The cleanest way to see what the market punished is to put Alphabet next to Amazon, which reported a day later.
Alphabet spent $51.4 billion on capex in the quarter against $30.7 billion of Cloud revenue: about $1.67 of spending for every dollar of cloud sales. Amazon spent roughly $77 billion on property and equipment against $46 billion of AWS revenue: also about $1.67. Amazon's figure includes warehouses, so its AI ratio is somewhat lower, but the order is the same. Both reported negative free cash flow. Alphabet fell 12.6% on Thursday. Amazon closed Friday down 1.4%.
So the market didn't punish the spending. It punished the paradox in your title. Queries grew at the fastest pace in five years, while Search revenue growth slowed from 17% to 12%. Currency was the biggest drag, but management also named AI Mode's lower monetization. More queries with slower revenue means each search earns less. Amazon's AI doesn't eat its retail margin. Google's AI competes with its own ad slot.
The number I'd watch is the gap between query growth and Search revenue growth. If AI Mode ads catch up, the gap closes and Thursday looks like an overreaction. If it widens while capex climbs again in 2027, Cloud has to pay for both the build-out and the hole in Search, and $11.1 billion of quarterly Cloud operating income isn't enough for that yet.
On the investment portfolio distorting earnings, the Q3 10-Q has the scale. Alphabet held $128.2B of private-company stakes at September 30, "primarily" one private company, with $116.3B of unrealized gains on them over nine months but only $3.0B in Q3, so most of this year's paper gain sits in the first half and the quarter's $8.7B equity loss came mostly from SpaceX. The 10-Q also lists $20.0B of further funding commitments to a private company, tied to milestones through 2030. Anthropic's S-1 lines up with those figures, compared filing by filing on Equibles: https://equibles.com/research/anthropics-s-1-checked-against-amazon-google-microsoft-and-broadcom-filings
Reddit's Q3 letter shows the other side of this paradox. Its search traffic fell by more than a million weekly visitors in October, and the CEO pinned it on AI Overviews. When people stop at the AI answer, Reddit loses the visit and Google loses a slot to sell ads against, so the same shift hurts both, just at different speeds. I'd watch which bends first: Google's paid clicks or Reddit's daily users.
Query growth at a five year high while Search revenue slows from 17% to 12% is the number I'll be watching. Cloud doubling is great, but $51.4 billion of capex in one quarter with free cash flow at minus $9.1 billion means the cloud side has to keep carrying it for a while.
On the insider side, nobody at Alphabet has bought the post-earnings drop so far: the only Form 4s since Oct 25 are routine GSU vesting with tax withholding at $317.28 (Pichai and other officers) and director Frances Arnold's 82-share sale at $284 on Oct 29 under a 10b5-1 plan. With CapEx at a record and FCF negative, a discretionary open-market buy would stand out if one shows up.
Interesting that queries are growing at the fastest pace in five years while revenue growth is slowing. I wonder how much of that gap is AI Mode simply monetizing at lower rates today versus ad formats not yet being adapted to it. The $8.7B loss on equity investments also shows how much noise the SpaceX mark can add to reported EPS. Do you think investors will start looking at operating income and Cloud margins instead of GAAP EPS when judging Alphabet?
The cleanest way to see what the market punished is to put Alphabet next to Amazon, which reported a day later.
Alphabet spent $51.4 billion on capex in the quarter against $30.7 billion of Cloud revenue: about $1.67 of spending for every dollar of cloud sales. Amazon spent roughly $77 billion on property and equipment against $46 billion of AWS revenue: also about $1.67. Amazon's figure includes warehouses, so its AI ratio is somewhat lower, but the order is the same. Both reported negative free cash flow. Alphabet fell 12.6% on Thursday. Amazon closed Friday down 1.4%.
So the market didn't punish the spending. It punished the paradox in your title. Queries grew at the fastest pace in five years, while Search revenue growth slowed from 17% to 12%. Currency was the biggest drag, but management also named AI Mode's lower monetization. More queries with slower revenue means each search earns less. Amazon's AI doesn't eat its retail margin. Google's AI competes with its own ad slot.
The number I'd watch is the gap between query growth and Search revenue growth. If AI Mode ads catch up, the gap closes and Thursday looks like an overreaction. If it widens while capex climbs again in 2027, Cloud has to pay for both the build-out and the hole in Search, and $11.1 billion of quarterly Cloud operating income isn't enough for that yet.
On the investment portfolio distorting earnings, the Q3 10-Q has the scale. Alphabet held $128.2B of private-company stakes at September 30, "primarily" one private company, with $116.3B of unrealized gains on them over nine months but only $3.0B in Q3, so most of this year's paper gain sits in the first half and the quarter's $8.7B equity loss came mostly from SpaceX. The 10-Q also lists $20.0B of further funding commitments to a private company, tied to milestones through 2030. Anthropic's S-1 lines up with those figures, compared filing by filing on Equibles: https://equibles.com/research/anthropics-s-1-checked-against-amazon-google-microsoft-and-broadcom-filings
Reddit's Q3 letter shows the other side of this paradox. Its search traffic fell by more than a million weekly visitors in October, and the CEO pinned it on AI Overviews. When people stop at the AI answer, Reddit loses the visit and Google loses a slot to sell ads against, so the same shift hurts both, just at different speeds. I'd watch which bends first: Google's paid clicks or Reddit's daily users.
Query growth at a five year high while Search revenue slows from 17% to 12% is the number I'll be watching. Cloud doubling is great, but $51.4 billion of capex in one quarter with free cash flow at minus $9.1 billion means the cloud side has to keep carrying it for a while.
On the insider side, nobody at Alphabet has bought the post-earnings drop so far: the only Form 4s since Oct 25 are routine GSU vesting with tax withholding at $317.28 (Pichai and other officers) and director Frances Arnold's 82-share sale at $284 on Oct 29 under a 10b5-1 plan. With CapEx at a record and FCF negative, a discretionary open-market buy would stand out if one shows up.
Interesting that queries are growing at the fastest pace in five years while revenue growth is slowing. I wonder how much of that gap is AI Mode simply monetizing at lower rates today versus ad formats not yet being adapted to it. The $8.7B loss on equity investments also shows how much noise the SpaceX mark can add to reported EPS. Do you think investors will start looking at operating income and Cloud margins instead of GAAP EPS when judging Alphabet?