AMD Q2 2026 Call: Massive AI TAM Upgrade and 2027 Roadmap Visibility
AMD’s second-quarter call was defined by a significant upward revision of its long-term market outlook and the unveiling of its 2027 product roadmap. Management raised the 2030 Data Center AI accelerator TAM to $1.4 trillion and the Server CPU TAM to $220 billion, signaling a belief that AI adoption is still in its early innings. Beyond the immediate ramp of the MI450 'Helios' platform, the introduction of the 6th Gen EPYC 'Venice' and MI500 'Verona' series for 2027 provided investors with a clear multi-generati...
What changed on the call
Incremental information and investor implications beyond the release.
- 2030 TAM Expansion
AMD now expects the total high-performance and AI computing market to approach $2 trillion by 2030, with AI accelerators specifically reaching $1.4 trillion (up from previous estimates).
This massive TAM expansion justifies AMD's aggressive R&D spend and suggests the company sees a much longer runway for AI infrastructure investment than the current 'bubble' narrative suggests. - 2027 Roadmap: Venice and Verona
Management officially named the 2027 lineup: 6th Gen EPYC 'Venice' CPUs (Zen 6 on 2nm) and MI500 'Verona' GPUs. Venice is expected to deliver 3.3x performance-per-watt over ARM-based competitors.
Naming 2027 products now provides a 'bridge' for investors to look past 2026, emphasizing that AMD is moving to a yearly cadence for AI platforms. - Anthropic Strategic Partnership
Anthropic committed to deploying up to 2GW of MI450 'Helios' GPUs, with the first gigawatt starting in 1H 2027.
This is a major competitive win against Nvidia, securing a top-tier AI lab as a primary customer for the next-generation Helios platform.
Guidance and model clarifications
Analyst Q&A
Gross Margin Dynamics
Barclays, BofAJean Hu clarified that while AI GPUs are slightly below the corporate average margin, the massive volume and the accretive nature of high-core-count EPYC CPUs will allow AMD to manage the balance effectively.
Investor read-through: The mix shift toward Data Center is the primary driver of the 56% margin target, even with the GPU ramp.
Supply Constraints
Morgan Stanley, UBSLisa Su admitted server CPU supply is tight for the remainder of 2026 but emphasized that 2027 capacity for both wafers and HBM is well-aligned with their 'more than double' growth target.
Investor read-through: Supply, not demand, is the current ceiling for the server business in the near term.
Risks and watch items
- PC Market Softness
Management expects a 'softer' PC market in 2H26 due to higher memory and component costs weighing on consumer demand.
- HBM Cost Inflation
Analysts questioned the impact of 50% higher HBM content in Helios on margins; AMD is relying on yield improvements to offset these costs.
Management tone
- Long-term Visibility
Dr. Su noted that demand for the upcoming 'Venice' CPU is stronger than any prior EPYC generation at this stage.
Extremely bullish on the structural shift toward AI. - Supply Chain Execution
Jean Hu highlighted that while 2H26 server supply is tight due to 'unforecasted demand,' they have secured significantly more capacity for 2027.
Pragmatic but confident in scaling.
Reinforced on the call
- Yearly AI Cadence
AMD reiterated its commitment to launching a new rack-scale AI platform every year (MI350, MI450, MI500).
- Embedded Recovery
The embedded segment returned to strong YoY growth (19%), confirming the bottom is in for that high-margin business.
Primary source
TickerStar generated this brief from an official company, investor-relations, earnings-call, or regulatory source. Review the primary material before making an investment decision.
Open Earnings call webcastFollow AMD in real time
Get material earnings, filing, IR, and earnings-call alerts by email.