Ben Pouladian
Founder of BEP Research, AI Compute Economics
NVIDIA can retain demand despite newer chips and custom silicon because customers still rent H100s and Anthropic also secures its compute. Silicon Data's one-month snapshot has H100 rent up 10.8%, more than B200 and B300, though A100 rent fell 0.6%. Anthropic is pursuing custom silicon with Google, Broadcom, Amazon and AMD while securing NVIDIA-powered compute via Microsoft, SpaceXAI, Nscale and Lambda; the chart counts selected relationships, not spending share. x.com x.comÂ
Nvidia can capture more AI infrastructure spending by supplying Vera CPUs and networking alongside GPUs. Oracle outlined an initial supply of 150,000 Vera CPUs, and OpenAI says it is putting them into production for agent tool execution; reducing CPU waits could improve cost per task and GPU-memory use. Benefits from independently scaling CPUs and GPUs still need validation in utilization and completion times. x.comÂ
Cerebras should dominate the enormous fast-inference market because agentic workflows often run steps sequentially, making each step's speed critical. Its newly announced Oracle partnership brings its ultra-fast inference to Oracle Cloud Infrastructure. Cerebras delivers inference speeds up to 13x faster than Nvidia GPUs, and its chips should eventually reach every major AI cloud. x.comÂ
Low Anthropic spending relative to customer expenses points to an enterprise AI adoption gap that leaves Nvidia reliant on intermediaries able to build their own chips. Anthropic's S-1 says over half the Fortune 500 were $100k+ annual customers as of September, yet spending among its publicly traded customers averaged under 0.1% of expenses in Q3 2026. That measures one provider and cannot show whether AI deployments succeed; Nvidia needs more enterprises buying its compute directly. x.comÂ
Nvidia's system components can reduce the idle time that erodes returns on expensive AI chips. CPUs, accelerator interconnects and server networking help operators such as Nebius and IREN keep machines doing useful work; solving the next bottleneck could win Nvidia more of the data-center equipment budget even when a rival chip looks appealing alone. Buyers must compare the cost of completed work across the whole system. x.comÂ
Enterprise and neocloud demand is creating a second wave of AI infrastructure spending, with tailwinds for Nvidia and Super Micro. Dell forecasts 200 GW of AI data center growth through 2030, half driven by these customers, and estimates a $1 trillion AI infrastructure opportunity over the next four years. AI servers could enter a new growth phase in 2027, with higher shipment volumes and more revenue per deployment. x.comÂ
Nvidia's neocloud orders increasingly require its own credit support. Per The Information, it is in talks to guarantee billions of Nscale debt for a 460MW data center; Nscale needs an IPO to fund $24B in Nvidia systems committed for 2026-27. Separately, a quoted article says US officials identified Alice Huang, scrutinized for possibly routing Nvidia chips to China, as a Chinese intelligence asset with direct access to Jensen Huang. This is just the tip of a very big iceberg. x.com x.com x.comÂ
NVIDIA's CUDA may become the easiest platform for AI agents to accelerate because it has the deepest, broadest and most mature substrate to work on. OpenAI used NVIDIA's programming stack to unlock ultra-fast performance, and NVIDIA says GPT-6 Astra Ultrafast runs up to 8x faster on its platform. x.com x.comÂ
NVIDIA's advantage extends beyond GPUs into engineering work that improves customers' AI economics. Its work with Meta on Andromeda redesigned ad-retrieval models and pipelines, processing over 1 billion ads per second on one Hopper GPU server with 100 times less network traffic and 300 times fewer FLOPs per inference. Meta says further optimization is underway for Grace Blackwell. x.comÂ
Nvidia's push into AI models and agent software looks weak even though its GPUs and racks are great. Its Nemotron model is mediocre and its LangChain-based agent stack won't make it. If Nemotron were capable of frontier performance it would be there already, and Nvidia doesn't seem to know how to get there. x.com x.comÂ
NVIDIA is steering toward highly diversified revenue, with enterprises and neoclouds a growing share, reducing its concentration in the largest hyperscalers and frontier AI labs. Law firm Latham & Watkins is buying NVIDIA servers and having its engineers fine-tune Nemotron 3 open-weight models in-house. NVIDIA's open-source push, including the Poolside and Hugging Face acquisitions, accelerates that diversification. x.comÂ