↗️ AMD Acquires ZT Systems
↗️ AMD Acquires ZT Systems
A $5 billion purchase to take on NVIDIA
Welcome to the Free edition of How They Make Money.
Over 140,000 subscribers turn to us for business and investment insights.
In case you missed it:
Advanced Micro Devices (AMD) announced it would acquire hyperscale data center solutions provider ZT Systems (private) for nearly $5 billion.
With this move, AMD wants to take on NVIDIA in the race for AI infrastructure with advanced chips and the ecosystem around them.
Like NVIDIA, AMD is focused on high-performance computing (HCP) and AI to capture a share of the surging data center accelerator market.
AMD has been on an acquiring spree of late:
2022: Xilinx ($49 billion) and Pensando ($2 billion).
2023: Mipsology and Nod.ai.
2024: Silo AI ($0.7 billion) and ZT ($5 billion).
A critical component of AMD’s recent strategy is to go beyond advanced chips and offer software (for developers to access the capabilities through applications) and now tailored system solutions (to optimize the data center for performance).
Let’s put it all in context in a few minutes.
Today at a glance:
AMD Q2 FY24.
The ZT Systems acquisition.
What does it mean for the industry?
Check out our deep dive for a primer on the semiconductor industry and its jargon.
1. AMD Q2 FY24
Before diving into the ZT acquisition, let’s look at AMD’s latest quarter.
AMD has four main reportable segments:
⚙️ Data Center: Primarily server microprocessors (CPUs) and graphics processing units (GPUs), data processing units (DPUs), Field Programmable Gate Arrays (FPGAs), and Adaptive System-on-Chip (SoC) products for data centers. These components are essential for managing and processing vast amounts of data used by businesses and cloud services.
💻 Client: Primarily CPUs, accelerated processing units (APUs) that integrate CPUs and GPUs, and chipsets for desktop and notebook personal computers. The products deliver high-performance computing and graphics capabilities for everyday users.
🎮 Gaming: Primarily discrete GPUs, semi-custom SoC products, and development services. These are used in gaming consoles, PCs, and cloud gaming platforms to render advanced graphics and support high-speed gameplay.
🔧 Embedded: Primarily embedded CPUs and GPUs, APUs, FPGAs, and Adaptive SoC products. These are used in various devices and applications, from digital signage and medical imaging to industrial controls and automotive systems.
Income statement:
Revenue grew 9% year-over-year to $5.8 billion ($120 million beat).
⚙️ Data Center grew 115% to $2.8 billion. It has experienced remarkable growth, becoming the primary revenue driver, fueled by the success of EPYC server CPUs and the recent surge in Instinct GPU sales for AI applications.
💻 Client grew 49% to $1.5 billion. Like the rest of the industry, this segment faced a post-pandemic hangover in late 2022 and early 2023.
🎮 Gaming declined by 59% to $0.6 billion. It follows the console cycle, with PS5 and XBOX Series sales in a downtrend.
🔧 Embedded (including Xilinx) fell 41% to $0.9 billion.
Gross margin was 49% (+3pp Y/Y, +2pp Q/Q).
Operating margin was 5% (+5pp Y/Y, +4pp Q/Q).
Non-GAAP operating margin was 22% (+2pp Y/Y, +1pp Q/Q).
Non-GAAP EPS was $0.69 ($0.01 beat).
Cash flow:
Operating cash flow grew 56% to $0.6 billion.
Balance sheet:
Cash and cash equivalent: $5.3 billion.
Debt: $2.2 billion.
Guidance:
Q3 FY24 revenue up 15% sequentially $6.7 billion ($0.1 billion beat).
Expanding gross margin in the second half.
So what to make of all this?
Strong Q2 results: Revenue beat expectations, driven by record data center sales and AI adoption.
Data Center booming: Revenue more than doubled year-over-year and surged 21% sequentially to $2.8 billion, fueled by the Instinct MI300 GPU (AMD's powerful accelerator for large-scale AI and HPC workloads) and strong EPYC CPU sales. Cloud and enterprise adoption is expanding. Management expects EPYC market share gains with the upcoming Turin launch (next-gen family).
AI at the forefront: The MI300 achieved over $1 billion in quarterly revenue, with major cloud providers embracing the platform. AMD's expanding AI software ecosystem and strategic acquisitions have started showing evidence of a stronger position in the AI space. To be sure, NVIDIA is in another ballpark with nearly $23 billion in Data Center revenue during the quarter ending in April. We’ll learn more about NVIDIA’s July quarter next week.
Client segment strength: Revenue was up 49% year-over-year, driven by Ryzen sales and initial Zen 5 shipments. Ryzen AI notebooks received positive reviews. AI PCs could be a growth catalyst despite intense competition.
Gaming & Embedded challenges: Gaming revenue declined due to the gaming console cycle and inventory normalization, but Embedded showed early recovery signs, slightly growing sequentially.
Positive Outlook: Management expects a strong second half with rapid revenue growth and gross margin expansion, led by the Data Center and Client segments. They are ramping up MI300 with a tight supply chain through 2025. Data center GPU revenue is projected to exceed $4.5 billion in 2024. The company is also increasing investments in AI to capitalize on the expanding market opportunities.
2. AMD acquires ZT Systems
What is ZT Systems?
Revenue and clientele: ZT Systems generates an estimated $10 billion annually, primarily serving large hyperscalers like Amazon Web Services (AWS) and Azure.
Company size and history: Founded in 1994, ZT Systems is a private company headquartered in Secaucus, New Jersey, with approximately 1,000 employees in design and customer support and another 1,000 in manufacturing.
Evolution and focus: ZT has evolved from desktop PCs and pedestal servers to focus on data center servers, rack-scale design and integration, and ultimately hyperscale solutions, shipping "hundreds of thousands of servers annually."
Deal Overview
Value: $4.9 billion, comprising 75% cash and 25% stock, including a $1 billion investment AMD made in ZT over the past year.
Strategic focus: The acquisition primarily targets ZT's rack-scale systems design capabilities, which AMD plans to integrate into its Data Center Solutions Business Group.
Manufacturing divestment: AMD plans to spin out and sell ZT's manufacturing assets, retaining the higher-margin system design expertise.
Leadership structure: ZT's CEO, Frank Zhang, will lead the manufacturing business (to be sold), while President Doug Huang will oversee design and customer enablement, reporting to Forrest Norrod, head of AMD's Data Center Solutions Business Group.
Deal closure and financial impact: The deal is expected to close in the first half of 2025, subject to regulatory approvals, and is projected to be accretive to AMD's non-GAAP earnings by the end of 2025.
AMD's Strategic Rationale
Accelerated AI growth: AMD aims to leverage ZT's expertise to rapidly deploy leading training and inferencing solutions at scale, capturing a larger share of the burgeoning AI market.
Enhanced customer experience: ZT's deep understanding of hyperscale customer needs will enable AMD to provide tailored solutions and superior support, fostering stronger relationships with its largest customers.
Complementary capabilities: ZT's system design expertise complements AMD's strengths in chips and software, creating a more comprehensive AI infrastructure offering.
Cultural alignment: The existing collaboration between AMD and ZT and their shared customer base suggests a smooth integration process.
Here’s AMD CEO Dr. Lisa Su talking about the deal on CNBC.
3. What does it mean for the industry?
AMD's acquisition of ZT Systems is a calculated move to expand its footprint in the booming AI infrastructure market. ZT brings valuable expertise in rack-level solutions and deep relationships with major cloud providers.
This acquisition empowers AMD to:
Challenge NVIDIA's dominance: By offering end-to-end AI solutions encompassing chips, software, and systems, AMD can more directly compete with NVIDIA's vertically integrated approach.
Focus on full-stack AI solutions: AMD aims to provide a comprehensive suite of AI offerings, accelerating the deployment of AI infrastructure at scale and catering to the specific needs of hyperscalers and enterprises.
Deliver customer benefits: The acquisition is expected to translate into faster deployment times, optimized performance, and potential cost savings for customers as AMD streamlines the process of building and scaling AI infrastructure.
Key Takeaways:
ZT Systems adds crucial system design expertise and customer relationships to AMD's arsenal.
AMD focuses on providing tailored, bespoke AI solutions to meet the unique needs of hyperscalers and enterprises.
The acquisition could shift the competitive landscape, potentially impacting NVIDIA's market share as ZT Systems designed NVIDIA-based solutions for major cloud providers (until now).
AMD's acquisition of ZT Systems represents a bold step towards establishing itself as a major player in the AI infrastructure market. By combining its chip and software capabilities with ZT's system design expertise, AMD could accelerate the deployment of AI solutions and ride the big AI wave behind NVIDIA.
Seizing the AI Opportunity
Based on AMD's estimates, the total adressable market (TAM) for large data center AI accelerator is projected to skyrocket from $45 billion in 2023 to $400 billion by 2027.
Based on the latest round of earnings we’ve discussed, Amazon, Microsoft, and Google plan to ramp up significantly their AI Capex for the foreseeable future. AMD can capitalize on this opportunity by bolstering its AI capabilities and offering bespoke integrated solutions.
In the AI gold rush, not all picks and shovels have to come from the same player.
Stay tuned for our breakdown of NVIDIA’s earnings next Friday!
That’s it for today.
Stay healthy and invest on!
Disclosure: I am long AMD, AMZN, ASML, GOOG, META, NVDA and TSM in the 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.