AMD announced that it has entered into a definitive agreement to acquire New Jersey-based ZT Systems. Expected to close by the first half of 2025, the deal is a cash and stock transaction valued at $4.9 billion, which includes $400 million of performance-based, post-acquisition payouts.

ZT provides hyperscale and enterprise server solutions for general purpose cloud computing and accelerated AI infrastructure. These solutions are brought to bear through ZT’s design and customer enablement teams and their manufacturing teams. With no desire to compete with its OEM and ODM partners, AMD has indicated that it is looking for a strategic partner to divest the manufacturing team and resources following the transaction close.

In the interim, design and customer enablement will be led by current ZT President Doug Huang and manufacturing will be led by ZT CEO Frank Zhang, with both reporting to Forrest Norrod, the AMD EVP and general manager of the Data Center Solutions Group.

This is the second major AI acquisition by AMD in as many months. According to AMD Chair and CEO Lisa Su, “Our acquisition of ZT Systems is the next major step in our long-term AI strategy. ZT adds world-class systems design and rack-scale solutions expertise that will significantly strengthen our data center AI systems and customer enablement capabilities.”

With this acquisition, along with the recently completed Silo AI acquisition, AMD takes another significant step towards providing full stack, vertically integrated AI system solutions.

ZT’s expertise in providing rack-scale systems is clearly a major component of this strategy. Accelerated AI infrastructure is a system-level problem when it comes to optimizing for total cost of operations driven by performance, power and cooling. The goal with this acquisition is to ultimately meet the changing needs of AMD’s customers through solutions that start at the system level from a design perspective and go down through the silicon design, which is AMDs core expertise. Designing solutions from the system level will help them not only address the issues holistically but also minimize the time between next generation silicon qualification and rack solution availability.

Any analysis of this acquisition would be remiss if its customer enablement aspect is not also highlighted. It should be noted that AMD’s other recent significant AI acquisition, Silo AI, also has world-class customer enablement capabilities. Application engineering and consulting services are critical when trying to challenge an entrenched incumbent. In competitive situations, it is critical to demonstrate technical and performance parity or even superiority (which arguably AMD has done and will continue to do) and to make it as easy as possible for the customer to make the switch. Strong customer enablement expertise and resources are valuable assets in both AMD AI acquisitions.

As indicated by the announcement, the transaction is expected to close in the first half of 2025. Assuming there are no major regulatory obstacles, and based on AMDs track record, it would not be surprising if it is completed closer to the first quarter as opposed to the first half.

However, even with that, it still takes time and considerable effort to properly integrate an acquisition of this size. TIRIAS Research expects that with the long silicon development cycle, it may be several generations before there is a tangible impact on the AMD product roadmap. On the customer enablement front, however, this can happen very quickly from day one. It would be surprising if strategy conversations around that topic were not already happening given that ZT is already an AMD partner.

Finally, acquisitions of this magnitude are not made without some level of vetting with both current and potential customers and AMD is clearly seeing demand for a more vertically integrated approach to address the evolving requirements associated with next-generation accelerated AI infrastructures. As with most things, it will all come down to execution and if there is one thing Su and the team have proven during her tenure, it is that AMD excels at execution.

By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved