The Year of Convergence

In this issue:

• The Year of Convergence—This year marks the biggest outperformance of European equities relative to the US in almost two decades. And the driver is, indirectly, an echo of the dominant macro thesis of the pre-crisis years.
• Ads—Apple can monetize in some of the same ways Google can, but doesn't have the right feedback loop to know how to maximize returns.
• Maximalism—Every moment you spend doing something other than interact with AI is, from OpenAI's perspective, an opportunity.
• Comparative Advantage—Sometimes the economically efficient outcome is to avoid certain economic activities.
• Trader-in-Chief—Scott Bessent ascends to macro fund manager Valhalla.
• GenAI—Where generative video ads start.

The Diff October 27th 2025 13:35 • Click to play audio

The Year of Convergence

If you looked at global market caps circa 2010, you'd see that the US was about a third of the total value of global equities. That made a certain amount of sense: the US economy was about a quarter of global output, but America has a disproportionate listing share among global companies, and the Anglosphere has historically been more fond of financing businesses with equity, compared to the global norm of closely-held businesses financed with bank debt. So, it's not crazy for the US to be overrepresented. But now it's nearly two thirds of the MSCI ACWI index, whose constituents are ~85% of global market cap. This isn't completely crazy, because for the last few decades the US has had a comparative advantage in attracting global talent and global capital, which have a complementary relationship ($, Diff).

It's hard for one country's equity market to outperform the rest of the world over an extended period, and 2025 is an exceptional year: it's the first time since 2017 that the US markets are up and Europe has outperformed, and this year so far marks European equities' biggest outperformance relative to the US since 2006. And it's useful to compare the circumstances that held in the early 2000s to today, because both periods were arguably a realignment trade, during which investors still liked the US but liked the rest of the world a lot more.

If you look at a surprisingly strong, unsustainable secular trend, there's a good chance that you're looking at a long epicycle, and that the trend you're paying attention to starts at the point where the opposite trend had been running for unsustainably long. In the early 2000s, there were two ways to describe the big bet investors really wanted to make:

1. The Convergence Trade: GDP growth would be higher in poorer countries than in richer ones, and as these countries grew, their institutions and culture would converge on rich-world norms. So, not only did you want to own cheap emerging-market equities rather than expensive US ones, but you also wanted to lend to these countries, and own their currencies. Within the US and other developed markets, the spread between high-yield bonds and investment-grade credit was collapsing, which was also a sort of convergence bet. Even equity volatility was low. People didn't start making "nothing ever happens" jokes until a few years ago (though the catchphrase is older, produced by the Breeder Reactor of Content that is anonymous imageboards).
2. The China Trade: the countries that were growing fastest in this period were China beneficiaries, one way or another: developing countries tended to have economies heavily weighted towards natural resource extraction, and China needed a lot of that to build out. Big global exporters were also buyers of low-risk, dollar-denominated assets, which they used to maintain cheap currencies. Within Europe, Chinese exports priced low-value added manufacturing out of the continent, but those economies shifted to sectors where they had a bigger comparative advantage ( higher-margin manufacturing in some cases, housing/tourism/finance in unhappier ones). This was nicely redistributive to US consumers, who needed to consume more than they produced in order to satisfy the world's voracious demand for dollar-denominated debt, but it wasn't great for the relative standing of America's companies (though some companies, like American oil and gas E&Ps benefit from higher global oil prices).

These frameworks are two ways of looking at the same outcome: a world where China's GDP per capita keeps rising and eventually becomes comparable to that of the US and Western Europe is one where Chinese companies can compete globally on quality as well as cost, and it's a world where the places that supply inputs to China’s buildout get a constant influx of profits that they can use to diversify away from depending on natural resources.^[1] It would also presumably require the US to rearrange some of its economic policies to be more similar to China's, or vice-versa. Commodity bulls in the 2000s were fond of doing the quick mental math on just how much oil the world would need if the average person in China and India drove as much as the average American, and to compare that to the pace at which new oil discoveries showed up.

That convergence trade fell apart in a big way, and the economic story since then has involved the US pulling ever further ahead of the rest of the world, often because of America's existing advantages. A reserve currency issuer can afford to bail out its financial system, and has a strong incentive to do so because the financial system’s strength is part of what makes reserve currency status so sticky.^[2] The country that basically invented the original-issue high-yield bond market was a great place for frackers to get funding. And of course today the extremely capital-intensive AI buildout happens because of deals that get dreamed up in San Francisco and have their implications priced out in Midtown.

So if global economic growth is more a story of intensive rather than extensive growth&dmash;zero-to-one, not one-to-N&dmash;it means that more of the world's future mega-caps will be founded in the US, and that over time more market cap will accrue to the US.

But right now, the world is betting on a different convergence trade, where neither the US nor China are as trustworthy as they used to be for providing either critical links in the supply chain or the defense/legal/logistical network that ties the world's economies together. In the US, baroque foreign policy constructions like "strategic ambiguity" or the "rules-based order" are being pushed aside in favor of a jazzy, improvisational approach. China is responding in kind. Tariffs, sanctions, and strategic embargoes mean that there are more countries that need to be able to make things domestically, and if national borders are sometimes subject to a very kinetic debate, that supply chain question extends to military hardware. European countries are rearming, Japan's new prime minister is aiming for the country's military spending to hit 2% of GDP (not too long ago, the Japanese government insisted that 1.3% was plenty ($, FT)).

If this isn't a prudent response to a new set of risks, it straightforwardly makes everyone poorer. For scale-driven supply chains, the cost-optimal approach is specialization, and global monopolies, but this creates chokepoints, and those create leverage for whoever's willing to exploit them. Free trade as a global norm actually makes economic issues more political, even though it's a policy that nominally reduces governments' role in the economy. What it does in practice is to make more economic issues foreign policy issues. It's only through globalization that there can be protests over food prices in Africa ($, FT) as a result of a land war in Europe, or a global spike in cooking oil costs driven by domestic unrest in Indonesia, or a London commodity exchange getting sued by an American hedge fund because a Chinese nickel producer got a margin call. The global village is a place where your day can be ruined by the whims of someone in a place you can't locate on a map—especially if you're in the bottom half of the global income distribution and have to spend a lot of your income on food and fuel.

When countries insource more production, they create globally redundancy and excess capacity, and they reduce the upside from sharing tacit knowledge. But this policy also entails a wealth transfer from governments to the private sector: if your country is going to be able to produce drones for defense, small modular reactors for energy, robots to offset a worsening dependency ratio, etc., private companies will tend to get paid to do it. So more countries whose equities are cheap because they're the fifth-best place to do something that only needs to be done well in one or two places will do better in a closed borders scenario. Many of the globally dominant companies displaced by this trend aren't American, but they're still part of a US-centric system, where the US is a big, standards-setting customer, and purchases are invoiced in US dollars. Rolling globalization back a bit isn't great for consumers who were used to getting the basics cheaply, and who collectively ensured that the R&D burden for more advanced products was amortized across as many customers as possible. But for the companies that were less competitive before and are more competitive now, it's a boon, and the relative outperformance of Europe shows this.

US equity market fundamentals seem pretty intact; earnings continue to grow, margins are expanding, etc. Though one reason for that is an accounting detail: some of the biggest consumers of capital right now are private companies whose accounting losses aren't part of the S&P's aggregate numbers, but whose losses take the form of spending that disproportionately flows to high-margin S&P 500 members. AI-adjacent earnings growth is a little over half of the S&P's earnings growth this year (as a rough guess, $1bn in incremental capital raised by labs leads to ~$150m in additional net income for the S&P, which at 30x earnings means it produces $4.5bn in additional market cap). But the S&P is a unique index; outside of a tiny handful of firms, other countries' major indices are closer to the "S&P 493" non-Magnificent 7 instead. And for them, a world with higher trade barriers and more government intervention in the economy is just fine.

Diff Jobs

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

• A growing pod at a multi-manager platform is looking for new quantitative researchers, no prior finance experience necessary, exceptional coding and stats skills required. 250k+ (NYC)
• Well-funded, fast-moving team is looking for a full-stack engineer to help build the best AI powered video editor for marketers. Tackle advanced media pipelines, LLM applications, and more. TypeScript/React expertise required. (Austin, Remote)
• A transformative company that’s bringing AI-powered, personalized education to a billion+ students is looking for elite, AI-native generalists to build and scale the operational systems that will enable 100 schools next year and a 1000 schools the year after that. If you want to design and deploy AI-first operational systems that eliminate manual effort, compress complexity, and drive scalable execution, please reach out. Experience in product, operational, or commercially-oriented roles in the software industry preferred. (Remote)
• A leading AI transformation & PE investment firm (think private equity meets Palantir) that’s been focused on investing in and transforming businesses with AI long before ChatGPT (100+ successful portfolio company AI transformations since 2019) is hiring Associates, VPs, and Principals to lead AI transformations at portfolio companies starting from investment underwriting through AI deployment. If you’re a generalist with deal/client-facing experience in top-tier consulting, product management, PE, IB, etc. and a technical degree (e.g., CS/EE/Engineering/Math) or comparable experience this is for you. (Remote)
• Deerfield-backed Series A company building agents for healthcare administration (prior authorization, eligibility checks, patient scheduling) is looking for a senior software/AI engineer to build backend services and LLM agents. Experience building and monitoring production-quality ML and AI systems preferred. (NYC, Hybrid)
• Ex-Citadel/D.E. Shaw team is building a new financial market within a multi-trillion dollar asset class and looking for a junior data scientist with some classical and generative ML experience.  (NYC, Boston)

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.

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.

Elsewhere

Ads

One of the funny load-bearing features of the consumer Internet economy is that Apple 1) sells its products to high-income, free-spending consumers, 2) makes really great software and hardware that presents as few obstacles as possible to browsing, shopping, and buying, and 3) has been mostly uninterested in ads. So they've basically outsourced the ad side of the business to companies that own apps and sites instead. But they're working to fix that, and plan to include ads in Apple Maps soon. One of the problems they have is feedback. If Google (disclosure: long) has excessive ad load in its own maps app, they'll start to see usage dip, and losing Maps users means losing a lot of context that can be used to show those users ads in other contexts. Apple is a little less sensitive to this, because selling devices is still the core business, and they probably won't have much churn from monetizing their maps product the exact way their biggest competitor does. But that means that it's hard for Apple to understand what the potential of the business is, or to read the tradeoffs. Whatever they get will still be a nice windfall, but they'll never know how much they missed.

Maximalism

OpenAI is working on a generative music product ($, The Information). AI-generated music fits in with many of the other directly consumer-facing AI use cases: it's great when there's a high production cost for mere mediocrity (e.g. commercial jingles, parodies, novelty songs, etc.). This is not a huge economic niche, but one way to look at it is that the limit of generative AI's impact is set by how much time people spend consuming different media. If there's a context where it's hard for them to actively interact with AI but possible for them to passively consume it—driving, doing chores, working out—that's one more avenue for OpenAI to capture every spare second of attention.

Comparative Advantage

The Economist has a fun piece looking at how much more lucrative it can be for Brazil to preserve the carbon-sequestering Amazon rather than converting it into agricultural land ($). One reason that Brazil deforested so much turns out to be insufficient rather than excessive capitalism: forged land deeds and a binary legal treatment for productive land implicitly subsidized people expropriating forests for unproductive ranching, whereas a more consistent and well-executed system of property rights would have left them alone. Sometimes, optimizing for growth really is an accounting question: if you put too low a dollar value on hard-to-measure outcomes, and then put a higher dollar value on measurable ones, you'll encourage exactly the wrong tradeoffs.

Trader-in-Chief

Life is never easy for global macro investors, because their job is to understand both economic fundamentals and how polymakers will respond. Sometimes, the core thesis can be exactly right, but the policy response can more than offset it. The dream would be to make the trade and then make the appropriate policy, and that makes Scott Bessent the luckiest macro trader on the planet; he can load up on Argentinian assets that are cheap because of fears that Peronists will win an election, and then see those assets instantly appreciate when, thanks in part to US economic support, the election swings the other way ($, FT). Granted, he's not personally earning those profits, but even before his government career, Bessent was rich enough that he was clearly doing this for the love of the game. And now, late in his career, he gets to play it on easy mode.

GenAI

Packaged food maker Mondelez (Oreos, Cadbury, Ritz, etc.) says it's cut the cost of new ads by 30-50% by using AI. Ads featuring humans are still tricky, both because of the remaining uncanny-valley effects and because of the inevitable-for-now backlash from replacing human actors. But snack foods already exist in the uncanny valley. Nothing resembling an Oreo exists in nature, and the easy availability of flavorful and energy-dense foods is also evolutionarily novel. So there's a comparative advantage in this kind of advertising. And that same evolutionary argument holds true here, too: it's hard to know exactly what will get someone to consume an incremental few-hundred calories right away, so it's valuable to test lots of commercial messages, see which ones resonate, and then continue to riff on those. Now that's cheaper; we'll see more unique ads, better targeted to individual consumers, and everyone's experience of ad-supported media will be a little less comparable than it was before.