NVIDIA has introduced NVHBM, a new custom high-bandwidth memory technology, and folded it into its NVLink Fusion ecosystem as AI infrastructure demands shift beyond raw compute toward tighter coordination across processors, memory, interconnects, and cooling. AWS chip unit Annapurna Labs is the first partner to adopt the technology and said its next-generation in-house Trainium4 chip will support both NVLink Fusion and NVHBM. The announcement came alongside a separate commercial commitment: Amazon confirmed plans to buy up to 2 million next-generation high-end GPUs from NVIDIA between 2027 and 2028, covering Blackwell Ultra, Rubin, and Rubin Ultra. According to ABMedia, the arrangement reflects a changing relationship between hyperscalers and NVIDIA, where internal chip development and external GPU procurement now sit side by side rather than in direct opposition. ABMedia said NVHBM moves the memory controller into the base die of a 3D HBM stack, freeing die space on the XPU for more compute cores or cache. The report cited gains of up to 30% more memory bandwidth, about 15% lower HBM power consumption, and up to 25% more chip area released for compute. NVIDIA is also pushing the design toward a standardized specification across suppliers including SK hynix, Samsung, and Micron.

Amazon founder Jeff Bezos said in an interview with Fortune that the company’s chip business is set to become its next long-term pillar after Marketplace, Prime, and Amazon Web Services, or AWS. He said Amazon plans to roll out its next-generation chip, Trainium4, in 2027. The company’s chip effort is primarily built around Annapurna Labs, the chip design company Amazon acquired in 2015. Annapurna Labs is now responsible for designing Amazon’s AI chips, making it central to the group’s push to expand its in-house semiconductor capabilities. The update follows comments made earlier by Amazon CEO Andy Jassy during the company’s second-quarter 2026 earnings call. Jassy said both Amazon’s AI chip business and its broader in-house chip operation had reached annualized revenue run rates of more than $25 billion each. He also said those businesses held more than $225 billion in multi-year customer commitments. The remarks were cited by TheStreet and carried by Odaily.

Amazon has fully funded its $50 billion commitment to OpenAI and received roughly a 5% equity stake, implying a valuation of about $852 billion. The unusual part is timing: $35 billion of the commitment had originally been tied to conditions such as an IPO or a technical breakthrough, yet Amazon paid the full amount even though neither milestone has been met. The move came after OpenAI and Microsoft rewrote their agreement on April 27, loosening earlier exclusivity around API hosting and cloud access. Following that change, Amazon expanded an AWS deal from $38 billion to $100 billion over eight years and became the exclusive external cloud provider for OpenAI’s Frontier platform. The agreement also includes up to 2 gigawatts of Trainium compute capacity, covering Trainium3 and the planned Trainium4. Amazon is also backing Anthropic, with commitments there reaching as much as $33 billion and actual investment at $18 billion so far, showing how large its AI infrastructure bet has become.