Loading...
Loading...
AMD, led by CEO Lisa Su, is accelerating its data-center AI strategy as it begins sampling the new Instinct MI450 GPU to key customers and plans volume Helios rack shipments in H2 2026. Strong Q1 results—driven by a 57% year-over-year data-center revenue jump—boost investor confidence and underpin guidance above Street estimates. AMD forecasts data-center AI revenue reaching the tens of billions by 2027, citing demand outpacing prior internal forecasts and multi-gigawatt deployment talks with major partners. The company will showcase roadmaps and developer tools at a May 2026 AI Developer Day in Shanghai and reveal MI500 details at a July event.
AMD's accelerated AI push affects data-center architecture choices and procurement cycles for cloud, enterprise, and AI startups. Tech professionals should track new GPU and rack offerings that influence performance, software stacks, and deployment timelines.
Dossier last updated: 2026-05-12 12:13:01
AMD chair and CEO Lisa Su will attend the 2026 AMD AI Developer Day in Shanghai on May 19, the company confirmed. The event will focus on AI compute, system architecture, open-source software ecosystems and real engineering challenges, positioning itself as a hands-on technical day for developers working on models, systems, inference, training, toolchains and deployment. Programming includes expert-led workshops, leader keynotes, face-to-face sessions with AMD engineers and networking with AI practitioners. AMD says the goal is to present its roadmap for AI and demonstrate long-term support for developers rather than a marketing-only showcase. The event signals AMD’s push to engage developers and showcase its AI platform strategy in China.
AMD reported stronger-than-expected Q1 results, sending its stock up over 20% premarket after revenue of $10.25 billion and adjusted EPS $1.37 beat LSEG consensus. Data center revenue jumped 57% year-over-year to $5.8 billion, making the division the primary growth driver; AMD guided Q2 revenue to about $11.2 billion, above analyst estimates. CEO Lisa Su said AMD expects data center AI revenue to reach “hundreds of billions” next year and aims for sustained high growth as capacity expands. AMD plans to ship its rack-scale Helios AI systems later this year, with OpenAI and Meta signed as customers. The results underscore investor optimism around AI-driven chip demand and competition with NVIDIA in GPUs and CPUs.
AMD CEO Lisa Su said during the Q1 2026 earnings call that AMD has begun sampling the Instinct MI450 GPU to key customers and will ramp shipments of Helios AI racks in the second half of 2026. She reported customer demand outpacing AMD’s internal 2027 forecasts, with major customers and new partners negotiating large-scale deployments; OpenAI and Anthropic were named as linked parties. The MI450 (CDNA 5) doubles key metrics versus MI350: up to 40 PFLOP FP4, 20 PFLOP FP8, 432 GB HBM4 at 19.6 TB/s, 300 GB/s lateral interconnect, and variants MI455X (AI) and MI430X (HPC/sovereign AI). AMD positions MI450 against NVIDIA’s Vera Rubin and will reveal MI500 details at its July event.
AMD CEO and Chair Lisa Su said on the company’s Q1 2026 earnings call that the Helios data center AI platform is on track, with chip, software and system development meeting key milestones. AMD has already begun sampling the MI450 series GPUs to core customers and plans volume production in H2 2026. Customer demand has outpaced initial forecasts, with more large-scale deployment talks and additional multi-gigawatt opportunities emerging. Based on expanded demand visibility, Su expressed confidence that AMD’s data center AI business could reach annual revenue in the “tens of billions of dollars” by 2027 and exceed its long-term growth targets of over 80% in coming years. The update signals AMD’s stronger push into AI accelerators and server markets.