Nvidia CEO Jensen Huang says tokens are now profitable for AI companies, and he’s not talking about crypto

Jensen Huang just dropped a line that, stripped of context, could send crypto Twitter into a frenzy. During Nvidia’s fiscal Q1 2027 earnings call, the CEO declared that “tokens are now profitable” for AI companies. Before anyone starts minting celebratory NFTs, here’s the thing: he’s talking about AI output tokens, not blockchain tokens.

The distinction matters enormously. In AI, a “token” is a unit of model output, a chunk of text, code, or reasoning that a large language model generates. Huang’s point is that producing these outputs has crossed a profitability threshold, meaning AI companies can now charge more for their model outputs than it costs to generate them.

Nvidia’s numbers tell the real story

Nvidia reported quarterly revenue of $81.6 billion for Q1 FY2027, an 85% increase compared to the same period a year earlier. The data center segment, which houses the GPU infrastructure powering most of the world’s AI workloads, brought in $75.2 billion. That’s a 92% year-over-year jump.

Huang described the current demand environment as having “gone parabolic.” His exact framing: “Tokens are now profitable. So model makers are in a race to produce more.” In other words, the economics have flipped. AI inference, the process of running trained models to produce useful outputs, is no longer just a cost of doing business. It’s a revenue engine.

What Huang means by ‘token budgets’

This isn’t the first time Huang has floated the concept of AI tokens as an economic unit. At Nvidia’s GTC event in March 2026, he suggested that companies could allocate AI token budgets to their engineers, potentially worth about half of an engineer’s salary. For someone earning $500K, that’s a $250K annual budget for AI-generated assistance.

Huang has framed this as part of a broader shift toward what he calls the “AI factory” economy. In this model, compute capacity itself becomes a form of revenue generation. Data centers aren’t just infrastructure costs on a balance sheet. They’re factories producing valuable output measured in tokens.

He also declared during the earnings call that “agentic AI has arrived,” signaling that AI models have moved beyond chatbot novelty into territory where they can produce actionable, revenue-generating work.

Why crypto investors should read the fine print

At no point during the earnings call or at the GTC event did Huang reference blockchain technology, cryptocurrency, or digital assets. The tokens he’s discussing are purely computational units. They live on GPU clusters, not on-chain.

For investors watching both sectors, the more useful takeaway is structural. Nvidia’s dominance in AI infrastructure creates a gravitational pull that affects every adjacent market. Decentralized GPU networks, which often position themselves as alternatives to centralized cloud providers, now face a competitive landscape where centralized AI is demonstrably profitable. If token generation is a money-printing operation for companies with access to Nvidia’s latest hardware, the value proposition for decentralized alternatives needs to be about more than just cost savings.

For crypto-native investors tempted to trade on the “tokens are profitable” headline, the move isn’t to ape into AI-themed altcoins. It’s to understand that the AI economy is developing its own token economics, entirely separate from blockchain, and that the infrastructure layer is where the real money is being made right now.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.