Jensen Huang Says AI Has Shifted From Unprofitable to Highly Profitable in Six Months

Jensen Huang says the economics of AI have changed dramatically in just six months.
According to the $NVDA CEO, AI has moved from an experimental and often unprofitable technology into something increasingly productive and profitable across real-world workloads.
That shift helps explain why AI labs continue demanding more compute.
The basic flywheel is straightforward.
More compute allows models to generate more tokens, and if demand for those tokens keeps rising, higher inference volume can translate into greater revenue and profits.
That creates an incentive for AI companies to keep expanding GPU infrastructure rather than slowing capital spending.
Huang also argues that future IPOs could provide AI labs with significantly more capital to accelerate that buildout.
The two most visible AI labs may only represent the front edge of a much broader wave of companies competing for compute.
For $NVDA, that is the key long-term thesis.
If AI workloads are becoming economically productive rather than remaining purely experimental, demand for GPUs, networking and data-center infrastructure could be supported by actual customer returns rather than speculation alone.