AMD Q2 2026 Earnings Analysis
AMD delivered record Q2 revenue of $11.5B (+50% YoY) with gross margin expansion to 56%, driven by data center strength, and guided Q3 revenue to $13B ±$300M.
Key Metrics
要点总结
- Record Q2 revenue $11.5B (+50% YoY) with data center segment doubling YoY to $6.7B now 58% of total revenue.
- Gross margin expanded 200 basis points to 56% driven by favorable product mix and data center scale benefits.
- Q3 guidance $13B ±$300M revenue; expects data center AI to more than double and server CPU 70%+ growth in 2027.
收听平台
可用语言
Transcript
// Full episode scriptBETA FINCH — AMD Q2 2026 EARNINGS BREAKDOWN
Welcome to Beta Finch, your AI-powered earnings breakdown. I'm Alex, joined as always by Jordan, and today we're digging into Advanced Micro Devices — AMD — and their second quarter 2026 results. Jordan, this one's a big one.
It really is. Record revenue, and the guidance for next year is honestly kind of stunning. Let's get into it.
Before we do — quick disclaimer. This podcast is AI-generated content for educational and entertainment purposes only. Nothing we discuss should be considered investment advice. Always do your own research and consult a qualified financial advisor before making any investment decisions.
Good to have that up front. Okay, let's talk numbers.
AMD posted revenue of $11.5 billion for Q2, up 50% year-over-year and 13% sequentially. That's their sixth straight quarter of greater-than-30% growth, which is a pretty remarkable streak.
And the story underneath that number is really about data center. It's now 58% of total revenue, up from 42% a year ago. Data center segment revenue was $6.7 billion, more than doubling year-over-year. Split that in two: server CPUs — that's EPYC — grew over 70% in both cloud and enterprise, fifth consecutive quarter of record server CPU revenue. And Instinct, their AI accelerator line, more than doubled.
Profitability kept pace too. Gross margin hit 56%, up over 200 basis points year-over-year. Operating income was $3.1 billion, a 27% operating margin. And non-GAAP diluted EPS grew about 82% year-over-year — meaningfully outpacing revenue growth.
That's the operating leverage story playing out in real time.
Client and gaming was more mixed. Client — that's PCs — grew 23% to $3.1 billion on record mobile processor sales and strong commercial Ryzen PRO adoption. But gaming dropped 31% to $779 million, mostly semi-custom console revenue fading late in that console cycle, plus some pricing pressure on graphics cards from component costs.
Embedded was a bright spot too — up 19% to $977 million, their strongest growth in three years, with networking and defense customers leading the way.
Now let's talk strategy, because this is where CEO Lisa Su really leaned in. The headline announcement was Helios — their new rack-scale AI platform combining EPYC Venice CPUs, MI450 GPUs, networking, and their ROCm software stack.
And the customer list here is the story. They already had multi-generation, gigawatt-scale commitments with OpenAI and Meta. Now add a brand-new strategic partnership with Anthropic — up to 2 gigawatts of MI450 GPUs, with the first gigawatt deploying in early 2027. There's also a joint engineering piece where Anthropic's Claude models help optimize workloads for AMD's chips.
And Microsoft is expanding its footprint too, deploying Helios at scale on Azure for frontier model inferencing.
What stood out to me is how much AMD raised its own long-term targets. They now expect the AI accelerator market to hit roughly $1.4 trillion by 2030, and the server CPU market around $220 billion. Combined, they're calling the total high-performance and AI computing market close to $2 trillion by 2030.
And Lisa Su said flatly that AMD is tracking materially ahead of the long-term model they laid out at last November's Analyst Day — revenue growth well above their prior 35% target, and EPS expected to "significantly exceed" their $20 target.
That's a big statement to make less than a year after setting those targets.
Let's get into guidance. For Q3, AMD is guiding to about $13 billion in revenue, plus or minus $300 million — that's 41% year-over-year growth at the midpoint. Gross margin guided at approximately 56%.
And looking further out, this is where it gets really interesting. They now expect server revenue to grow more than 80% year-over-year in the second half of 2026, and the full data center segment — CPUs plus AI accelerators combined — to more than double in 2027.
There was a great back-and-forth in the Q&A about that "more than double" phrasing. Analyst Stacy Rasgon basically asked, how much work is the word "more" doing in that sentence? And CFO Jean Hu and Lisa Su both hinted the AI accelerator piece specifically could be "well over 100%" growth — stronger than the headline number suggests.
There was also a useful nugget on the supply side. CEO Su said the server CPU supply chain has been tight all year because demand outpaced forecasts, but with 2027 demand now better understood, she expects supply to actually improve — and potentially allow growth to come in even higher than the 70%-plus target.
One more thread worth flagging — the questions about gross margin trade-offs as AI accelerators ramp. Jean Hu was pretty clear that AI accelerator margins run slightly below corporate average, but the growing server CPU mix and a recovering embedded business should help offset that. Her framing was that it's a "puts and takes" balance they feel confident managing.
So what does this all mean for someone watching the stock? A few threads to pull on. First, the diversification story — this isn't just an AI GPU bet anymore. Server CPUs, embedded, and even PCs are all contributing meaningfully to growth, which reduces reliance on any single product ramp.
Second, the customer concentration in AI — OpenAI, Meta, Anthropic, and Microsoft are the names doing the heavy lifting on Instinct and Helios demand. That's a strength in terms of visibility, but it's also worth watching how broad that customer base gets over time, since Su did mention more customers are expected to contribute through 2027.
And third, component and memory costs are a real watch item — both for the softening PC market in the second half and for HBM costs feeding into GPU margins. Management sounded confident they've got supply locked in, but it's a variable worth tracking each quarter.
Before we wrap, Jordan, the closing disclaimer.
Right — everything discussed is AI-generated analysis for educational purposes. Past performance doesn't guarantee future results. Please do your own due diligence.
AMD heads into the second half of 2026 with a lot of momentum — record numbers, a raised long-term outlook, and some serious AI partnerships now locked in.
The real test comes over the next few quarters as Helios actually ramps into production and we see whether that "well over 100%" language turns into real numbers. We'll be watching closely.
That's it for this episode of Beta Finch. Thanks for listening, and we'll catch you next time.
See you then.