The New National Champions

Plus! Open Source AI; Meme Private Credit; Alpha Generators as an Asset Class; Going Hostile; Defense

Byrne Hobart Byrne Hobart

In this issue:

This issue of The Diff is brought to you by our sponsor, Intercom.

audio-thumbnail
The Diff January 27th 2025
0:00
/709.433469

The New National Champions

One of the unique features of the US economy is that while we have some very big companies, the biggest US companies are a comparatively smaller share of national output than the biggest companies in other places. Samsung is about a fifth of the Korean economy, and last year they were about half the growth. Volkswagen's annual revenue is almost 8% of German GDP, TotalEnergie is almost 7% for France, TSM is close to 12% of Taiwan's, and Novo Nordisk is about a tenth of Danish output.

A lot of this arises from statistical artifacts: these big companies all sell globally, so it's cheating a bit to compare them to local GDPs—Samsung isn't selling 20% of all goods and services within Korea, though in some sense roughly 20% of those goods and services are either made by Samsung or paid for with the proceeds of things Samsung sells elsewhere. If multinationals run into constraints on size—only so many cars or chips can be built and sold—then the ones that happen to be headquartered in other countries will tend to be bigger relative to those countries. The US also has a big, complex economy, and many of its brands and distribution networks were built piecemeal, a long time ago, so there was room for multiple competitors.

Another reason for all of this is policy: "pro-business" can mean different things in different contexts If you're describing someone you like, what you probably have in mind is minimal red tape for starting new companies, low taxes, and the like; if you're talking about someone you hate, you probably mean crafting policies that favor big corporate donors at the expense of the common man.

Donald Trump is often described by his supporters and detractors as pro-business.

There's a particular brand of "pro-business" that the US does a bit less of than other countries, though it was once quite fond of it—selecting a handful of companies and making them partly policy arms of the state (or, equivalently, big companies selecting politicians because it would be nice to have a flag, a currency printing press, a navy, etc., while outsourcing the messy and hard-to-scale business of keeping constituents happy). There are some national interests that are best served with a combination of state-provided capital and protection and some layer of market-based incentives guiding the details.

The US did plenty of this in the late 19th century, granting railroads land in exchange for building and operating railroads that would otherwise be unprofitable. Aside from the massive corruption throughout this process, it had a certain kind of elegance: the land was worth a lot more if it had easy access to affordable transportation for both goods and people, so the asset the government gave away was worth less than the asset the railroads got.[1] High tariffs were more of a blanket version of this policy, taxing American consumers in order to make manufacturers richer, but those manufacturers had a high propensity to reinvest and America had plenty of investment opportunities at the time, so that might have been the high-ROI choice.[2]

The US model has moved away from this in most sectors, but still practices certain kinds of protectionism:

Boeing probably counts as a national champion by the same standard, but to a much smaller degree, than a business like Samsung. It is a symbol of American economic might and engineering prowess (for better or for worse, at times), and the US is hardly the first country to tolerate a bit of deadweight loss in exchange for global bragging rights. In fact, Boeing is also the recipient of funds misallocated in exactly the same way; flag carriers are basically a transfer to the home country's elite ($, Diff), and that transfer entails buying planes from one of two companies. Between that, and foreign countries' energy consumption subsidies reducing their share of energy exports, there are some forms of economic nationalism practiced by other countries that net benefit the US.

America is, or at least is rumored to be, considering a few new national champions. Trump has suggested that either Oracle or Elon Musk could acquire TikTok, perhaps in partnership with the US government. Musk is, of course, a massive Trump donor; Ellison actually spent $35m supporting Tim Scott as an alternative during the primaries. (Political donations turn out to be a fun variant on Polymarket for people who don't recall offhand how many vacation homes they own.) But both groups, Oracle as a company and Musk as the center of a constellation of them, have skills that the government lacks. Destroying TikTok might serve some US policy goals, but those goals are mostly achievable with the company under US control.

As with many other instances of national champions, the most value-creating policy, or at a minimum the least value-destructive one, requires some cooperation between the public and private sectors. The US government probably can't run a massive consumer-facing app (though they do employ many of the people you'd want scouring it for backdoors). Oracle probably didn't have that kind of diversification in mind, and Elon's bankers are presumably praying that he doesn't decide to add TikTok to his portfolio. But these groups could conceivably keep the business running (and in Elon's case, he's hopefully gotten some of the obvious mistakes out of his system).

Governments have access to certain kinds of capital that aren't fungible with the ones the private sector uses, but that are complementary to it. This can go in dark directions, of course: there were ample synergies between gunboats and United Fruit. But it's created plenty of valuable industries; the entire consumer tech experience is shaped by products that were subsidized by the government due to their early military applications, telcos have natural ties to government that benefit both sides, and the banking system doesn't make sense without a government backstop, and can't function well without regulations that tamp down the behaviors that are unavoidably incentivized by that backstop. A world where everyone carries a computer in their pocket that has access to all of their information, and where opinions are shaped by social media feeds, is a world where consumer-facing software drifts into the same half-public/half-private category.

Disclosure: Long TSM. Some national champions do just fine.

  1. This is a gilded age version of Paul Graham's equity equation, which is basically the argument that you should sell equity to investors or offer it to employees if the value of the remaining shares rises as a result. This is not a perfect rule—you don't want to take a near-breakeven deal at the expense of doing a better one in a few weeks, and if your company generates enough cash to grow and has a small addressable market, selling the equity is expensive unless the company has a lower return than alternative investments you could make (in which case maybe you're better off as an investor than as a founder, anyway). But it does help people get past the simple mental block of thinking about how much that outside investor would multiply their money in the event of a good outcome. ↩︎

  2. If a country has a big market for its manufactured products, and is technologically behind other places, and has high wages—as the US has had since the colonial era—tariffs are basically a way to shift output away from already-adequate consumption and towards high-return investment, which ends up paying off over time. The East Asian countries that applied a similar model did not have high wages, but they did tend to have authoritarian governments that could stay in power even if the average worker was annoyed that they didn't have an especially good standard of living. ↩︎

A Word From Our Sponsors

Do you know what would make your customer service platform dramatically better? Dumping it and switching to Intercom. But maybe you’re not ready to make that big of a change. We get it! That’s why we made Fin, our leading AI customer service agent available on every helpdesk.

Fin instantly resolves up to 80% of customer questions around the clock in 45+ languages. That means you can get off the customer service rep hiring treadmill because your tough-to-fill shifts are covered. With Fin handling frontline support, your human agents are freed up to do higher impact work—like managing complex questions and building deeper relationships with customers. And the best part is that you can experience all of Fin’s benefits from the comfort of your current platform—no switch required.

See why more than 3,700 companies trust Fin with their customers and business.

Learn More

Elsewhere

Open Source AI

There's some debate about whether DeepSeek's training cost numbers are real, but there's enough evidence that Meta, at least, is quite worried ($, The Information). (For more on DeepSeek, Jordan Schneider and Ben Thompson have good rundowns.) It's an open-weight model like Llama, and obviously Meta sees it as a serious threat to their own open-weight adoption. But given Meta's general AI strategy, that's not an imperative. They want to commoditize models because they expect to use theirs at more scale than anyone else, making them the biggest beneficiary of any improvements other developers make. As long as they own distribution, the best open model is good news whether or not it's theirs. Another feature of open models is that if the model is a commodity, that raises the value of its complement, the hardware. And it especially raises the impact of different access to hardware; it's possible to end up in a world where China develops the best model, but that model is monetized by the US companies that can produce the cheapest tokens.

Disclosure: Long META.

Meme Private Credit

The valuation of Elon Musk's public company has been on a tear since the election. Does that extend to his less frequently-marked private assets? And, if so, does it extend throughout the capital structure? His bankers appear to believe so; they're planning to sell some of the loans they made to fund the acquisition, with senior ones trading at 90-95 cents on the dollar ($, WSJ). In an internal memo, Musk (who denies sending said memo) is quoted saying "Our user growth is stagnant, revenue is unimpressive, and we're barely breaking even." That's a fairly rich valuation for a company that isn't currently growing or making money, after dramatic headcount reductions. Lending in the first place was something the banks did more to preserve a valuable relationship than because they were delighted by the risk; the haircut the banks take ends up being a measure of how much they valued that relationship.

Alpha Generators as an Asset Class

Hedge fund Eisler Capital is, for a second year, asking its traders to return bonuses if they leave the fund the year they receive the bonus. The market for portfolio managers remains extremely tight, and as bigger firms build on their economies of scale, they end up being the most profitable place to run any given strategy. In that environment, smaller funds have an incentive to lock them down for longer, by raising the transaction cost involved in switching jobs. One elegant feature of this model is that it imposes adverse selection on competitors: if a trader has a bad year, and gets a terrible bonus, they take a smaller hit when they leave. But their employer feels like it’s taken a smaller hit from that, too.

Going Hostile

QXO is a planned roll-up in the building products industry, founded by Brad Jacobs, who has previously run several successful rollups in other spaces. They've been negotiating for a few weeks to acquire Beacon Roofing Supply, a publicly traded competitor, but have been rebuffed—so they've gone hostile, making an unsolicited bid at $124.25/share ($, WSJ). Beacon's up a bit today, but trading below the value of that offer. For a while now, purely hostile offers have gone out of fashion; both corporate acquirers and PE firms see M&A as an iterated game and know that they'll spend more for the next deal if they're too aggressive for the first one. But hostile takeovers are part of the standard M&A repertoire for a reason: sometimes, boards and management have bad reasons not to sell, while shareholders have good reasons to support a deal. When antitrust enforcement is strict, a hostile tender offer can lead to owning a material chunk of a company, at a premium, but not being able to acquire the whole thing. If the average deal is more likely to go through, that's relatively beneficial for the ones that go through despite management's objections.

Defense

Lithuania and Estonia have agreed to raise their target defense spending to 5% of GDP ($, FT), up from 2.9% and 3.7%, respectively. One feature of a more isolationist foreign policy is that it shuffles around some spending in an accounting sense without necessarily changing the aggregates: if the US were providing exactly the same defense investment to protect European countries, it would mean a larger US deficit funded by exports from those countries. If the same countries are doing the defense spending, then at least some of it will involve buying US military equipment. But that's replacing consumption foregone for exports with consumption foregone to pay for non-consumption imports, and for the US defense contractors whose equipment is part of this buildup, it changes who's getting invoiced with a smaller impact on what's getting sold.

Diff Jobs

Companies in the Diff network are actively looking for talent. See a sampling of current open roles below:

Even if you don't see an exact match for your skills and interests right now, we're happy to talk early so we can let you know if a good opportunity comes up.

Find a Role

If you’re at a company that's looking for talent, we should talk! Diff Jobs works with companies across fintech, hard tech, consumer software, enterprise software, and other areas—any company where finding unusually effective people is a top priority.

Find Talent with Diff Jobs