Talking to the assembled throngs at Code Conference, Vole’s chief technology officer Kevin Scott said that procuring Nvidia's compute GPUs for artificial intelligence and high-performance computing applications is less challenging than it was a few months ago.
"Demand was far exceeding the supply of GPU capacity that the whole ecosystem could produce. That is resolving. It is still tight, but it is getting better every week, and we have got more good news ahead of us than bad on that front, which is great. It is easier now than when we talked last time," Scott said.
Nvidia experienced a demand spike when Microsoft began integrating generative AI into its offerings. Such services primarily hinged on Nvidia's compute GPUs to train large language models and run inference, leading to a noticeable supply crunch. Microsoft and its counterparts' reliance on these chips led Nvidia to see a stock price increase of 190 per cent in 2023.
Kevin Scott, who manages GPU allocations at Microsoft, described the role as tricky over the past few quarters. However, he observed that the availability of compute GPUs is steadily getting better.
Scott did not comment on rumours about Microsoft building its own custom silicon for AI workloads. Yet, he accentuated the company's significant endeavours in the silicon domain. For the past several years, Nvidia has remained the top choice for Microsoft's needs in the AI sector, the company's CTO said.
"I am not confirming anything, but I will say that we have got a pretty substantial silicon investment that we have had for years," Scott said. "And the thing that we will do is we will make sure that we are making the best choices for how we build these systems, using whatever options we have available. And the best option that’s been available during the last handful of years has been Nvidia."