the tournament, sold in tiers from global partners down to regional supporters. Sponsors spend several times more on advertising, though that money goes to agencies and broadcasters, not FIFA.

• 🏷️ Licensing ($0.4 billion, ~3%): FIFA-branded merchandise, video games, and royalties on the FIFA name. Tiny next to the rest, and almost a rounding error.

FIFA runs a deficit three years running, then the World Cup arrives and pays for everything — nearly $9 billion in a single year. The Club World Cup debut in 2025 also created a revenue boost. The lumpiness is the design, with a budget built with that timeframe in mind.

Takeaway: For a non-profit, near-breakeven is the whole point. The cycle is engineered to look like FIFA reinvests almost everything. Which raises the real question: where does the money actually go?

2. 🌍 Where the Money Goes

Almost none of that $13 billion stays with FIFA, and that is the entire point.

The 2026 World Cup is unusually asset-light for FIFA. Every match will be played in an existing stadium, many of them NFL venues. Unlike Qatar’s reported $200 billion buildout, the tournament does not require a countrywide stadium construction boom. More of the revenue can flow back into football.

Here’s where it goes:

• 🏟️ Competitions & Events ($7.6 billion, ~58%): The largest line by far. This is the cost of staging the tournaments themselves: running the World Cup and FIFA’s other events, including the prize money paid to teams.

• 🌱 Development & Education ($3.9 billion, ~30%): The money FIFA puts back into the game. Through its FIFA Forward programme ($2.25 billion of it), FIFA funds all 211 member federations: pitches, youth programs, and grants in places the commercial market overlooks.

• 🏛️ Governance & Admin ($0.9 billion, ~7%): FIFA’s own running costs. Salaries, governance bodies, and the day-to-day expenses of operating the organization itself. That includes the compensation package of roughly $6 million annually for FIFA President Gianni Infantino.

The World Cup Prize money pool is $871 million for 2026, including $655 million allocated to performance, a 49% increase over Qatar, with the champion taking home $50 million and even the first team eliminated leaving with over $10 million. The expansion of the competition from 32 to 48 teams was the main reason behind the increase.

Every member federation that receives a check also has a vote for the FIFA president. Infantino won in 2016 partly by promising to more than double those grants. He runs for a fourth term in 2027. Critics see a patronage machine. FIFA calls it growing the game.

In 2015, US prosecutors charged dozens of officials and sports-marketing executives over more than $200 million in bribes, almost all of it paid for the media and marketing rights to FIFA's tournaments. That history matters because media and marketing rights are still FIFA’s two largest revenue engines. The same machine that funds football also asks the world to take its redistribution on trust.

The books are built to break even on paper. But when revenue routinely beats budget, the surplus compounds into a reserve now approaching $2.7 billion.

Takeaway: FIFA runs a redistribution machine mixing sports and politics. The same dollars that grow the game secure the votes that keep the machine and its president in place. No rival governing body has built a flywheel this durable.

3. 🎟️ Pricing the People's Game

A real-time auction for seats

For the first time, the tournament uses dynamic pricing, ticket prices that float with demand, like airline fares. The contrast with the last US-hosted World Cup is brutal.

In 1994, a World Cup ticket ran $25 to $475. In 2026, the floor is $60, roughly in line with inflation. But the ceiling has detonated: $6,730 for the final at list, with premium seats that opened there climbing past $32,000 by May.

The cheap seats held. The expensive ones went up as much as 70×.

Some estimates put the average 2026 World Cup ticket around $1,300. That implies an increase of about 1,000% after inflation compred to 1994. Real median household income over the same span rose just 32%.

Football is the most popular sport on earth, played and followed by people of every income level. The World Cup carries public and national meaning, and pushing prices out of reach can erode its cultural value. The asset that makes the rights so valuable is football’s universality — and pricing out ordinary fans threatens exactly that.

FIFA pushes this hard for a reason. This is likely the last men’s World Cup on US soil for decades. FIFA’s own budget projects that hospitality and ticketing will decline by $938 million in the 2027-2030 cycle, even as total revenue climbs. The 2026 gate is a peak FIFA doesn’t expect to see again.

The backlash arrived on cue. New York and New Jersey attorneys general have subpoenaed FIFA over its ticket practices. Meanwhile, host cities like Boston and Kansas City are covering tens of millions in security and transport while FIFA keeps the upside at the gate.

Cities take the bet for a reason. Each US host is promised between $160 million and $620 million in local spending from visiting fans, plus local jobs and a month in the global spotlight. A summer of full hotels and busy restaurants is real money.

But the payoff may not arrive as advertised. Victor Matheson, a sports economist at Holy Cross, calls FIFA’s projections closer to a press release than serious analysis. Much of the spending that does show up may have simply shifted from somewhere else. For an economy of America’s size, the whole tournament barely registers, well under 0.1% of GDP.

The controversy goes beyond prices and public costs. Human-rights groups have also warned about risks around immigration enforcement, policing, freedom of expression, and the treatment of local communities during the tournament. FIFA markets the World Cup as a global celebration, but the 2026 edition arrives with a heavier political backdrop than usual.

Beyond the men’s World Cup

A major growth driver for FIFA has been the 2025 Club World Cup. FIFA expanded it to 32 teams and staged it in the US, creating a second marquee event to smooth the four-year sawtooth.

The $1 billion Women’s World Cup revenue target in Brazil in 2027 points in the same direction: more global events, more premium pricing, and more predictable revenue between men’s World Cups.

Takeaway: The 2023-2026 cycle is where FIFA became less dependent on a single event and more like a portfolio of scarce global sports rights. But maximizing revenue comes with controversy. Teams collect record checks. Fans are asked to pay more. FIFA may keep finding new ways to monetize football, but its own brand could be the one paying the price.

That’s it for today!

Happy investing!

Want to sponsor this newsletter? Get in touch here.

Thanks to Fiscal.ai for being our official data partner. Create your own charts and pull key metrics from 50,000+ companies directly on Fiscal.ai. Save 15% with this link.

Disclosure: I do not own any stock discussed in this article in App Economy Portfolio. I share my ratings (BUY, SELL, or HOLD) with App Economy Portfolio members.

Author's Note (Bertrand here 👋🏼): The views and opinions expressed in this newsletter are solely my own and should not be considered financial advice or any other organization's views.

Over 300,000 subscribers turn to us for business and investment insights.

In case you missed it:

• 🤖 NVIDIA: Gone Parabolic

• 🚀 How SpaceX Makes Money

• 🔎 Google: The Anthropic Paradox

• 📈 Stock Buybacks & Dividends Explained

Subscribe now

Alphabet sold $85 billion of stock in two days

That’s more than what the largest IPO in history will raise.

SpaceX is about to price the largest initial public offering ever at $75 billion, valuing the company at ~$1.8 trillion. But Alphabet, the parent company of Google, quietly topped that last week by selling new shares — upsizing a planned $80 billion raise to $85 billion after demand poured in. Berkshire Hathaway anchored it with a $10 billion investment, buying shares at a 6–8% discount.

Here’s the paradox. Alphabet sits on well over $120 billion in cash and short-term investments on its balance sheet, just generated $165 billion in operating cash flow last year, and can borrow at rates few companies on earth can match. Firms in that position almost never sell equity. Instead, they use their cash flow to buy back stock. Diluting your own shareholders is the most expensive money there is, a slice of every future dollar of profit, forever.

So why do it? Because the AI buildout has grown so large it requires every available tool. Alphabet expects to spend $190 billion in 2026 alone, and that figure could increase by year-end.

Every company racing to build compute now stands in front of three doors:

• Pay with the cash you already generate.

• Borrow it (issue debt).

• Sell a piece of the company (issue equity).

Each door sends a different signal about how confident you are and how much risk you are willing to take (more on this in a minute).

Today, we are looking at the three ways Big Tech is currently funding the trillion-dollar AI buildout. We’ll review which ones are good or bad for investors and what it says about where we are in this market cycle.

1. 💵 Self-funding

2. 🏦 Debt issuance

3. 📉 Selling more stock

1. 💵 Self-funding

The cleanest way to fund a data center is to pay for it with money the business already makes. No interest, no dilution, no banker fees. For most of this cycle, that’s exactly what Big Tech did.

If you need a reminder, free cash flow (FCF) is the cash generated from the business, less capital expenditures (funds to buy, upgrade, or maintain physical assets). As long as FCF is positive, you are effectively self-funding your business ventures.

Free Cash Flow = [Cash from Operations – Capital Expenditures]

Amazon is the textbook case for how far that model stretches. The company is on track for roughly $200 billion in capital expenditures in 2026, up from a trailing-twelve-month figure of $147 billion (+67% Y/Y).

Meanwhile, trailing-12-month operating cash flow grew to $149 billion (+30% Y/Y) at the end of March. It’s a staggering number that, a few years ago, would have funded the whole thing with room to spare. It no longer does.

Amazon is now spending essentially every dollar it generates. It was already the case in 2021 and 2022, when the company invested in fulfillment to meet pandemic-driven demand. As a result, FCF turned slightly negative in 2021 and 2022. FCF recovered to a record $38 billion in 2024 before the AI buildout cut it to $11 billion in 2025

At the end of Q1 2026, trailing FCF worsened to just ~$1 billion.

“The faster AWS grows, the more short-term CapEx we’ll spend.”

That was CEO Andy Jassy on the Q1 call, and it captures the squeeze. AWS reaccelerated to +28% Y/Y, its fastest in nearly four years, but Amazon has to lay out cash for land, power, chips, and servers 6 to 24 months before it can bill a customer. Growth and free cash flow are pulling in opposite directions.

Self-funding is a position of strength: you don’t need anyone’s permission to build. The catch is that it caps your spend at what the business produces today. The AI bill is now testing that ceiling everywhere. Microsoft, Meta, and Alphabet still fund most of their CapEx from existing operations. The open question is what happens when one great business can no longer feed an even hungrier one.

Takeaway: Free cash flow is the cheapest capital and the proudest signal — it says the core business can pay its own way. But when CapEx consumes more than 100% of operating cash flow, the wallet is empty, and the next dollar has to come from somewhere else.

2. 🏦 Debt issuance

Read more