Intel’s CEO on Turning Skeptics Into Believers

Amid CEO Pat Gelsinger’s ambitious effort to restore the company to its former glory, Intel has found itself caught up in global forces: geopolitical competition with China, the AI boom, and President Joe Biden’s efforts to reshore manufacturing. Gelsinger is positioning Intel as a national champion and it has been rewarded accordingly—the company has received more CHIPS Act funding than any other. In March, Biden unveiled nearly $20 billion in grants and loans for Intel's domestic chip-making plants. 

Beyond the U.S., Intel struck a deal last June to build a government-subsidized fabrication plant in Germany, and it’s reportedly in talks about building a facility in Ireland. Judging by the company’s share price, investors are bearish on Intel’s prospects, but Gelsinger welcomes their skepticism. “They cause us to wake up just a little bit more energetic each day, because we're going to prove them wrong.”

Gelsinger spoke with TIME on May 3 about Intel’s turnaround strategy, and its role as a U.S. chipmaking champion.

This interview has been condensed and edited for clarity.

You joined Intel as a quality-control technician aged 18 in 1979 and worked there until 2009. Between leaving in 2009 and rejoining as CEO in 2021, Intel went from being arguably the world’s leading semiconductor company to being in a state of rapid decline. How did it feel to watch from the outside?

It was painful. First, it was painful to leave the company, as I was, I'll say, nudged out the door at that period of time. It took me years to get over it. I joke that I used to do the Intel bong when I went to sleep at night. To watch its decline and some of the decisions that were made, it was hard, it was emotional.

When you re-joined Intel as CEO over three years ago, you came bearing an ambitious strategy that you asked the company’s board of directors to agree to as a condition for your taking the role. One of the key pillars of that strategy is to retake the lead in terms of manufacturing the most advanced chips from your main manufacturing competitor, TSMC, by 2025. As I understand it, things are going well so far, but you’re not there yet. Is catching up with TSMC by 2025 do-or-die for Intel?

This is a really important piece of the overall strategy—get back to process technology, competitiveness, and leadership. We see the end of that [being] what we call our “five nodes in four year” journey. Every day, our confidence builds in it. When we laid that strategy out three years ago, people were like: “Wow, that's so audacious, how are you ever going to pull that off?” Here we are three years into it, and we see the end of making all of those things that we said come true. So our confidence has grown substantially in our ability to make it happen. 

You seem pretty confident that you can continue to turn things around at Intel. But the markets don’t seem to share your confidence—your shares fell 31% in April, their worst month in more than 20 years, after disappointing quarterly results. Is there something that you know that the markets don’t? 

Well clearly, we are on a multi-year journey, and stock markets are generally looking at what I like to say is 90-day shot clocks: what's your next quarter going to look like? A very near term basis. We've given now tremendous transparency to [Wall Street] of what the foundry business costs are, and what the losses are, and how we're going to bring that back to profitability. If you're looking and judging us on a 90-day or near term, financial return, hey, I understand the skepticism. 

At the same time. If it takes five years to build a factory, launch a new technology, and bring it online, you can't measure that on a 90-day basis. This is a multi-year journey. That's what I asked the board of directors to sign up for. That's what I signed up for. That's why we said five nodes in four years. It takes that long to rebuild this process. We're very confident. I remain absolutely committed that we are on this journey to make that happen.

Could it not be that investors are more skeptical than you are of your ability to execute this turnaround strategy, which is very ambitious?

I do admit there is skepticism from some of the investors. Some of them, when they looked at us, thought: “Wow, this is a long term strategy. It's an expensive long term strategy, it has risks associated with it as well.” We accept that those are the case. 

But here we are three years in. I'm much more confident today than I was when I launched that strategy three years ago. Because we see the evidence—the wafers are popping out of fab [fabrication plants], we're proving the transistors work, we're seeing the advanced packaging leadership position, we're seeing the AI products start to come to life as well. 

So I respect the critics. But at the same time, they don't deter me. They embolden us. They challenge us. They cause us to wake up just a little bit more energetic each day, because we're going to prove them wrong.

You’ve said before that there will need to be another CHIPS Act to bring semiconductor manufacturing back to the U.S. If Congress decides against this, is Intel’s turnaround doomed?

This was an industry that was lost over three to four decades. It wasn't like we just needed one shot in the arm to fix 30 to 40 years of bad industrial policy. This was very explicit on the part of the Asian countries. There was never a vote in Congress to get rid of this industry. There were votes throughout Asia, to attract this industry.

I do believe there needs to be sustainable policy adjustments if we're going to finish the job. Getting the fabs built here is great. But we also have to have the supply chains—both the supply chains coming into the fabs, plus the supply chains that come out of the fabs into the systems that are built around those pieces of silicon. We also believe there needs to be sustainable tax policy that encourages these long term capital investments, there needs to be long term R&D policies. That's the seed corn of the industry. We need long term talent as well, as we rebuild the industry. 

So for us, it's less about the grant programme for the long term, and establishing the entire ecosystem that's required to have a healthy semiconductor ecosystem. Now, when we measure the CHIPS Act today, it's like: “Huh, this is pretty good.” Because we've seen projects from Intel and Micron and TSMC, and Samsung, and others across the industry starting to return to the U.S. It is having that domino effect. 

But, as I've always said, if we're ever going to go before Congress for further actions required to support this industry, the best way to do that is to have success from CHIPS One. And the success patterns that we've already demonstrated, you're clearly seeing the economics, the job creation, and the migration of the industry is already well underway. That's the best thing—that we can go before Congress and say: “Look at the momentum that we've already achieved.” Today we're somewhere in the order of 10% of leading edge semiconductors in the US—a hideous state. And all the projections in the announced projects today, say that would double by the end of the decade, going from a decline to doubling in a five to six year period. That's pretty stunning.

Intel has often positioned itself as a sort of national semiconductor manufacturing champion. Many geopolitically important companies have very patriotic internal cultures. Would you say that's the case at Intel?

Intel is somewhat, I'll say, a patriotic company. It was founded that way by Grove, Moore and Noyce. So there's a strong lineage there of the benefits of U.S. Andy Grove, personally, was a survivor coming out of the World War Two period—a Hungarian refugee coming into New York. Bob Noyce was the first Sematech, rebuilding in the Japanese-U.S. battle. So I'd say there's a long history here. 

At the same time, we're very much a global company. I have large businesses across the world that are globally distributed. 

Finally, the vast majority of our R&D is right here in the U.S. When we consider Silicon Valley, Intel is the company that puts silicon into Silicon Valley. We believe in this wonderful, chaotic, long term R&D, bubbling cauldron of innovation that occurs here. And we view our role in uniquely helping to enable that. Obviously, we have R&D in all locations around the world. But the core of the company, it has always been here in the U.S., with the preponderance of our R&D here in the U.S.

With some of the CHIPS promotion activities often the refrain was something like: “If the CHIPS Act gets passed, we can build our fabs in the U.S., but otherwise, I'm just going to have to build them in East Asia.” From the outside, it sometimes appears like Intel positioning itself as a national champion is slightly opportunistic. After all, Intel is just a publicly traded company, and is therefore meant to serve its shareholders, not the national interest.

Let's put this a bit more carefully in context. Lots of studies have shown—and there's studies by the SIA, studies by BCG—that prior to the CHIPS Act, it was on the order of 30% to 40% more expensive to build manufacturing facilities in the U.S. Now, if I'm supporting my shareholders and I'm going to make a $30 billion new factory investment, that's 30% to 40% uncompetitive. How do I compete? It's simply impossible to compete. 

The entire thesis of the CHIPS Act was: rebalance that gap, close that gap, so investments in the U.S. were not so punitively affected by decades of shift of this industry to Asia—decades of industrial policy in Asia. Level the playing field so I could continue to invest in the U.S. with R&D and manufacturing. That's what the CHIPS Act was out to fix, and we're very proud of that legislation. 

And then obviously, as we've seen, the Samsung and TSMC and Micron and Intel announcements, a major factory. It's worked. We do see those factory locations returning to the U.S., and it complements the R&D that we've always had here.

I spent a significant amount of my time helping to lobby to bring this about, because we didn't want to be competitively penalized to continue the long heritage that we had in the U.S. I'm proud of this. And we do think that the effects are going to be meaningful for our business long term. But more importantly, for the industry long term.

We have created supply chains that are simply precarious. Geopolitics has been defined by where the oil reserves have been for the last five decades. Where the technology supply chains are is more important than that for the next five decades. Let's not create a precarious situation. Let's start building resilience in the supply chain and U.S. and in Europe. That's what our strategy has always been since we began this journey, and we're proud of the progress that we've got to participate in helping to bring about.

How do you think about the impact that AI will have on the world? 

I think of it on par with what the internet did 20 plus years ago. It is that transformative. Where we connected every human on the planet, drove a cycle of innovation, of new users, new use cases. We think of AI very similarly. 

As we think about that, we've described our strategy as AI everywhere. That it will be in your personal computer—and we launched the AI PC late last year. We expect to ship more than 40 million of those this year, as your personal computer in the past was defined by the QWERTY keyboard—that great innovation of 1886 still defines your user interface. AI is going to change that. We're going to drive that into existence, the AI edge, as you think about edge use cases, retail locations get smarter, supply chain locations become more intelligent, healthcare becomes better as we go forward. 

Clearly we see, and we just launched our open platform for enterprise AI, where almost all corporations today have not yet deployed major AI use cases internally. We think of this as transformational for business. I just demonstrated how we're using AI in our factories, to drive yields and productivity of our semiconductor factories. And obviously, the training environment has been super incredible. What's been able to achieve with the high end training markets. But what's more exciting to us is—how do we deploy those models. There's a lot of exuberance in creating the models, but it's sort of like weather models. Not many people create them, lots of people use them. We're very much the company that wants to drive those use cases. 

But finally, our strategy is also to be the semiconductor supplier for all of the other chips that get built. We're proud of the chips that we're going to build for AI. But we also want to be the foundry, the factory, the manufacturer of everybody else's chips as well. And in that I say that we get to participate in 100% of the AI market, both through our products, as well as through our foundry.