Business

AMD launches Helios AI rack system, securing Microsoft as customer

Microsoft becomes the newest client for the new Helios platform, joining a roster of existing technology clients including Meta, OpenAI, and Oracle.

Author
Owen Mercer
Markets and Finance Editor
Published
Draft
Source: CNBC · original
AMD launches Helios, its first rack AI system to rival Nvidia, adding Microsoft as newest buyer
Semiconductor giant enters rack-based AI infrastructure market with direct challenge to Nvidia

Advanced Micro Devices (AMD) has officially launched Helios, its first rack-based artificial intelligence system, explicitly positioning the new hardware to compete with market leader Nvidia. The launch marks a significant strategic move for the semiconductor manufacturer as it seeks to expand its footprint in the rapidly growing infrastructure sector.

Microsoft has secured a purchase of the Helios system, becoming the newest customer for the platform. This acquisition places Microsoft alongside a growing list of established clients that already utilise AMD’s AI hardware, including Meta, OpenAI, and Oracle.

The introduction of Helios represents AMD’s formal entry into the rack-based AI system market. By offering a complete rack solution rather than solely individual chips or accelerators, the company is attempting to capture a larger share of the capital expenditure spent by major technology firms on data centre infrastructure.

While the system is designed to rival Nvidia’s dominant position in the AI chip market, specific technical specifications, performance benchmarks, and pricing details for the Helios system were not disclosed in the announcement. Similarly, the scale of Microsoft’s purchase, including the number of racks or total contract value, remains unspecified.

The launch occurs against a backdrop of intense competition in the AI infrastructure space. Nvidia currently holds the incumbent market leadership, and the entry of AMD’s integrated rack system signals a continued effort by competitors to diversify the supply chain for large language model training and inference workloads.

Continue reading

More from Business

Read next: US House passes War Powers resolution to constrain Trump’s Iran authority
Read next: Global equities tumble as oil breaches $100 amid US-Iran tensions
Read next: Comer Alleges Staley Defended Epstein at JPMorgan Despite Internal Concerns