U.S. senator and frequent Trump critic Bernie Sanders has come out in support of his government taking a stake in chipmaker Intel in exchange for CHIPS Act grants. TaiwanPlus spoke to Stacy Rasgon, Managing Director and Senior Analyst at Bernstein Research, about the prospects for the deal, whether it's enough to fix Intel's problems and what the beleaguered chipmaker really needs.
00:00U.S. Senator Bernie Sanders, normally a Trump administration critic, has come out in favor of the White House plan to take about a 10% stake in the U.S. chipmaker Intel in exchange for CHIPS Act grants.
00:13That would seem to give it at least some bipartisan cover. Do you think this makes the plan materially more likely in the near term?
00:20To be honest, I doubt that anything Bernie Sanders has to say is going to influence this administration one way or the other.
00:26It is kind of interesting, though, because the CHIPS Act itself was quite bipartisan.
00:31It was actually one of the very few true pieces of bipartisan legislation that actually made it through.
00:37And there's a fair amount of grants that were awarded in quite a few states.
00:44So it actually makes this whole thing very interesting, you know, because if the Trump administration is saying they want to swap grants for equity now, I don't even know, frankly, if they're allowed to do that.
00:54I mean, the CHIPS Act, it was an act of Congress.
00:59So I don't know what the Trump administration actually can do.
01:01Although, look, I'd say, you know, they don't always care, right?
01:05I mean, they typically will go and, you know, they'll try to do whatever they can do and, I guess, see if anybody tries to stop them.
01:13Even if the deal goes through, do you think the 10.9 billion U.S. dollar injection would be enough to fix the problems that Intel is facing?
01:22I don't think so.
01:23It's not big enough.
01:25I mean, so, again, just for some context, the Ohio project that they delayed, I mean, that was Pat Gelsinger's baby, the old CEO.
01:32And now they've delayed that, you know, 2030 or even beyond.
01:35They don't need it.
01:36They don't need the capacity.
01:37That project was originally spec'd at $28 billion.
01:42So, I mean, the CHIPS Act money was, again, that $10.9 billion.
01:45That was actually supposed to go across, I don't know, three or four different locations, Oregon, New Mexico, Arizona, and Ohio.
01:53This $10 billion or $11 billion or whatever it is, it isn't enough.
01:57They need more.
01:58If you're Intel CEO Lipu Tan right now, do you take this deal?
02:03Well, again, Lipu may or may or may not have a choice, right?
02:06But it's still, you know, we should talk about what they need.
02:09They need more than money.
02:10So they need three things, in my opinion.
02:12Number one, they do need money, sure.
02:15Number two, they need customers.
02:17And, like, this is one of the other.
02:18Let's say they build out Ohio.
02:19Let's say they have enough.
02:20Let's say they hand them $28 billion and they go build out Ohio.
02:22Great.
02:23What are they going to fill it with?
02:25Right?
02:26If you don't have customers, you're just digging a hole and throwing $28 billion into it and then setting it on fire, right?
02:32I mean, customers.
02:33And some folks have speculated that maybe the Trump administration would, like, force or encourage customers to use Intel and help to get the business.
02:42The problem with that is that they need a third thing, capability.
02:47And this is something that they have not yet proven that they have.
02:50And my own view is you don't need to force customers to use Intel.
02:53If Intel can prove that they can deliver products at these advanced manufacturing nodes in spec at a good cost structure and on time, they will have customers lined up around the block to use it.
03:06If they cannot prove that they can do those things, and those are table stakes for running a foundry business.
03:11If they can't prove that they can do that, nobody's going to put any material volume there.
03:16How could you bet your own roadmap on Intel's ability to execute if you're not convinced that they can?
Be the first to comment