Former Intel CEO Pat Gelsinger: 'Of course' we're in an AI bubble

Government Investment in Semiconductor Manufacturing [0:34]

• The primary metric for evaluating the U.S. government's stake in Intel is whether it leads to the building and filling of more Intel fabs in the U.S.
• If more fabs are built and filled, it is considered good; otherwise, it is not.
• Market response to the announcement of government equity was slightly positive, but past equity investments have not included commitments to fill fabs.
• The passage of the bipartisan Chips Act was a positive development, but the slow deployment of funds by the Department of Commerce was disappointing.
• The government's intervention to rethink the program by taking equity signifies a readiness to address issues and reconceive strategies.

To me, the only metric that matters is does it cause the building and filling of Intel fabs? And if it causes more fabs to be built, more fabs to be filled in the U.S., then it's good. If that doesn't occur, then it's not right.

Past Grant Program and Execution Concerns [2:05]

• The Chips Act began under a previous administration and was structured as a grant program.
• The speaker was disappointed by the significant time it took for the Department of Commerce to disburse the funds to help Intel.
• The delay of two and a half years meant that critical progress did not occur.
• The speaker views the initial grant approach as a good deal if it drives the industry to build and fill fabs and reshore supply chains for leading-edge R&D and technology.

I was super disappointed, you know, that uh the GINA and Department of Commerce took so long to get the money deployed, so long to get it across the line to help Intel.

Company Stock Performance and Industry Shifts [3:46]

• While the company's shares are up 54% over the last year, they are still down more than 33% over five years.
• The shift of the industry to AI and the lengthy, expensive process of rebuilding the technology base for the company (estimated at five-plus years) are the two major factors contributing to this situation.
• The company experienced a loss of technical leadership and was not led by technologists for many years.
• Rebuilding core technology, manufacturing supply chains, and process technology has been a long journey.
• The company was late on AI, a consequence of long-standing decisions made over 15 years that put it in a difficult position.
• The recent milestone of 18A is seen as important for the nation, emphasizing the need for leading R&D and manufacturing at scale in the U.S. for supply chains.

Clearly, you know, two things happened. One is, you know, it's a long, heavy, and expensive, and I said it was five plus years to rebuild, you know, the technology base for Intel. We haven't even gotten there yet.

The AI Market: Bubble and Long-Term Outlook [5:30]

• The current situation in AI is described as a bubble, with a lot of leverage and cash in the system.
• Despite the bubble characteristics, the speaker does not see the AI boom ending for several years.
• Businesses have not yet begun to materially benefit from AI, and the current phase involves displacing existing internet and service provider industries.
• There is a long way to go in realizing the full potential of AI.
• Disruptive technologies are on the horizon, promising significant improvements in power and performance, with materialization expected in the latter part of the decade.

And you know, are we in an AI bubble? Of course. Right. You know, we were in that bubble. You know, of course we are. I mean, you know, we're hyped. We're accelerating. We're putting enormous leverage uh into the system. That said, I don't see it ending for several years.