Dara Khosrowshahi on why Uber will stop hiring engineers and instead buy agents and GPUs.

“Maybe 5 years from as the engineers get more and more productive, I may not decide to add engineering headcount because at that point, instead of adding an engineer, I should add agents and buy some more GPUs from Nvidia.”

Micron and Nvidia together accounted for 45% of total S&P 500 earnings growth in the measured period.

“45% of the entire S&P 500's growth was from Micron and Nvidia.”

Jensen Huang on Nvidia's cost advantage over free competitor chips.

“Our competitors building A6 they could literally all their chips are cheaper already today but they could literally price them at zero our objective is they could price them at zero and you would still buy an Nvidia system because the total cost of operating that system power data center land etc the intelligence out is still a better bet than buying a chip even if it's given to you for free.”

Nvidia product specs for B300 and beyond show FP4 delivering only 3x the throughput of FP8, short of the 4x that quadratic scaling of matrix multiply hardware would predict.

“Nvidia's product specs have started acknowledging that in B300 and beyond, where the FP4 is three times faster than the FP8. Though it should be 4x.”

Jensen Huang contends that even at zero price, competitor AI chips would lose to Nvidia on total cost of ownership when factoring in power, data center, and land expenses.

“Our competitors building A6 they could literally all their chips are cheaper already today but they could literally price them at zero our objective is they could price them at zero and you would still buy an Nvidia system because the total cost of operating that system power data center land etc the intelligence out is still a better bet than buying a chip even if it's given to you for free.”

Nvidia: Uber's CEO signals the company may shift capital spend from engineering headcount to GPU purchases as AI agents take over coding work, making Nvidia a direct beneficiary of enterprise AI adoption.

“Maybe 5 years from as the engineers get more and more productive, I may not decide to add engineering headcount because at that point, instead of adding an engineer, I should add agents and buy some more GPUs from Nvidia.”

Halving bit precision should yield more than double the FLOP count due to quadratic scaling in matrix multiply hardware, meaning the historical 2x rule understates the true gain.

“Historically, up until B100 or B200, every time you halved the bit precision, you doubled the FLOP count. For the reason you said, because of this quadratic scaling, that ratio is actually slightly wrong. You should get an even bigger speedup than you might otherwise think.”

Nvidia's Blackwell delivers 30x performance over Hopper, making a 15% gross margin discount from competitors insufficient to close the gap.

“If somebody gave them a 15% discount, you know, the difference between our gross margins, which is called the 75 points, and somebody else's gross margins, call it the 50 to 65 points, is not so much as to make up for the 30 times difference between Black Wall and Hopper.”

Jensen Huang explains that the end of Moore's Law (flat transistor costs) forces a shift to accelerated computing and algorithmic efficiency as the new drivers of performance improvement.

“Mo's law is dead, right? Because transistors basically cost the same every single year now.”

Jensen Huang told Brad that Nvidia plans to triple in size with little to no headcount growth, signaling AI-driven operating leverage even at the hardware layer.

“Brad, I'm going to 3x the company and our headcount may not grow or only grow a little bit.”