AMD Pledges Annual Rack-Scale AI Launches Through 2030, Lisa Su Confirms
AMD will launch a brand-new rack-scale AI system every single year from here on, CEO Lisa Su told investors on the company’s Q2 2026 earnings call on August 4. Shacknews reported her exact words: “Looking beyond Helios, we plan to launch a new rack-scale AI platform every year,” said AMD CEO Lisa Su on the company’s latest quarterly earnings call. She added that “each generation” would be delivering “significant performance, efficiency, and TCO gains.”
The commitment turns AMD’s Helios rack — the company’s first fully integrated GPU-CPU-networking system — into the opening chapter of a much longer roadmap, rather than a one-off flagship. It also lands the same week AMD posted its strongest quarter yet, giving Su’s pledge real financial weight behind it.
Record $11.5 Billion Quarter Backs Up the AI Bet
AMD’s official Q2 2026 results, published by AMD Investor Relations, show revenue increased 50% year-over-year to a record $11.5 billion, driven by continued strength in AMD’s Data Center business, which represented 58% of company revenue in the quarter. Su framed the quarter as a launchpad for what comes next, saying “we enter the second half with strong momentum as EPYC demand accelerates, Instinct deployments scale and Helios begins to ramp.”
According to CNBC’s coverage of the call, Su said the company expects its data center sales to double in 2027, and that server revenue will be up more than 80% on an annual basis in the second half of fiscal 2026. Net income for the quarter hit $2.3 billion, or $1.38 per diluted share, up sharply from $872 million, or 54 cents per diluted share, in the year-ago period, CNBC reported.
TheStreet’s live coverage of the earnings call noted that Su wrapped up by hinting Wall Street’s 2027 expectations could actually be too low, not taking into account strength from Instinct AI GPU sales or Data Center segment growth tied to the new rack-scale Helios system. In the Q&A, Su told analysts that server revenue would grow over 80% year-on-year in the second half of 2026, and that the “volume ramp of Helios across major strategic customers in 2027” would accelerate GPU growth.
Helios Ships This Quarter, MI500 Already Locked for 2027
The annual-cadence pledge isn’t just an earnings-call soundbite. AMD’s own newsroom statement tied to its Advancing AI 2026 conference in San Francisco spells out specific silicon for specific years. AMD confirmed that it is extending its annual cadence of CPU, GPU, networking and rack-scale innovation through 2030, with next-generation EPYC server CPUs based on the “Zen 7” architecture coming in 2028.
Those Zen 7 chips carry the codenames “Florence,” “Ferrara” and “Fidenza,” expected to extend AMD’s leadership in density, performance, performance-per-system dollar and performance-per-watt, while a Zen 8 generation called “Ravenna” is coming in 2030, designed to continue AMD server CPU leadership. On the GPU side, AMD confirmed next-generation Instinct MI500 Series GPUs are coming in 2027, followed by Instinct MI600 Series GPUs in 2028.
The current-generation Helios rack itself — built around 72 Instinct MI455X GPUs and 18 sixth-generation EPYC “Venice” CPUs — is already moving through production. AMD has said shipments to customers, including Microsoft Azure, begin before the end of Q3 2026, with volume ramping into 2027.
Read also: AMD’s Helios AI Rack Enters Full Production, Microsoft Locks In Azure Deployment
Client and Gaming Segments Feel the Squeeze as AI Gets the Spotlight
The pivot to a yearly AI hardware cycle isn’t without trade-offs for AMD’s PC-facing business. TheStreet’s earnings coverage noted that the aggressive data-center growth outlook “will likely come at the expense of the company’s more modest Client and Gaming segments.” That’s a detail worth watching for PC builders and gamers who rely on Ryzen and Radeon parts, since it signals where AMD’s engineering and manufacturing priorities are trending.
Su was more upbeat about near-term PC demand on the call itself. According to CNBC, she told analysts “our first half performance has been very strong and although we are expecting that the market will decline in the second half, the market has actually held up better than most people would have thought,” referring to AMD’s consumer PC business. For Australian and New Zealand PC gamers watching component pricing and Ryzen availability, that mixed signal — booming data-center demand paired with a softer but resilient consumer PC market — is likely to shape how AMD allocates chip supply and wafer capacity through the rest of 2026.
Why AMD Is Betting the Company on a Yearly AI Rhythm
By committing publicly to a fixed annual release rhythm through 2030, AMD is trying to give hyperscaler customers — and its own investors — a predictable runway to plan multi-year AI infrastructure builds around. That mirrors the cadence Nvidia has used with its own GPU generations, and it puts pressure on AMD’s roadmap execution: missing a single year’s rack-scale launch would now read as a broken promise rather than a delay.
For now, the numbers back the confidence. A 50% jump in quarterly revenue, a data-center segment approaching 60% of total sales, and a fully specified silicon roadmap running to the end of the decade give AMD’s annual-cadence pledge more substance than a typical earnings-call soundbite — even as the company’s gaming and client hardware lines wait to see how much of that AI windfall trickles down to consumer products.
Join the Conversation