Ted Audio Collective Welcome everyone.
You're listening to After Hours.
I'm Felix. I'm Mi -Hyr.
And I'm Charlotte from The Economist.
Welcome back Charlotte.
It's great to have you.
Thanks for having me on again.
And you brought the topic for today.
So I want to talk about chips.
When business people, politicians, when they talk about critical supply chains, semiconductors are probably the good that is deemed most critical.
And people used to be concerned about access to oil and now governments are worried about a whole range of goods and technologies.
And it's challenging for the governments and also for the companies involved.
So that's what I'm hoping to talk about.
Fantastic. I got to tell you, Charlotte, before we started taping, I saw my 11 -year -old, and she asked me what we were talking about.
And I said, chips. And she said, what kind of chips?
And I said, guess. And she said, chocolate chips?
And I was like, no.
Peeta chips? No. She guessed Takis, which I was quite proud of.
And then finally she said, computer chips.
And I said, yes. Oh, good for her.
Yeah. She literally said, that is so boring.
It's not boring. I'm here to prove her wrong.
I'm delighted. Please prove her wrong.
You know, one thing that is funny is we keep changing our minds.
First data was the new oil.
Now chips is the new oil.
And oil is now the new oil.
Yes, maybe. Did you bring a topic, me here also?
I did. So there's been a lot of reports coming out recently from companies, earnings, but the ones that really struck me were in ride -sharing and mobility.
So Lyft lost a third of its value.
Uber reported pretty large losses.
I want to just check in on ride -sharing and these two -sided platform business and what its future is.
Wonderful. That sounds interesting.
By the way, Charlotte, I have to confess.
One of the reasons I wanted to do that is anytime I have any thoughts about two -sided platforms, I desperately want to check in with Felix.
I basically want to sit at the knee of my two -sided platform, Guruji, who is, by the way, Felix Oberholzerji.
Oh God, I'm in trouble already.
So Charlotte, chips.
Yeah. So I think it's worthwhile just to start by explaining what a chip is.
People talk about them all the time, but just as a brief reminder.
But a chip is basically a silicon wafer, right?
And it has tiny electric circuits scratched into it.
And those circuits carry electrons that do calculations for all kinds of things that we take for granted.
The chip is kind of the essential tool for many, many other machines.
And different governments see it as a really big strategic industry.
China was trying to build up a semiconductor industry, really pouring a lot of money and support into this sector.
And America, even before the pandemic, was concerned about this industry that it viewed as strategic.
And then COVID hit.
And there was this huge demand in electronics that were dependent on chips, and that coincided with supply chain problems.
And the result is countries falling over themselves to think about how to secure access to these chips.
And that raises all kinds of questions, both what government policy should be, but also how companies decide how to allocate capital in this environment.
I see two things that seem particularly important to think about.
As you point out, there's huge demand for these chips, because they're basically everywhere.
Every product almost seems to have a chip.
And as the supply chain gets strained, we actually discover even more products than I ever knew, for some reason, contained chips.
So there are two changes, I think, that are really interesting.
The first one is that we get many new plans.
So in the next two years, we think something like 90 new factories will come online.
And that just reflects the importance of the industry and overall demand.
We have, for some reason, decided that we need to subsidize these plans.
So both the EU and the United States now have plans, roughly $50 billion each, to subsidize particular companies.
The second and maybe even more important question is around limiting export opportunities, both in terms of advanced chips.
So there are many restrictions in place, the kinds of chips that can be exported, the kinds of chips that can be used, and then maybe even more important restrictions around the equipment makers.
Because while it's true that many of the semiconductor manufacturers, they have formidable market positions, I think the most amazing source of market power is actually in the companies that make the tools, that then allow Intel and Samsung to make the chips in the first place.
So these are companies, we don't actually talk about them all that much, but in the US, it's companies like Applied Materials, KLA.
The Dutch company ASML is a company that produces lithography machines.
And for many of these companies, I think it's true that there's no substitute.
They're literally monopolies.
There's no one else who can do what they can do.
And so we're thinking about how to constrain their business opportunities, in particular because we're concerned that anywhere from 10 to 30 % of the revenue now comes from China.
So subsidies and restrictions, to my mind, those are the two big questions that we need to think about as we think about this space.
Well, so let me add one more to that, Felix, which is I think you're absolutely right.
There is this sense in which these sources of critical parts of this are highly concentrated, whether it be in Taiwan or the Netherlands.
It just seems very risky in this world to have them be so concentrated.
But I confess all of this feels a little bit like, you know, this time is different.
What do we know about this industry?
We know this industry is a capital goods provider to a pretty volatile industry, and it is boom and bust, and it's prone to that and has been for the last 20 plus years.
And why is that? Well, demand starts to rise, capital spending lags, and then before you know it, you have overcapacity, crashing prices, and it happens again and again.
And now the story this time seems to be this time is different, which I'm always skeptical about.
And this time is different because, well, it's the Internet of Things.
There's going to be chips in the wooden floors.
There's going to be chips in the walls.
There's going to be chips everywhere.
So this time is different.
And this time is different because governments are involved and very worried about having sources that are actually closer by.
My instinct is like usual, which is this time is not different.
And you have the recipe for a bunch of non -economic actors, that is governments making non -economic decisions.
You have this sense of unending consumer demand for chips, which may be true, but if the economy falters, this could all come undone.
So I guess the other interesting part to me is, is this time really different or is this just a recipe for kind of massive overspending?
Yeah, I think both are actually true.
I think there is a recipe for massive overspending.
I think that clearly is what's already underway, frankly.
Also, this time is different because there's rising distrust of globalization.
There's rising distrust in an enormous economic force, i .e.
China, and an interest in, if not completely decoupling, reducing reliance on Chinese supply chains.
And when you have a mistrust of global supply chains, that inherently leads to a decline in efficiency.
And you see this not just in chip making, but all across different sectors, where companies are building redundancies into their supply chains.
And, Trell, the implication of that view, I think, is that ultimately consumer prices are a lot higher for all these products, like significantly higher if we end up with a much less efficient production system.
Either consumers are going to pay the price or investors are going to eat it by having funded a lot of capital that is kind of inefficient.
You see this playing out pretty dramatically in the area of solar panels, where you have, right now, an American company that is trying to restrict and has restricted import of solar panels that are Chinese companies often sourced through different parts of Asia.
So, they're trying to crack down on very low price solar panels.
And the result is that you have an enormous slowdown in installation of clean energy products.
And the cost of those projects is going up.
And so, you have to think about these things in context, where there are knock -on effects that might undermine goals that the administration has and that the country might have, in this case, clean energy, that are in the service of trying to protect an American company.
And I think it raises, at the same time, an important question of how best to do it, even if it's true.
Say, for a moment, we all agree that semiconductors are critical and we somehow need to maintain control.
There's still a really interesting question, what can you do?
What should you allow?
What should you not allow?
Maybe I'll throw out two examples.
There's a Chinese semiconductor company, YMTC.
It is now the leading Chinese company to produce flash memory.
They are in trouble because one of their chips found its way into a Huawei phone.
And Huawei, obviously, is this old enemy.
We decided we don't like the company.
We're using something called the Foreign Direct Product Rule, which essentially says, even if you're not an American company, if you happen to buy equipment from America and then somehow that equipment is used to supply another company that happens, to be on the entity list that the U .S.
government maintains, meaning I cannot buy American products unless there's an export license.
So in an indirect way, YMTC now gets hit for having supplied these chips.
Now, that, to me, is super hard to justify.
What's the reason that we care about whether a Huawei phone, this is not even the telecoms equipment.
This is literally just a cell phone.
If you don't like the idea of having a Chinese cell phone, don't buy a Chinese cell phone.
What's the role of the U .S.
government to restrict Huawei's access to these particular chips?
I think that's a great example of, there's this general sense, it's strategic, it's important, and then it plays out in ways that, to your point, drive up consumer prices, limit consumer choices, and frankly, make no sense.
Even if you're concerned about the security of that supply chain overall.
I'm really glad you brought up Huawei because it was sort of the tip of the spear here.
So the government banned American microchips that Huawei needed for its equipment.
That was 2018, the export of those.
Yes. It did make one thing clear, which is that if you act alone, as the Trump administration did, you don't actually secure your supply chain, you just diminish your control of it because other people start to fill in the gaps.
And you're seeing that right now with some of the machines that you talked about earlier, Felix, that are really, really highly specialized, that are the tools to help make semiconductors, the tools that are used in different factories.
And America has these, China would like them in order to build up its semiconductor industry.
But what you hear these American companies saying to the administration is that if America tries to clamp down on exports of these types of goods, if it acts alone, it doesn't actually depress China's power.
It just depresses American companies market share.
So they are losing business in China and other people might try to fill in the gap.
There's one last piece to this, Charlotte, which I think is really interesting, which is, as you point out, all of these are fairly blunt instruments.
And as you point out, Felix, there's really weird perverse and knock -on effects, right?
But there's one last dimension to this, which is these blunt instruments are used in a space that is actually quite heterogeneous.
So we talk about chips as if it was one thing.
Yeah. In reality, chips are many, many, many things.
So as one example, Felix, you mentioned memory and DRAM.
And so that market is already kind of saturated, is already a little bit overbuilt relative to some other markets where we have tremendous undercapacity.
To get this right in this sector is really, really hard with blunt instruments because there is so much heterogeneity.
And so you have to make the right call on the 9 nanometer versus the 7 nanometer.
I don't even know the variety and the heterogeneity versus the memory chips versus the ones that go in the smartphones, whether the ones that go in the floors.
And it feels like, wow, we're taking a blunt instrument to a highly nuanced space.
And that also makes no sense, just to kind of add to the chaos of this.
No, I totally agree with you.
And the industry changes really quickly and you want to support that level of innovation.
So the government says, oh, we're going to support X.
And the industry is like, you know, actually, we've moved on from X.
We skipped over Y. We're now at Z.
It doesn't line up.
So another way to say that is products become obsolete quickly.
So you end up adding capacity in the wrong category.
Right. And then prices crash and we're into this overspending cycle.
So it just feels like an area which is prone for, well, maybe not just a mess, maybe some big winners, but also some really big losers conceivably.
In part, it's also reflective of, I think, a misunderstanding of how the industry works in the first place.
So take the equipment manufacturers that have these amazing market positions, in particular in congressional testimony.
The idea often is that if you sell the most advanced machines to the Chinese, the Chinese will somehow magically copy them.
And as a result, we should restrict these export opportunities.
But if you look a little under the hood, a company like the Dutch ASML company, for instance, it's more like Boeing than compared to an integrated manufacturer.
So ASML has three, 400 suppliers, mostly in the Netherlands, some in Germany, some in the US.
And they supply something like 1 ,500 components that go into each individual machine.
So even if you somehow had the final machine without having this really complicated supply chain, you will never be able to replicate what ASML has done.
And the idea that we sell one of these machines to the Chinese and then magically the Chinese will copy that machine, that is, I think, ludicrous.
We hurt the companies by restraining their export possibilities without really contributing to the security of the supply chains that we care about.
So, Charlotte, let me ask you, since you brought up this topic, if you go to Congress and they ask you what we should do in this space, would you simply say, don't even try because you're chasing something that is ephemeral and just let private market participants figure this out.
Or would you say, yeah, no, you're right.
We should do something like this.
You know, as we were talking about how idiotic some of these policies are, I was thinking, oh, gosh, how much easier is it to criticize these policies than to devise them?
I don't have sympathy for the idea that you need to throw a lot of money at semiconductor plants domestically.
I think the idea that you protect supply chains by being protectionist is completely and totally misguided and that you need to have really robust agreements and trade relationships with countries that you view as either closely or even not that closely aligned with your broader geopolitical interests.
That's how you actually create a secure supply chain.
What do you think, Felix?
What would you tell them?
I do think it was good to block the Chinese attempt to take over ASML, for instance.
So you remember a couple of years ago, there were in negotiations and early talks about getting control of the company.
I think blocking takeover, blocking ownership of these critical pieces of the supply chain, I think that makes a lot of sense.
Like you, Charlotte, I'm super critical about the government subsidies.
There's no evidence that they're needed at this point.
The industry is profitable.
It's not wildly profitable for all the reasons that me here has pointed out.
When you look at returns and invested capital, nothing suggests that there's a big fundamental problem here that needs government intervention and that needs taxpayer subsidies.
I am mostly critical of limiting the markets of these companies.
Apple is now interested in buying flash memory from the Chinese semiconductor company, YMTC.
And I don't see a reason why we should not allow Apple to do this.
If they look around and if they find, oh my God, here's a Chinese supplier that is now cost competitive, that will force Samsung to double down on its effort to be the leading company in the industry.
That can only be good for consumers.
That can only be good for America.
Even on the equipment side, I'm much less nervous about China catching up there relatively quickly.
One thing I'm taking away from this conversation is there's at least three ways to think about this.
There's domestic capital subsidies.
There's protection on trade in its various forms.
And then there's the market for corporate control, which is kind of like your ASML purchase.
And I think the protectionist tendencies, as you and Charlotte have articulated, make little sense.
I think the domestic capital subsidies make little sense.
Restricting ownership, I'm coming to understand maybe okay, you know, per your argument.
That's different, right?
Maybe that's just different because it's ownership of IP.
These are all big questions and the degree to which you respond with government intervention, how much intervention is necessary if it's necessary at all.
How does the private sector figure out how to invest in really long term projects?
Building a factory is not something you put up and then put down in a year.
Investing in a mine is not something you do in a matter of months.
These are really long term capital decisions.
So it's complex, both for the politics of it and certainly for the decisions around long term capital spending.
And I think this is only just beginning.
So lots more to come.
So Mihir, you wanted to talk about ride sharing.
So there's been some big developments.
So as you'll remember, we had these two big companies go public almost three years ago, Uber and Lyft.
We have these adjacent kind of gig economy companies like DoorDash and Just Eat.
And there was this remarkable pandemic that ensued in the middle, which totally disrupted their business in some ways, providing some opportunities and then causing chaos.
And now we're seeing the return and we're starting to see some of the metrics coming out of it.
And it's a little bit interesting.
So Lyft, as one example, lost a third of its value last week after announcing results that were OK, but struggling with a little bit of driver incentives.
Uber reported a very large loss, although operationally they were doing well.
So I think it's just a really good time to revisit these two sided platforms, understand what they're doing.
It seems now there's this remarkable concern on the supply side, which is in this labor market, relying on independent contractors who are more loosely affiliated with you actually becomes a problem because driver incentives become really costly.
So I'm curious what you think of these business models now that we're five or ten years or more into them.
Is this a transitory phenomena where, yeah, now they got to work out at the supply side and it's a tight labor market.
It'll all be OK. Or is there something more deep that is wrong in these business models?
Well, so the original promise and the original excitement around these business models, I think, was twofold.
So this is your classic two sided market, which many people at that time believed leads to winner take all or winner take most outcomes.
And so, of course, now we know time after time after time this is just not true.
And I think it's exacerbated in the case of ride sharing because you have to capture these markets one after the other.
If I have a really amazing competitive position in Boston, that doesn't really mean anything for my competitive position elsewhere.
The second thing, as you alluded to me here, of course, had to do with this independent contractor idea.
So I don't have employees.
Whenever you feel like driving, whenever you feel like delivering food, you do it for me.
But this is an arm's length relationship, and it meant that I would have a much more favorable cost structure.
And I think even before the pandemic, we saw two issues.
The first issue is an issue around quality.
The moment someone is not an employee, you essentially cannot really tell them how to do their job.
And so in many ways, the quality of these products, the quality of these services suffered because there was very little handholding for legal reasons.
And then the second big problem that I think continues today is what people sometimes call multi -homing, the ability of workers to work for several companies that want them at the same time.
So what's happened in reality, and you see it play out right now, is that this is essentially an ongoing auction for labor.
As a driver, I'm on the Lyft app, I'm on the Uber app, maybe I'm on the DoorDash app at one at the same time.
And I'm basically looking around thinking about who's going to offer me the most money, and the data suggests that the likelihood that if you send out a job that it will be accepted by someone the first time around, that likelihood is actually surprisingly small.
And so what the companies then have to do is they have to up their offers.
They have to say, I'm going to pay you $20, I'm going to pay you $22, $24, $25 an hour for a particular job, and otherwise no one is going to do it.
And so this is sort of the hidden cost of independent contracting.
I want to ask you whether you think it's a temporary phenomenon.
So right now it's clearly a seller's market in terms of the labor force.
So we saw the jobs numbers for April.
They showed continued strong hiring more than most economists were expecting, and a labor force participation rate that dipped.
So do you think that this situation where you have independent contractors really wielding their might, do you think that that's temporary and that when the job market cools, that this will turn out to be a good model again for these companies?
So I think it's, of course, true that the situation is exacerbated because we lack so many workers in the economy.
The labor market is really, really hot right now.
And so that makes it difficult for Uber and Lyft, and that makes it difficult for just about every company in the United States.
But I don't think this is a transitory phenomenon.
And I say that for two reasons.
On the one hand, when you look at workers' incomes during the pandemic, actually having opportunities like delivering food, like delivering meals to individual households, that was a really big opportunity.
So in some sense, it's of course true that some workers quit because they didn't feel comfortable driving others because you never really know who's sick, who's not sick.
But it's also true that this was a really stable employment opportunity that drew many workers during a time when there weren't many other employment opportunities.
So when the companies now say, well, it's all transitory and it's going to go away, I'm not convinced that this idea that gig work is maybe less attractive right now, that this is something that's transitory.
I think the issues around gig work have been issues of it's just really expensive to maintain a workforce the size you need in order to be competitive in these markets.
The other piece of it that's interesting to me, Felix and Charlotte, is now Uber and Lyft look different because Uber allows drivers to make money in different ways.
So they're making a big deal about Uber Eats.
So they're saying that Lyft, you only do ride sharing.
And so drivers have only that piece of the puzzle.
So just imagine on a rainy day, maybe ride sharing goes down, but Uber Eats goes up.
And so for drivers, you're kind of insured with Uber, but you're not with Lyft.
And so I'm curious what you make of that logic and then this tantalizing idea that somebody has put out recently, which is that DoorDash should buy Lyft.
They should buy it because then the drivers will be able to do what they're able to do with Uber, which is basically have multiple different types of work within one platform.
I don't know if that runs counter to your idea, Felix, which is, doesn't that multi homing problem still exist or does it ameliorate that multi homing problem?
Right. I think it's a flawed argument because in fact, I can work for DoorDash and I can work for Lyft and I can work for Uber right now.
So it might solve a corporate problem in the sense that Lyft now has narrower scope.
And if in fact that's a big competitive advantage, you might think about investing in DoorDash or DoorDash buying Lyft, I guess, because it's the more profitable company.
But it doesn't solve anything on the worker side because, say, I mostly drive for Lyft.
Well, if there's downtime, I can always pick up a job from DoorDash.
I don't actually need Uber in order to create these kinds of combinations.
What do you both make of the deal that Uber announced in March that it was going to partner with New York City taxis largely in a bid to get more drivers?
So they announced in March that you could use the app to hail a yellow cab.
I was struck by it both because how antagonistic traditional cab companies have been towards Uber historically, but also just as a reminder that as much as people might want to criticize the terms for drivers in Uber and Lyft.
Let's not forget how deeply dysfunctional traditional taxi companies are.
Really, really, really expensive medallions.
But what do you guys make of that as a strategy of partnering with traditional cab companies to help deal with this worker problem?
Yeah, so it's a super interesting move.
And in a way, from Uber's perspective, it makes perfect sense, right?
All you need is supply of cars.
You don't really care where they come from.
Is it a taxi? Is it an individual?
All of that is roughly the same.
I think that's exactly right, Felix.
And I'm glad you brought this up, Charlotte, because there's an element of all this, which is it's kind of deja vu all over again.
Like in a lot of these settings, we unbundle the package and then we re -bundle it.
Everybody gets a sub -stack thing and then we re -bundle it into magazines.
It feels like deja vu all over again.
But I guess the big question for me is there was this remarkable thing that Uber did, which was loosen up the constraint on capacity that was dominated by these medallions.
So you had restricted supply and you had a lot of rents accruing to the medallion owners in a lot of urban markets.
And so Uber just busted that open.
And that process transferred a lot of value from the owners of medallions effectively to consumers.
But that game is now done.
And then the question becomes, is this a good place to deploy capital?
The question is like, where are these businesses?
Just to be concrete, Uber is at 40 % below its IPO price, Lyft is as well, well off their highs.
And I guess in part, I'm reminded of our conversation about Zoom.
One of the things that struck me about that conversation about Zoom was a company that creates enormous amounts of value but actually doesn't capture that much for investors.
And I'm curious if Uber and Lyft are like that too, ultimately things that change the world, but ultimately not that great a place to deploy capital.
Yeah, so I agree with that.
How much value created for investors?
Not totally clear. And in fact, it was probably always true that the low prices of Uber and Lyft had more to do with investor enthusiasm because they believed in network effects.
They believed in the market power of two -sided platforms, which was not a very smart rationale to begin with.
But what's really important not to forget is that they created a type of employment that wasn't easily accessible before the gig economy companies came along.
I have research with a doctoral student at Harvard Business School that looks at the value created for DoorDash drivers.
And we find that the flexibility that these models create, I can work an hour, I can work two hours, I can decide that I will help around lunchtime and then I will have a second job, or maybe this is the one job that I have.
There is enormous value created by this kind of flexibility.
So for Uber driver, the evidence is that this type of flexibility has almost as much value as driving for an extra day a week.
So it's really important.
And I know this is an old conversation between the two of us here.
Yes, of course, not much value created for investors because it remains a competitive market.
But is the world a better place because these companies exist?
Absolutely. All right.
Recommendations. Felix, are you taking us back to chips?
I am taking us back to chips and in particular to a series of articles that the magazine Bon Appetit published on junk food.
It's a collection of things and they call it junk food redefined.
And one of my favorite recommendations is the best moment to eat chips is after a long night out.
You come home, you're exhausted, you lie in bed and you eat potato chips.
Oh my God, too good to be true.
The salt, the crisp, the texture, absolutely.
Yeah, it's just perfect.
It has a very interesting story about lays and how they became dominant, not only in the U .S.
but basically around the globe.
And I was struck they have 200 different flavors.
In each country, you only see a small selection.
I knew from the article, you go to Greece and they have something like crab flavored potato chips.
I don't know if that's a good idea or not.
So there's lots of things about junk food for lots of people with very different tastes.
I highly recommend the collection of articles.
It's by Bon Appetit junk food redefined.
I'll make sure that my 11 -year -old is all over that, Felix.
She was right on it.
Nicely done. Charlotte, what do you have?
Me here, a few weeks ago, you and I were talking about falafel places and how we have a preference for taim.
Yes. But not the taim on 22nd Street, the taim on Waverly Place.
Yes. Are you even living in New York or if you don't have strong views about the competing branches of a given falafel chain?
So I recommend taim on Waverly Place.
And just a little plug, my colleague John Fazman wrote an excellent food column on junk food and authenticity about how just because something's inauthentic doesn't mean it's not totally delicious.
So I recommend looking that up.
So true. It's the junk food episode.
Who knew? Yeah. Chips and chips.
Well, so I have to get on this bandwagon and just make sure that everyone tries Takis because Takis are the best snack chip ever.
Tell me about this.
I'm so impressed. Oh my God.
What is a Taki? A Taki, especially the Fuego Takis, are these rolled kind of like Cheeto style, slightly crispier, super spicy.
They're called Fuego.
And then the Taki Blues are also fantastic.
But that's the lowbrow version.
The highbrow version is I just want to draw attention to an author who has a new book out.
But whose older book I just recently reread and I loved, who is Carlo Rovelli.
So Carlo Rovelli is a physicist and writes books largely about physics.
And he's like a serious physics scholar.
But he wrote in particular seven brief lessons on physics.
And it is magical to see somebody demystify something like physics.
I don't even like physics.
And reading his writing is beautiful.
And you feel like you're in the hands of somebody who deeply understands something but can make black holes and quantum physics poetic.
And it's good for people who aren't interested in physics, but it's good for anybody who wants to see how someone can demystify a world and make it completely accessible in a great way.
So Carlo Rovelli is my recommendation.
That sounds really good.
And this is it for tonight.
Thank you for listening.
This was After Hours from the TED Audio Collective.
