- Nvidia's data centre revenue rose 18% quarter-on-quarter to $8.9 billion
- Compute demand per AI agent is 15x to 100x higher than human usage
- Future AI agents will outnumber employees by millions and run continuously
Nvidia's founder and chief executive Jensen Huang is rarely short of a quote, and the company's August 26 earnings call was no exception. Behind the headline numbers, a $96 billion quarter and a 70% revenue growth guide for fiscal 2028, Huang used his answers to analysts to sketch out a broader thesis on where AI is headed, from the compute cost of running autonomous agents to why he thinks the industry has already, in his words, achieved AGI.
The comments came as Nvidia reported that data centre revenue rose 18% quarter-on-quarter to $89 billion, with executives repeatedly stressing that the company's growth is now limited by supply, not by demand. It is against that backdrop, of a company trying to build fast enough to keep up with the AI buildout, that Huang's remarks offer a window into how Nvidia is thinking about the next phase of the AI cycle.
Here are five things he said that stood out.
1. Compute demand per AI agent dwarfs human usage
"The amount of compute necessary for an agent versus a human using it is probably 15x to 100x, depending on the type of problem you're trying to solve."
2. Agents will soon outnumber employees by orders of magnitude
"We have 40,000 employees roughly. In the future, we'll have 400,000 agents, 4 million agents, and those agents are running continuously."
ALSO READ: Nvidia Says Memory Pricing Beyond 'Expectations'
3. AGI is already here — the label just doesn't matter anymore
"In a lot of ways, for many tasks, we could say that we've already achieved AGI. I think all of those milestones... are kind of senseless at this point."
4. Open and closed models are both winning, and both run on Nvidia
"Most — I would say nearly all open models run on NVIDIA... Both closed models and open models are skyrocketing in use."
5. Even AI is teaching itself to get better, in real time
"Every time you run through an agent, it reflects on how it could do a better job next time, and it updates the skill file."
ALSO READ: Amazon To Buy Two Million Nvidia Chips For Data Center Build-Out
Essential Business Intelligence, Sharp Market Insights, Practical Personal Finance Advice, Daily Fuel, Gold and Silver Prices and Latest Stories — On NDTV Profit.