Full Breakdown
AMD and Rackspace Seal Multi-Year AI Compute Deal
6/17/2026, 12:18:53 AM
The Deal: 30 MW of AMD-Powered AI Infrastructure
AMD and Rackspace Technology announced a definitive agreement to install up to 30 megawatts of AMD-based compute across Rackspace’s global data-center network. The rollout begins in the fourth quarter of 2026 and is slated for completion by 2028. The hardware stack will combine AMD Instinct GPUs (MI355X, MI350P and future successors) with AMD EPYC CPUs, forming an “Enterprise AI Cloud” that routes workloads to the appropriate compute tier while maintaining end-to-end performance accountability.
Background & Context
A memorandum of understanding signed on May 7 2026 outlined a collaborative approach to regulated-sector AI. The definitive agreement upgrades that MOU to a binding contract, positioning AMD as the silicon-layer partner for Rackspace’s governed AI stack and targeting compliance-heavy markets such as healthcare.
Key Figures & Organizations
- Rackspace Technology, Inc. – Cloud-services provider pursuing AI infrastructure for regulated enterprises.
- Advanced Micro Devices (AMD) – Semiconductor company supplying GPUs and CPUs for the deployment.
- Gajen Kandiah, CEO, Rackspace Technology.
- Dan McNamara, Senior Vice President of Compute and Enterprise AI, AMD.
Timeline
- May 7 2026 – MOU announced.
- June 16 2026 – Definitive agreement disclosed; shares react.
- Q4 2026 – Installation of the first AMD compute capacity.
- 2028 – Full 30 MW deployment expected.
Data & Statistics
- 30 MW of dedicated AMD compute.
- Rackspace shares rose 13 % to $6.66 in early trading and 21 % to $7.15 in pre-market.
- Market capitalization reached roughly $1.7 billion.
- FY 2026 revenue guidance: $2.6 billion–$2.7 billion; adjusted EBITDA projected at $305 million–$315 million.
- Q1 adjusted EBITDA: $71.2 million; debt $2.71 billion, cash $93.6 million.
- Planned staff reduction: 15 % of global workforce, with one-time charges of $14 million–$19 million and anticipated run-rate savings of $75 million–$85 million.
Why It Matters
The partnership creates a purpose-built AI platform for sectors where data governance and regulatory oversight are paramount. Early interest from healthcare providers suggests a market for large-scale clinical AI inference. For investors, the deal signals a concrete hardware roadmap for Rackspace, potentially stabilizing revenue streams amid broader cloud-service competition.
Official Statements & Responses
Rackspace described AMD as its “principal chip provider” and emphasized that the collaboration delivers a governed AI stack with a single accountable partner from silicon to outcomes. AMD highlighted the need for a balanced mix of accelerated and general-purpose compute to meet enterprise requirements. Both companies committed dedicated sales and marketing resources to pursue regulated-industry customers.
Criticism & Opposition
Analysts noted execution risk: financing for each deployment must be secured, and commercial sign-offs remain pending. The announced 15 % workforce reduction raises concerns about operational capacity during the rollout. Rackspace’s debt load and modest cash balance were cited as constraints on scaling the initiative.
Conflicting Reports & Gaps
While the agreement outlines a 30 MW target, AMD has indicated that each deployment requires separate financing terms and that AMD is not obligated to accept any specific deal. The precise timeline for securing customer contracts and the impact of the layoffs on project staffing were not disclosed.
Verbatim Quotes
- “governed AI stack with one accountable partner from silicon to outcomes.” — Gajen Kandiah, CEO, Rackspace Technology
- “right mix of accelerated and general-purpose compute.” — Dan McNamara, SVP, Compute and Enterprise AI, AMD
- “governed from the ground up.” — Gajen Kandiah, quoted by Barron’s
- “high-performance AI infrastructure with the openness, scalability and accountability needed to run AI at enterprise scale.” — Dan McNamara, AMD
What’s Next
Rackspace must finalize financing and secure customer agreements for each phase of the deployment. The company plans to reinvest a portion of the projected run-rate savings into forward-deployed engineering and AI solutions delivery. Monitoring the progress of the staff reductions and the impact on project timelines will be critical for assessing the partnership’s long-term viability.
