Intel’s Gelsinger, OpenAI’s Altman Augur the Future of GenAI
SAN JOSE, Calif. – Intel CEO Pat Gelsinger was rather direct — but in a gracious way — with his special onstage guest, OpenAI CEO Sam Altman, at the Intel Foundry Direct Connect event here on Feb. 21.
“So what’s this $7 trillion thing about?” he asked, immediately after introducing Altman. There were giggles from the audience.
“I mean, anybody can report anything,” Altman said, somewhat in jest, referring to the way he sees the media (in this case, the Wall Street Journal) report the news. More giggles.
Altman was quoted earlier this month as saying the IT industry would need huge investments — perhaps as much as $7 trillion over several years — to produce the amount of power, processors, servers and data centers required to satisfy the world’s vastly growing thirst for new GenAI hardware and software. He didn’t qualify the actual figure.
“No, I guess the kernel of truth is that we think the world is going to need a lot more AI compute. And it’s not just semiconductors. That’s the power that it’ll take data centers to house these chips,” he said.
Gelsinger and Altman were alone on stage as the last act of the daylong Intel Foundry launch event at the McEnery Convention Center. A crowd of about 1,000 analysts, journalists, Intel partners, customers and potential customers were on hand to learn about Intel Foundry and to witness a couple of major-league IT influencers have their conversation.
“This is going to require a global investment of a lot of stuff. We are thinking about just chips today, but the total cost of the infrastructure will be high. How high? I don’t think we know yet. But it’ll be $7 trillion,” Altman said.
Many companies are going to need this technology, but mostly this is about so many people who want to use its services and can benefit from being able to run AI apps right now, Altman said.
‘Severe Compute Crunch’
“There’s obviously a severe compute crunch, and as the models get bigger and more powerful if they’re able to be used for more things, we’d like to make sure that becomes way more accessible, way more abundant, and way more affordable to everybody, and not what I think could happen — which is that it goes very much the other direction and gets to be so expensive that only the richest people can afford it. That’d be really bad,” Altman said.
Should governments be concerned about domestic AI chips and AI technology capabilities? Gelsinger asked.
“Obviously, I think the answer to that is yes and yes,” Altman said. “Given what I expect for the geopolitical importance of AI in the coming decade and beyond, I think a resilient supply chain and U.S. capability is critically important. So I’m happy the U.S. government is doing this. I think it’s great.”
“Do you see countries viewing this as a strategic differentiator? How much of this is going to be globally available models versus becoming driven by different countries or industries?” Gelsinger queried.
Altman said he believes that “different governments are going to take different approaches, and we’ll get to see how different things work. I think there are a lot of ideas, and people are starting to try stuff.
“There’s obviously a huge amount of understandable anxiety,” Altman said. “But exactly what people are going to do and what the right approach is going to be … ah, I’m just happy I’m having the conversation now while the experiments, regulatory ideas and the stakes are relatively low, and while the model is relatively weak.”
Altman is all for democracy in developing GenAI models and subsequent applications and hardware to take advantage of it.
“A big part of our strategy is what we call iterative deployment. What you shouldn’t want would be to build AI in secret in the basement, and then push a button and drop it on the world all at once,” Altman said.
“What you want instead is like, after a very embarrassing GPT4, a slightly less embarrassing GPT5, is a pretty good GPT6 up to GPT7. You want time to have society and our institutions gradually evolve and think about and try some different approaches for how to manage this technology in this socio-economic change. So I think it’s great. And it’s happening now.”
As for what the long-term future holds for the nascent GenAI business and for OpenAI, Altman was particularly optimistic.
“I think it’s really good news for our collective adaptability,” he said. “People update very quickly to new technology, people increase their expectations for what the world should deliver for them very quickly. So I think that’s great. We will continue to deliver, and I think people will continue to get used to it. And what people will be capable of achieving with these tools in another five or 10 years will be quite remarkable.
“It’s not even if people will just be able to do more work faster; we’ll be able to do things that we just couldn’t do before. What we just weren’t smart enough to do before on our own — and what these new tools will enable — is gonna be great.”
