# Financial Entrepreneurship in the Indexing Era > Source: > Published: 2026-08-17 15:18:24+00:00 In this issue: - Financial Entrepreneurship in the Indexing Era—If it's trivial to put money into indices and get decent returns, and it's hard to get access to or replicate the approaches of strategies that produce uncorrelated alpha, what are market participants actually doing? One of their important functions today is to get companies ready to be part of the index. - Easy As—It's good to have visible fine-grained distinctions among intellectual elites, because the alternative is that those distinctions still exist but they're harder to understand. - Missing Estimates—Anthropic once again produces incredible financial results that indicate that they aren't spending as fast as they should have. - Oil—The US and China export financial stability in very different ways. - Off-Balance Sheet—The biggest obligations are the ones it makes the least sense to worry about. - Turnover—There aren't many people who are equally comfortable at an early-stage company and at a very visible publicly-traded megacap. [Talk to this post on Read.Haus](https://read.haus/pi/vItMx6pY?ref=thediff.co). ## Financial Entrepreneurship in the Indexing Era It's fair to divide the world of asset management into three over-broad categories: - Many flavors of passive investing, where you make one general bet, like "I like large-cap US stocks" or "I want to own a little of every equity" or "I'd like 110%-minus-my-age of my money in equities, and the rest in bonds," and then, ideally, never think about it again. There are passive vehicles that can be actively-managed, like sector-specific ETFs, and you're also interacting with the passive ecosystem if you use S&P futures to make massively levered intraday bets. - There are investment vehicles, marketed primarily to institutions, that try to net out basically any risk that you could cheaply take in category 1. They still take some correlated risks; [last month's multistrategy hedge fund numbers show that most of them were on the same side of the AI trade](https://www.bloomberg.com/news/articles/2026-08-05/hedge-funds-take-big-hit-in-july-after-bruising-ai-selloff?ref=thediff.co)(and Citadel had some nice factor timing). - Some investments take on factor exposure, but either because of their structure or mandate, they don't hedge it out, and bundle it with their offering. This encompasses lots of long-only strategies across different products, which are sold as a way to get the best kind of exposure to the category. In some cases, there isn't a good way to hedge. In other cases, there's natural investor demand for the hedgeable piece of the risk. Some of these products get sold to individuals, but institutions are in a better position to evaluate when it makes sense to pay tens of basis points in extra fees in order to get something that correlates with the S&P but is also trying to beat it.[[1]](#fn1) Passive investors [aren't completely passive](https://www.economist.com/finance-and-economics/2026/08/11/hooray-for-index-funds-just-dont-call-them-passive?ref=thediff.co) ($, *Economist*). They're making choices when they allocate to different asset classes, or choose different ways to represent a given asset class—buying the S&P 500 is an allocation to large-cap US companies that report in dollars, not to the global economy, though the distribution of market cap by company and country is so skewed that buying the S&P gets you exposure to about 40% of global equity market cap. Still, it isn't the complete portfolio of every asset; it's underweight housing, precious metals and other commodities, and fixed income. It also contains an implicit political economy bet: one reason American companies are so valuable is that they aren't strictly American, and get [28% of their revenue from overseas](https://www.marketwatch.com/story/heres-how-much-revenue-s-p-500-companies-make-from-overseas-ce10f434?ref=thediff.co). That international expansion tends to have a higher incremental return on capital than their home market, because there are some expenses they don't need to pay twice. But that means that the S&P is, implicitly, a bet on other countries' continued willingness to do business with the US, and for the profits from that to disproportionately go to Americans. Still, indexing is popular for a good reason: [we can't all be above-average](https://web.stanford.edu/~wfsharpe/art/active/active.htm?ref=thediff.co), and every active manager with above-average returns must be matched by some active manager with below-average returns. Put another way, for an imperfectly efficient market, alpha can be defined equally well as either investment skill or as a measure of the sales and marketing talent of below-average active managers, or the avoidable losses from individual investors who pick stocks, founders and employees who have poor timing when they sell their equity stakes. But another piece of this is that active management is a subsidy for passive, and passive is a subsidy to companies that make it into an index. When we talk about investor skill, we're talking about ability relative to whoever is the marginal price-setter. The more efficient labor markets are, the higher the skill threshold for being good enough to beat the market, versus being better than the average person at reading up on a company and putting a fair value on it. Stock prices are often counterintuitive: if you can look at a company's financials and history and guess the value of their shares to within 20% of where they actually trade, that's pretty impressive, but being off by 20% is still a big margin of error. [2] If an investor bought a broad index fund in 2000, that index fund allocated some money to energy, and that means it allocated some money to Enron; if [short-selling hedge funds](https://fortune.com/article/is-enron-overpriced-fortune-2001/?ref=thediff.co)had had a bigger market impact at the time, that weighting would have been smaller and the cost of Enron's collapse to everyday investors would have been smaller. When prices bob around randomly, it gives overpriced companies an opportunity to issue more stock (increasing their share of the index) while underpriced ones might buy back shares (decreasing their weighting). The more efficiently they're priced, the less indexers bleed returns from this dynamic. Less liquid marketplaces aren't subject to this kind of discipline, so they tend to have more persistent mispricings. That also makes them worth less. Investors in private companies get higher returns on average in part because, unlike public markets, there isn't a trivial way to own "on average." It happens a deal at a time, and someone who's investing in these markets as an asset allocation decision, rather than a directional view about the future state of the world, is a good target for adverse selection. This also means that you can look at many illiquid corners of the market as, basically, a collective effort to make assets legible to capital markets so they can fit into an index, at which point active public market investors will compete to make their pricing as efficient as possible. This is obviously a big part of what venture capital does. It's hard to take a company public when it's three friends writing code on laptops in whoever's apartment can fit the biggest desk (unless it's late Q4 1999 or Q3 2021, when *anything* with a good enough story can go public). Plenty of VC advice will be useless, but VCs are very attuned to what people think about a stage earlier or a stage later than where they participate, and late-stage investors are more or less writing the informative parts of the S-1 as their investment memo when they do a deal. All of these investors are picking up some excess return because BlackRock and State Street aren't bidding for the same assets they do, and they're underwriting the risk that the company they invest in won't end up on the list of companies that those firms automatically buy more of every time they land some new assets. Private equity does this, too. The PE model is to take a company that's either private and not ready to go public, or public but not understood by the market, and clean it up to the point that it's something Baly, Citadel, Millennium, and all the rest will have an opinion on every quarter. That often involves cost cuts, divestitures, and the like, but PE firms don't tend to take companies public when they're growing revenue slower than inflation and will eventually be a zero—they try to build a growth story and go public with *that*. The net result is that the financial system ends up functioning like a utility. Today's investors can, if they choose to, passively and tax-efficiently harvest various risk premia for decades, and, depending on exactly what asset they choose; their lifetime fee burden as a share of assets is less than their grandparents might have paid a mutual fund manager for a single year. That's an extraordinary accomplishment for the industry, and like any other product that's universally available and Just Works, it's the result of staggering collective efforts by many brilliant people (roughly all of whom were paid well, or, for the academics, had the option of getting paid well by the industry if they wanted it). There are people who complain about their electric bills this summer who, in the absence of cheap electricity powering climate control, would not have been complaining about anything at all because they'd be dead. That's how you know an industry has achieved its mission: if what it offers is so ubiquitously accessible that there's room to complain about what previous generations would have seen as a miracle. For example, if you're investing in early-stage startups, the only publicly-traded proxies are very loosely related to them. The Nasdaq 100 might go up because of general optimism about tech, which correlates with startups' fortunes. But it can also go up because investors get more convinced about the durability of public incumbents, meaning that a startup investor loses on both sides of the trade. And there's also the risk that the liquid leg of the trade gets a margin call that can't be satisfied by borrowing against or selling the illiquid one. For any equity, somebody has to have the net long exposure, and sometimes it's you. In other asset classes, there can be similar constraints: the smaller stocks get, the more differentiated long and short investing are: smaller companies are more likely to have accounting issues or governance problems bad enough to eventually zero the stock, but betting against those only loosely hedges a portfolio of small-and-misunderstood businesses. [↩︎](#fnref1)To illustrate how hard this is, consider two companies that both do $4-5bn in revenue, have roughly 30% EBITDA margins, and are growing revenue above 5% a year. One makes industrial scales, one has a brand name synonymous with video chat. The scale company, Mettler-Toledo, has slightly lower EBITDA, but its enterprise value is 30% higher than Zoom's. [↩︎](#fnref2) You're on the free list for * The Diff*. This week, paying subscribers read about [Mark Zuckerberg's AI manifesto](https://www.thediff.co/archive/toll-collectors-on-the-road-to-abundance/)($), [how a nuclear fuel disposal company is deploying SMRs](https://www.thediff.co/archive/holtecs-very-roundabout-drop-in-replacement-model/)($), and [using customers as a source of R&D](https://www.thediff.co/archive/customers-as-r-d/)($). Upgrade today for full access. [Upgrade Today](#/portal/signup) **Diff Jobs** Companies in the *Diff* network are actively looking for talent. See a sampling of current open roles below: - High-growth startup building dev tools that help highly technical organizations autonomously test and debug complex codebases is looking for someone who can help scale and manage their rapidly expanding facilities footprint: everything from lease negotiations, to proactively anticipating / catalyzing new office acquisition, to making sure those offices are stocked with the best snacks. If you want a seat on a rocketship and enjoy fixing things that break–literally and figurately–this one is for you. (SF, DC, London) - Well-funded, frontier AI neolab working on video pretraining and computer action models as the path to general intelligence is looking for researchers who are excited about creating machines that learn from experience, not text. Ideally you have zero-to-one pre-training experience and/or are a high-slope generalist who’s frustrated that the big labs aren't doing this. (SF) - Lightspeed-backed team building the engineering services firm of the future is looking for founding members of technical staff excited about working alongside civil engineers to translate their domain expertise into the operating system that powers the next era of great American infrastructure. If you’re an engineer with strong product intuition, who's energized by access to users, and excited by the prospect of transforming how we design and construct our built world with frontier AI, this is for you. (NYC, SF or Remote) - Ex-Anduril, Ex-Abnormal Security, Ex-Bridgewater, fast growing startup bringing agentic cybersecurity to 99% of businesses via MSPs is looking for platform and machine learning engineers. Startup experience preferred; what matters most is that you've grown in scope and handled ambiguity over the last few years. (SF) - AI Transformation firm with an ambition to build an economic world model to run swathes of the private, unstructured economy is looking for Systems Engineers, Platform Engineers, and business generalists who understand how to solve problems. 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. And: we're now actively deploying capital into early-stage companies through [Anomaly](https://www.anomalyfund.vc/?ref=thediff.co). Our focus is on defense, logistics, robotics, and energy. If you'd like to chat, please reach out. ## Elsewhere ### Easy As Harvard [is making it harder to get an A, which will make life harder for their undergraduates](https://www.bloomberg.com/news/features/2026-08-12/harvard-won-t-limit-a-s-until-fall-2027-but-faculty-are-making-changes-now?ref=thediff.co). To an extent, employers have already patched some of the deficiencies in how well grading distinguishes between different students. For some career tracks, internships are effectively an extra point or two of GPA, but that means that getting a quick assessment of ability in some domain means knowing which interships are the most prestigious, and that having an even slightly out-of-date view gives a misleading impression. It will be nice to have a better read on how hardworking the smartest young people in the country are, and how smart the most academically compliant are. It's going to be hard for those students, if they've gone through life taking lots of tests that aren't sensitive to their position within the intellectual elite and suddenly have to cope with a context where they can achieve a B average if they work hard for it. But that's a positive externality: everyone has to encounter their first serious psychological challenge some day, and a semester's GPA is the lowest-stakes context imaginable. ### Missing Estimates Anthropic's revenue last quarter was [up 14x year-over-year, and they once again reported an operating profit](https://www.bloomberg.com/news/articles/2026-08-14/anthropic-revenue-ahead-of-ipo-surges-over-14-fold-in-second-quarter?ref=thediff.co). That profit illustrates how hard it is for labs to calibrate their spending, and why their capex is so aggressive: a business growing that fast should absolutely be a net consumer of capital, because it can quickly turn that capital into something more valuable. Anthropic probably aimed for this, but underestimated how much compute they'd be able to profitably sell. It would be strange if the end result were superhuman intelligence in all domains and the company that made it somehow managed to earn a profit beforehand; almost anything they could have spent to accelerate the development of that software would have been worth it. But since they weren't spending money that way, they presumably underestimated just how much money they could make selling faster software development even if it isn't superhuman at everything else. ### Oil It's striking [just how much of the Iran war-induced oil shortage has been offset by less consumption from China](https://www.economist.com/finance-and-economics/2026/08/09/china-is-now-the-worlds-great-oil-power?ref=thediff.co) ($, *Economist*). When there's a global recession, the US tends to export growth to the rest of the world by running a deficit and cutting interest rates until Americans start to consume. Since some of that consumption consists of imports, the US exports demand (and dollars) when there's a shortage. China is apparently playing a similar role for real resources: they have enough state capacity that it makes sense to have contingency plans for a sudden drop in the supply of oil, so they end up being a swing demander of oil in the same way that Saudi Arabia has historically been the swing producer. ### Off-Balance Sheet The *WSJ* [has a piece about big tech companies' $3tr in off-balance sheet liabilities](https://www.wsj.com/tech/ai/why-big-techs-ai-spending-is-3-trillion-higher-than-it-seems-e1067bb2?ref=thediff.co) ($). The last time that term got thrown around a lot, it was in the early days of the financial crisis, when banks turned out to have moved some of their levered structured product trades off their balance sheet while still being committed to funding them. It was generally abd news. But this is a bit different, and it's easy to misread the numbers: these liabilities are the total sum a company has committed to paying, but they aren't discounted back to their present value, and they're only capturing gross costs, not the returns companies might get from them. In general, the longer-term the obligation, the more fungible it will be: it can make sense to commit to land and power over multi-decade periods, but it doesn't make sense to precommit to buying a certain amount of compute or a certain number of chips ten years from now. There's just no good way to know what will be available. So while these numbers look intimidating, they're basically some mix of COGS, opex, and capex that will be realized over many years in the future (and every time they get realized, that converts the abstract liability into an actual cost, shrinking the original problem). ### Turnover OpenAI [has lost a few senior executives and reshuffled others ahead of its IPO](https://www.axios.com/2026/08/14/openai-executive-greg-brockman-ipo?ref=thediff.co). This is pretty close to a tech tradition at this point, and there are a few drivers. One is that the market reacts negatively to executive turnover, so if it's going to happen anyway, it's a good idea to get it out of the way when there aren't meaningful daily quotes. Second, while there are some people at early-stage companies who feel like they were born to be executives at a big, public company, there are others who work at private companies in part because that's the environment they prefer. If there's some order of magnitude of headcount or revenue that someone is most comfortable with, it doesn't take much time for a company growing at OpenAI's pace to blow through that and end up bigger. What you'll sometimes see in an S-1 is that the company has assembled an IPO-ready team in the six to twelve months before the IPO happens; it's a common pattern, just less newsworthy when the company is worth a few billion dollars rather than almost a trillion.