Nvidia has spent years selling the picks and shovels of the AI gold rush. Now it’s telling Wall Street the gold rush is about to become a full-blown economy.

The company is positioning AI as no longer an experimental technology confined to research labs and Silicon Valley demos, but something ready for deployment across industries at scale. And it’s putting serious capital behind that thesis, with commitments exceeding $40B in 2026 directed toward AI infrastructure partners spanning cloud computing and data center buildouts.

The OpenAI deal and the data center bet

Perhaps the most eye-catching signal of Nvidia’s confidence is its deepening relationship with OpenAI. The two companies have signed a letter of intent that could see Nvidia invest up to $100B, with the goal of deploying at least 10 GW of AI data center capacity by 2026.

To put that in perspective, 10 gigawatts is roughly the electricity consumption of a country like Jordan. That’s a staggering amount of compute power dedicated to training and running the next generation of AI models, and it tells you everything about how Nvidia sees the trajectory of demand.

This isn’t a moonshot research grant. It’s an infrastructure play, the kind of commitment companies make when they believe the market is about to shift from “promising” to “unavoidable.” Nvidia is essentially building the power grid for an AI-native economy.

The partnership also signals something subtler. Nvidia isn’t content to just sell chips to AI companies. It wants to be embedded in the infrastructure layer itself, making it harder for competitors to dislodge even as the market matures and alternative chip architectures emerge.

Beyond the cloud: robots that think locally

While data centers grab headlines, Nvidia is also making a quieter bet on the edge, meaning AI that runs on devices rather than in distant server farms.

Enter Jetson Thor, a $3,499 developer kit designed to let autonomous robots run multiple AI models without needing a cloud connection. Think of it as giving robots their own brain instead of making them phone home every time they need to make a decision.

This matters more than it might sound. Cloud-dependent AI introduces latency, privacy concerns, and single points of failure. A warehouse robot that loses its internet connection and freezes mid-task is, to put it mildly, not ideal. Nvidia is betting that on-device AI workloads will be a massive growth driver as robotics, autonomous vehicles, and edge computing applications scale up.

The Jetson Thor kit is aimed at developers, which is a classic Nvidia playbook. Get the tools into builders’ hands early, create an ecosystem, and then sell the production hardware when those projects go commercial. It’s the same strategy that made CUDA the dominant framework for GPU computing. Build the moat before anyone else realizes there’s a castle worth defending.

The $5 trillion question

Analyst Dan Ives has estimated that Nvidia could reach a $5T market cap as AI becomes more deeply integrated across industries. That’s a bold number, but it reflects a broader Wall Street consensus that Nvidia isn’t just riding a hype cycle. It’s positioned at the center of a structural shift in how computing resources are allocated globally.

The $40B-plus in 2026 infrastructure commitments isn’t just Nvidia spending its own money. It’s Nvidia co-investing alongside hyperscalers, cloud providers, and enterprise customers who are all racing to build out AI capacity. The company is functioning less like a traditional chipmaker and more like an orchestrator of the entire AI supply chain.

Nvidia is also collaborating with the US Department of Energy’s Genesis Mission, an effort focused on maintaining American leadership in open-source AI and advanced technologies. Government partnerships like this add a geopolitical dimension to Nvidia’s strategy. In a world where AI chip export controls and national AI strategies are becoming standard policy tools, being aligned with federal priorities is both a shield and a growth lever.

For crypto investors specifically, Nvidia’s mainstream AI push has second-order effects worth watching. The company’s GPUs remain central to proof-of-work mining operations, and its data center expansion could influence the economics of AI-adjacent crypto projects, particularly those focused on decentralized compute networks. If Nvidia succeeds in making AI ubiquitous, the demand for decentralized alternatives to centralized AI infrastructure could either accelerate or face existential pressure from the sheer scale of what Nvidia and its partners are building.

The risk, of course, is execution. Committing $40B-plus in a single year to infrastructure means Nvidia is betting that demand won’t plateau. If enterprise AI adoption hits regulatory headwinds, if the economics of large language models shift unfavorably, or if a competitor delivers comparable performance at lower cost, those investments become very expensive very fast. Nvidia has navigated boom-and-bust cycles before, notably in crypto mining hardware. Whether this AI cycle proves more durable is the question that separates a $5T valuation from a painful correction.