AMD secured a 6 gigawatt GPU partnership with Meta, one of the largest AI infrastructure deals disclosed to date. The chipmaker simultaneously invested in Nutanix to strengthen hybrid cloud AI capabilities, positioning against NVIDIA's enterprise dominance.
Red Hat and NVIDIA co-launched enterprise agentic AI platforms. Supermicro deployed Red Hat-certified systems for AI factories. "Our validated solutions for the Red Hat AI Factory with NVIDIA help ensure that customers can combine our high-performance systems with a robust, enterprise-grade software platform," said Vik Malyala of Supermicro.
Cryptocurrency mining operators are pivoting to AI infrastructure. DMG Blockchain Solutions adjusted equipment operations to prioritize profitability over hashrate generation after receiving a $1.5 million energy efficiency incentive. Mawson Infrastructure Group reported a preliminary $23.8 million net loss for 2025 while shifting focus to data center conversion.
Pure Storage rebranded as Everpure and announced acquisition activity expected to close in Q2 FY27, expanding storage infrastructure for AI workloads. Mawson adopted a limited-duration stockholder rights agreement but confirmed the board remains open to proposals in stockholder interests.
The AMD-Meta deal directly challenges NVIDIA's 80%+ data center GPU market share. Meta's 6 gigawatt commitment represents enough power for roughly 1.5 million high-end GPUs, a multi-billion dollar revenue opportunity extending through 2027-2028.
Traders should watch AMD's gross margins in coming quarters. GPU deals at hyperscaler scale typically compress margins 5-8 percentage points versus consumer chips. NVIDIA faces pressure if Meta's AMD adoption proves successful at scale.
The crypto-to-AI facility conversions matter for infrastructure REITs and power utilities. Data centers running AI workloads consume 3-5x more power per rack than traditional compute. Companies with stranded mining infrastructure and cheap power contracts gain conversion optionality.
Red Hat's enterprise AI platform launch with NVIDIA validates the hybrid cloud AI thesis. Enterprises want on-premises AI for data sovereignty. This creates a wedge for AMD and Intel to compete on price-performance in private cloud deployments where NVIDIA's CUDA moat is weaker.
Watch capital expenditure guidance from cloud providers in March earnings. Hyperscaler AI infrastructure spending is running $200B+ annually. Any allocation shift toward AMD chips moves market share billions at a time.