The Vera Rubin stack also includes the Groq LPU announced earlier in the quarter. NVIDIA's "revenue per gigawatt" metric — the company's preferred measure of compute productivity — has grown from $18 billion (Hopper) to $25 billion (Blackwell) to $40 billion (Vera Rubin, including CPU, GPU, NVLink, networking, and LPU). CEO Jensen Huang's framing on the call: customers "want to race to the next generation as fast as they can" because NVIDIA's compute is so productive that the tokens the GPUs generate are "insanely profitable" for them (Carson Group earnings call transcript, August 26, 2026).
AWS 2-million GPU expansion
The AWS expansion announced on the call is the largest single deployment commitment of its kind. AWS is deploying an additional 2 million NVIDIA GPUs starting in the current quarter through Q2 FY2029, along with Vera CPUs. The expansion builds on AWS's already vast installed base of NVIDIA compute. AWS's NVIDIA spend is concentrated in Trainium and Inferentia silicon as well, but the 2-million-GPU commitment is the largest disclosed GPU buildout from a single hyperscaler (Carson Group earnings call transcript, August 26, 2026).
The expansion matters for GPU pricing: more committed volume means NVIDIA can negotiate multi-year framework agreements at slightly lower per-unit margins, but the absolute revenue contribution is enormous. For neoclouds like CoreWeave and Nebius, the AWS scale-up increases the supply of alternative cloud capacity, which has historically pressured per-hour GPU rental rates on the secondary market. The Motley Fool estimates NVIDIA captures roughly 24% of hyperscaler CapEx, which implies about $315 billion in revenue from hyperscalers in calendar 2027 alone (Motley Fool, September 3, 2026).
MLPerf, Apple Private Cloud, and the supply story
NVIDIA Blackwell led across every category in the MLPerf Training 6.0 benchmarks and in AgentPerf, the industry's first agentic AI infrastructure benchmark. NVIDIA GPUs with Confidential Computing are now used for confidential inference in Apple's Private Cloud Compute — a flagship enterprise inference deployment. MLPerf wins are a recurring moat: enterprise buyers who follow the benchmark cycle default to the leader, and the gap between Blackwell and the prior generation typically translates to a multi-quarter pricing premium (NVIDIA Q2 FY27 press release, August 26, 2026).
On the supply side, NVIDIA's preliminary expectation is for FY2028 revenue to grow approximately 70% year-over-year, and management says supply is what limits that figure. CEO Jensen Huang put it more directly: "even though our demand is much greater than 70%, our supply allows us to confidently deliver 70%." Memory cost increases have driven AI server price increases above 15% in some cases, but NVIDIA's pricing power has protected profitability while still meeting demand from customers with urgent capacity needs (Futurum Group, August 2026).
What enterprise buyers should watch next
The next NVIDIA earnings call is November 17, 2026, for Q3 FY2027 results. Between now and then, the catalysts to watch are: (1) the Goldman Sachs Communacopia + Technology Conference in San Francisco on September 10 (Jensen Huang fireside keynote), (2) GTC Berlin on October 21 (Jensen Huang keynote), (3) further cloud-provider framework announcements, and (4) any HBM supply tightness that could shift the AI server pricing dynamic. For GPU cloud buyers, the practical implication of Vera Rubin going to 5 production partners is that Blackwell Ultra pricing should ease over Q4 2026 as Vera Rubin absorbs the highest-priority workloads.