AMD sees AI accelerator market reaching $1.4 trillion by 2030

AMD Chief Executive Officer Lisa Su projected the AI accelerator market will reach $1.4 trillion by 2030, a forecast that more than doubles prior industry estimates and points to sustained demand for the company's MI-series graphics processors.

"The AI infrastructure buildout is still in its early innings, and we see demand accelerating across training, inference, and agentic workloads," Su said Wednesday at AMD's Advancing AI 2026 event in San Francisco.

The forecast compares with AMD's previous estimate of roughly $500 billion for the total AI chip market and comes as the company prepares to ship its Helios rack-scale system to customers including Microsoft, Meta Platforms, OpenAI and Oracle later this year. AMD also raised its server CPU market outlook to more than $120 billion by 2030, citing growing demand from agentic AI workloads that require greater CPU-to-GPU parity in data centers.

The projection gives investors a new framework for valuing AMD's AI business as it challenges Nvidia's dominance in data center GPUs. AMD shares have gained 152.8% year to date, outperforming the Philadelphia Semiconductor Index's 72.7% advance, though the stock fell about 11% last week as semiconductor stocks broadly pulled back after Micron Technology triggered concerns about Chinese memory-chip competition.

The $1.4 trillion target encompasses graphics processors, custom ASICs, networking silicon and memory subsystems, Su said. AMD's MI450X accelerator and MI500 GPU family, expected to be detailed during the two-day event, will compete directly with Nvidia's Blackwell and Rubin architectures in the data center segment. The MI500 is expected to feature HBM4 memory, according to AMD's published roadmap, giving it a bandwidth advantage over Nvidia's current-generation offerings.

Helios and the Rack-Scale Push

AMD's Helios rack-scale system, introduced Monday, bundles the company's GPUs, EPYC server processors and networking into a single chassis, mirroring Nvidia's DGX platform strategy. The system is expected to begin shipping to customers in the second half of 2026. Microsoft, Meta, OpenAI and Oracle have committed to deployments, Su said. The rack-scale approach allows AMD to sell complete AI infrastructure rather than individual chips, potentially increasing its revenue per customer by an order of magnitude.

CPU Supply Tightens as AI Demand Spreads

The AI buildout is also reshaping the server CPU market. Intel and AMD have signed longer-term purchase commitments with Chinese data-center customers as server CPU prices surge, with some products rising more than 40% since the start of the year, Reuters reported Wednesday. The deals typically lock in volume without fixing prices, reflecting how AI demand has spread beyond accelerators to CPUs, memory and networking gear. Intel CEO Lip-Bu Tan told analysts in April that demand "continues to run ahead of supply," especially for Xeon server processors, with lead times stretching to six months for some products.

AMD reports second-quarter earnings on Aug. 4. The company's data center GPU business is expected to exceed its long-term target of an 80% compound annual revenue growth rate, AMD said in May. Nvidia, which commands an estimated 80% of the AI accelerator market, trades at about 35 times forward earnings. AMD trades at roughly 28 times forward earnings, reflecting the market's view that it remains the challenger despite gaining share. Cathie Wood's Ark Invest has sold about 145,550 AMD shares worth roughly $79 million this month, trimming its position even as the stock has outperformed.

This article is for informational purposes only and does not constitute investment advice.