Jensen Huang said it plainly at Nvidia's GTC keynote.
Running the software that runs agents on an AI factory costs about $2,000 per GPU per year.
That figure sits next to one that already exists.
Nvidia's AI Enterprise subscription lists at $4,500 per GPU for a single year of coverage.
Put both on a 16-GPU cluster and the arithmetic turns awkward.
$32,000 a year at the new number, $72,000 at the enterprise rate, and the cards were bought and paid for before either invoice arrived.
Huang sized the upside too, with a proposed scenario of five million paid licenses reaching roughly $10 billion a year.
Nvidia has not confirmed which products the $2,000 covers, so read it as a direction of travel rather than a price cut.
For a small team on a fixed runway, that line item is now a hiring decision.
One more renewal, or two more months of experiments.
This is where raw compute changes the math.
RTX 3090 capacity at $0.15/hr and storage at $2.49/TB, with no per-GPU license stacked onto the hourly rate.
Buy Now Pay Later, zero upfront, pay after one month, so a team starts on day one and settles the bill at the end of it.
Where does the software line sit on your GPU budget review, above or below the hardware line?
https://ai.bhkcloud.com/?utm_source=linkedin&utm_medium=social&utm_campaign=bhk_social_2026w44&utm_content=morning
Originally posted on linkedin_personal