Advanced Micro Devices Sees AI Boom Fueling MI450, Helios and Server CPU Growth

Key Points

• AMD expects a major AI product ramp: MI450 shipments are scheduled to begin at the end of the third quarter, with a sharp increase in the fourth quarter and further growth through 2027. Helios rack-scale systems are being developed with customers including Meta, OpenAI and Anthropic, although deployment involves significant supply-chain and manufacturing complexity.
• AI demand is broadening beyond GPUs: AMD sees inference and agentic AI driving demand for server CPUs, specialized accelerators, adaptive computing and AI PCs. The company raised its estimated 2030 addressable market to $2 trillion and expects its data-center business to double next year.
• Server CPU growth is strong but supply remains tight: AMD forecasts server CPU revenue growth of more than 80% in the second half and over 70% next year, supported by higher unit volumes and average selling prices. Capacity constraints span wafers, advanced packaging, high-bandwidth memory and other components, while the MI450 ramp may create some gross-margin dilution.

Advanced Micro Devices NASDAQ: AMD executives said artificial intelligence demand is broadening beyond graphics processors to include server CPUs, AI PCs and adaptive computing, while the company prepares for a significant ramp of its MI450 accelerator platform and Helios rack-scale systems.

Speaking at the Citi Global TMT Conference, Chief Financial Officer Jean Hu described AI as “the most consequential technology transformation” and said the market has shifted as inference demand has outpaced training demand. She also pointed to the rise of agentic AI, which she said is creating increased demand for both GPUs and CPUs.

AMD updated its estimate of its addressable market opportunity to $2 trillion by 2030 at its Advancing AI event, according to Hu. The company expects its data center business to double next year, driven by GPU and server CPU demand.

MI450 and Helios Ramp

Hu said AMD expects production shipments of MI450 products to begin at the end of the third quarter, with revenue beginning in the quarter. The company expects a “very significant step up” in the fourth quarter, followed by another increase in the first quarter of 2027 and continued ramping through that year.

She said deploying Helios at rack scale adds complexity beyond the availability of GPUs, CPUs and high-bandwidth memory. AMD is working with original design manufacturers, supply-chain partners and customers on component availability, manufacturing processes, mechanical systems and software.

“The demand and the volume for 2027 has certainly gone above our original initial expectation,” Hu said, adding that AMD is seeking to expand supply to meet customer requirements.

The company identified Meta, OpenAI and Anthropic as three anchor customers for multigigawatt-scale deployments and multigenerational engagements. Hu also said demand for the MI350 platform has been strong among model builders, newer AI companies and enterprise customers. AMD expects to work with neocloud providers to serve customers beyond its largest strategic accounts.

Matt Ramsay, AMD’s corporate vice president of financial strategy and investor relations, said the company’s technical engagement with its large model-building customers is helping shape future Instinct MI500 and MI600 products as well as subsequent rack designs.

Inference Strategy and Taalas Acquisition

Ramsay said inference has become the majority driver of AI computing demand, with a shift from chatbot applications toward agentic inference. He said this trend has created opportunities for inference silicon as well as CPUs used to run AI agents.

AMD expects GPU-led computing to account for the largest portion of the inference market, with its MI455 product ramping alongside Helios. However, Ramsay said the market also includes application-specific chips, other accelerators and an emerging segment for disaggregated inference requiring ultra-fast response times.

AMD has announced a partnership with Cerebras under which Helios systems will be deployed in Cerebras’ cloud alongside its wafer-scale engine racks. Ramsay said AMD also has internal silicon ambitions for ultra-low-latency inference using chiplets. The company’s acquisition of Taalas adds engineering talent and technology supporting that effort, he said.

Asked about competitive positioning, Ramsay said AMD believes it offers a tokens-per-dollar advantage for large-scale inference in its current generation of products. He said the company intends to expand training capabilities with the MI450 series and future generations, while participating across GPUs, CPUs, customized systems, optics, networking and system-level design.

CPU Demand, Pricing and Margins

AMD expects server CPU revenue to expand more than 80% year over year in the second half and by more than 70% next year, Hu said. The business remains supply-constrained, she added.

The company raised its 2030 total addressable market forecast for server CPUs to more than $220 billion, from $60 billion projected at its November 2025 financial analyst day. Hu said the expansion is being driven largely by agentic AI, which requires CPUs for data retrieval, workflow execution, orchestration and coordination with other computing resources.

Ramsay said AMD aims to reach 50% of the server CPU dollar market over time, including competition across instruction sets. He said the company’s enterprise server business grew more than 70% in the second quarter and characterized the upcoming Venice server platform as a particularly compelling program.

Hu said CPU growth is being supported by both unit gains and higher average selling prices, with unit growth exceeding price increases. Higher core counts in successive product generations have contributed to pricing, she said. AMD also expects to pass through increases in component costs, such as wafer pricing, but does not intend to raise prices solely to expand gross margin.

While AI accelerators remain below AMD’s corporate-average gross margin, Hu said their ramp should drive substantial revenue and gross-profit-dollar growth. She said stronger server CPU and embedded-business performance should help offset some margin dilution from data center AI products. AMD guided for a 56% gross margin in the third quarter and expects future quarterly results to vary as MI450 ramps.

Supply Chain and Manufacturing

Hu said supply remains tight across wafers, advanced process nodes, high-bandwidth memory, advanced packaging, substrates and other components. AMD has been preparing its supply chain for the MI450 and Helios ramp and is increasing capacity for its server CPU business, particularly ahead of the Venice platform launch.

The company has increased capital expenditures primarily to build capacity for CPU production, including purchasing equipment and using consignment arrangements, Hu said. AMD expects Taiwan Semiconductor Manufacturing Co. to remain its primary wafer supplier for the foreseeable future, citing a longstanding strategic and technology-development relationship. Hu also said AMD is diversifying geographically through TSMC’s Arizona fab.

Looking ahead, Ramsay said AMD plans to offer both copper-based and near-package optical connectivity options for larger scale-up domains in the MI400 series, expected to be introduced in the second half of 2027. He said copper and optical technologies are likely to coexist for multiple product generations rather than transition through a single industry-wide shift.

About Advanced Micro Devices (NASDAQ:AMD)

Advanced Micro Devices, Inc NASDAQ: AMD is a global semiconductor company that designs and sells microprocessors, graphics processors, chipsets and adaptive computing solutions for a broad set of markets. The company's product portfolio includes consumer and commercial CPUs under the Ryzen and Threadripper brands, data center processors under the EPYC brand, and Radeon graphics processing units for gaming and professional visualization. AMD also offers semi-custom system-on-chip (SoC) products for gaming consoles and other specialized applications, and provides supporting software and platform technologies for OEMs, cloud service providers and end users.

Founded in 1969, AMD has evolved from a supplier of logic chips into a diversified, fabless semiconductor designer.

This instant news alert was generated by narrative science technology and financial data from MarketBeat in order to provide readers with the fastest reporting and unbiased coverage. Please send any questions or comments about this story to contact@marketbeat.com.