Deal will see AMD will retain ZT Systems' AI systems design businesses.
AMD has sold the US-based data center infrastructure manufacturing business of ZT Systems to Sanmina Corporation.
Sanmina paid $3 billion in cash and stock, and under the terms of the deal, AMD will retain ZT Systems' AI systems design businesses while working with Sanmina as a US-based preferred new production introduction (NPI) manufacturing partner for its AI rack and cluster-scale systems.
The deal is expected to close towards the end of 2025, subject to regulatory approvals and customary closing conditions.
AMD completed its own $4.9bn acquisition of ZT Systems in March 2025, following a prior $1bn investment in the server maker. When the deal was completed, AMD said it was already in talks with 'multiple potential strategic partners’ regarding the sale of the US part of the business.
ZT Systems currently designs, integrates, manufactures, and deploys servers for companies like Amazon Web Services and Microsoft Azure. In a statement, Sanmina said ZT Systems’ annual revenue is approximately $5-6 billion, while its outlook is “promising as new accelerated compute platforms are launched.”
"The acquisition of ZT Systems' manufacturing operations positions Sanmina as an industry leader in the cloud and AI ecosystem and enables us to further capitalize on the significant growth opportunity of this market," stated Jure Sola, chairman and CEO of Sanmina Corporation.
"Combining Sanmina's global expertise and vertical integration capabilities with ZT Systems' high-quality manufacturing capacity in the US and Europe and extensive experience with critical cloud and AI infrastructure for hyperscalers will give us enhanced scale and deepen our customer relationships.”
EVP of AMD’s data center solutions business, Forrest Norrod, said: "We look forward to working with Sanmina as our preferred NPI manufacturing partner. This agreement will help accelerate the US-based manufacturing of AMD AI end-to-end training and inference solutions – which are optimized for our customers' unique environments, ready-to-deploy at scale, and based on our open approach.”
