This series has spent a lot of time reading the ledgers of the dead, the foundations and estates that still follow written orders long after the writer is gone. The Howard Hughes Medical Institute belongs to a different genus. It does not give money away. It spends it, on its own payroll, on scientists it employs directly, and it has arranged its affairs so that it can keep doing that essentially forever. Its most recent public Form 990, for the fiscal year that ended August 31, 2025, reports a balance sheet that reads less like a charity and more like a small sovereign nation: $29.5 billion in assets, $27.6 billion in net assets, and a portfolio that competes, dollar for dollar, in the same markets a schoolteacher's pension and a country's oil fund are trying to win. Nothing below is invented. All of it is arithmetic that has been sitting in a public file.
The tax dodge that became a cathedral
Howard Hughes founded the Institute in 1953, and the founding was not, at first, an act of obvious generosity. Hughes donated most of the stock of the Hughes Aircraft Company to his new medical institute, which had the convenient effect of turning a large defense contractor into a tax-exempt charity that he controlled as sole trustee. Congress noticed. In 1969 Representative Wright Patman called it a tax-evasion device and pointed out that the Institute had spent only about $5.7 million on operations across its first seven years while Hughes Aircraft piled up profits (Wikipedia, "Howard Hughes Medical Institute"). For two decades HHMI was a research charity that mostly owned an aerospace company.
The cathedral got its stone in 1985. The trustees put Hughes Aircraft up for sale, ran a five-month sealed-bid auction against Ford and Boeing, and General Motors won. The deal closed on December 20, 1985 for an estimated $5.2 billion, paid as $2.7 billion in cash plus 50 million shares of GM Class H stock (Wikipedia, "Hughes Aircraft Company"). Overnight, a charity that had been living on a few million dollars a year had a real endowment. Its budget went from roughly $4 million in 1975 to around $200 million a year by 1985, and it kept climbing. Forty years later that $5.2 billion has become $27.6 billion in net assets, and the Institute now spends more in a single year, $1.18 billion, than the entire company sold for in inflation-unadjusted terms would have covered for long.
People, not projects
The Institute states its method in three words that appear right in the filing: people, not projects. Rather than award research grants, HHMI hires scientists. An HHMI Investigator keeps a lab at a host university or hospital but becomes an HHMI employee, with salary, benefits, and a research budget paid by the Institute, on a seven-year appointment that renews indefinitely as long as the science holds up (HHMI, "Investigators"; HHMI, "For Host Institutions"). Each investigator represents roughly $11 million over a seven-year term (EurekAlert, via HHMI). The 990 says the model supports more than 265 current investigators, among them 16 Nobel laureates and nearly 150 members of the National Academy of Sciences, spread across more than 55 institutions. One of the names on the compensation schedule is Richard Axel, the Columbia Nobel laureate, listed as an investigator at $806,661 for the year.
The first program line on the return, the investigator program and the Janelia Research Campus, cost $832,476,143. Janelia, opened in Ashburn, Virginia in 2006, is the Institute's own lab, built to chase problems too interdisciplinary for a university, and it is where HHMI announced AI@HHMI in 2024, a ten-year, $500 million bet on artificial intelligence as a research partner (HHMI, "HHMI invests $500 million in AI-driven life sciences research"). The second program line, science education and fellowships, cost another $112 million and includes the SEA-PHAGES course that put 7,779 undergraduates into real research last year and BioInteractive, which reaches more than 70 percent of American high school and college biology classrooms.
How it invests, which is the whole story
Here is the part the user of this series asked to see plainly. HHMI's $27.6 billion is not sitting in a bank. Line 11 of the balance sheet reports $5,409,644,608 in publicly traded securities. Line 12 reports $21,916,648,565 in other securities, and the schedule behind that line names what "other" means: private equity, real assets, public equity held through funds, an "independent return" sleeve that is the polite name for hedge funds and absolute-return strategies, closely held equity interests, and financial derivatives. Roughly eighty cents of every invested dollar sits in private and alternative assets. That is not a savings account. That is the exact allocation you would find inside a large state pension or a national sovereign wealth fund.
The people who run it are paid like it, too. Read the compensation schedule and the surprise writes itself: the highest-paid officer at the Howard Hughes Medical Institute is not the president and not any Nobel laureate. It is Donald Koch, the Vice President and Chief Investment Officer, at $1,928,020 for the year. Below him sit a Managing Director of Private Markets at $1,207,419, a Managing Director of Public Markets at $1,160,055, and a head of investment operations at $1,043,689. The Institute's own materials describe an in-house team of about 30 investment professionals investing globally through long-term partnerships with external managers, split into a public markets desk and a private markets desk (HHMI, "Donald Koch"). The president, Erin O'Shea, the first woman to lead the Institute, earned $1,652,120, less than her own investment chief. Schedule J, which breaks the pay apart, sharpens the point: of Koch's compensation, roughly $1.14 million was base salary and the rest a performance bonus, and the two markets managing directors were likewise paid close to half in incentive pay, the structure of a fund manager rather than a foundation officer, while O'Shea drew an essentially flat salary with no bonus at all. The outside managers the Institute partners with surface on the vendor list too: three of its five highest-paid contractors are investment firms, Cohen and Steers, Whitehawk Capital Partners, and Accel, and the single largest is the operator of the conference center at its Janelia campus. Nine hundred and fifty-five people at the Institute were paid more than $100,000 in the year, but the ones paid like Wall Street are the ones managing the money. In this house the money comes first and the science is what the money buys.
That framing is not a criticism. It is the mechanism. In the year this return covers, investment income was $1,966,414,132. Contributions were $1,783,098. Program service revenue was $4,608,732. In other words, 98 percent of everything that came in the door came from the portfolio, not from donors and not from selling anything. The Institute earned nearly two billion dollars from its investments and spent a little under $1.2 billion running the science, which is how net assets grew by more than $2 billion in a single year and how the endowment is managed, in the Institute's own words in its audited statements, to fund the programs in perpetuity (HHMI FY2024 audited financial statements).
The same well everyone else drinks from
This is the point worth sitting with, because it connects HHMI to the retiree reading this. The Institute's $27 billion competes in the same equity and security markets that almost everyone's future depends on. When HHMI's public-markets desk buys the S&P 500, it is buying the same index in the same retiree's 401(k). When its private-markets desk fights for an allocation to a top venture fund, it is bidding against Yale's endowment, against the California teachers' pension, against Norway's oil fund and Singapore's GIC and Abu Dhabi's investment authority, for the same scarce seats in the same deals. The returns that fund a Nobel laureate's lab this year are drawn from the identical pool of market performance that funds a firefighter's pension and a couple's retirement account. There is one well. HHMI simply arrives at it with a very large bucket and a very good crew, and in a good market year it hauls up nearly two billion dollars. In a bad market year it will feel the drawdown the same way the pension and the 401(k) do, because it is invested in the same things they are.
What it owes, and what wears out
An endowment this size still carries liabilities, and HHMI's are instructive. Total liabilities are $1,879,493,442. The largest single piece is $935,242,011 in tax-exempt bond debt. Stop on that for a moment: an institution sitting on $27 billion still borrows nearly a billion dollars. That is deliberate. When you can borrow at low tax-exempt rates and your endowment is compounding at a higher rate, it is cheaper to issue bonds for your buildings than to sell investments that are busy earning more than the bonds cost. Large endowments do this on purpose. The rest of the liabilities include $274,775,084 in payables and $70,196,270 in grants payable, money already promised to scientists but not yet out the door.
Depreciation tells the physical half of the story. The Institute holds $2,230,992,106 of land, buildings, and equipment at cost, against $1,471,065,010 in accumulated depreciation, leaving $759,927,096 net. Two thirds of its capital base has already been depreciated on the books, which is what you would expect of a mature campus and a lab full of aging instruments. Science is not only salaries and portfolios. It is also microscopes and buildings quietly wearing out, and the ledger books that wear against the balance sheet every year.
For how long
Forever is the honest answer, or as close to forever as an institution can engineer. A perpetual endowment that earns more than it spends does not have an expiration date the way a spend-down foundation does. The risk is not that HHMI runs out of money. The risk is that it changes its mind, or has its mind changed for it. This filing already shows both. In February 2025 the Institute abruptly ended its $60 million Inclusive Excellence program, cutting off 104 institutions mid-award, a reversal of the $2 billion inclusion commitment it had made in 2021, reported against the backdrop of the new administration's orders against diversity programs (STAT, "HHMI ends STEM diversity program"). And the program narrative in this very 990 notes that in May the Institute paused competitions for two of its fellowship programs in order to redirect resources to its current scientists. A perpetual machine can run indefinitely and still turn on a dime, because perpetual means the money lasts, not that the mission is fixed. Howard Hughes proved that at the founding, when the same institution was a tax shelter. What the endowment guarantees is longevity. What it is used for is a decision somebody makes every year, and this year the decisions changed.
Related reading
- How Deathless Money Invests: the seven-style taxonomy of institutional money, and where a perpetual endowment sits in it.
- The Working Ledgers: HHMI set beside six other research institutions, from the pure contract shops to the charity that runs on toys.
- The Broad Institute's Billion-Dollar Endowment: the other Cambridge biomedical giant, and a very different funding mix.
- The Quiet Shelf, Volume Two: how ordinary foundations invest, sorted by how long the money has worked.
Fact-check notes and sources
- All financial figures (total assets $29,497,763,012; net assets $27,618,269,570; total revenue $2,004,784,057; investment income $1,966,414,132; contributions $1,783,098; program service revenue $4,608,732; total expenses $1,178,202,544; salaries and benefits $490,580,362; grants $63,319,232; publicly traded securities $5,409,644,608; other securities $21,916,648,565; tax-exempt bond liabilities $935,242,011; total liabilities $1,879,493,442; grants payable $70,196,270; land, buildings, and equipment $2,230,992,106 at cost less $1,471,065,010 accumulated depreciation; the program costs of $832,476,143 and $112,031,924; and the officer compensation for Donald Koch, Erin O'Shea, and the investment managing directors): read directly from the Howard Hughes Medical Institute's IRS Form 990 for the fiscal year ending August 31, 2025 (EIN 59-0735717), available free at ProPublica's Nonprofit Explorer. The investment-category names and Richard Axel's compensation are from the same return's Schedule D and Part VII. The split of compensation into base salary and bonus (Koch's roughly $1.14 million base, and O'Shea's flat no-bonus salary) is from the return's Schedule J; the top outside contractors, including Cohen and Steers, Whitehawk Capital Partners, and Accel, and the count of 955 individuals paid more than $100,000, are from Part VII, Sections A and B.
- The 1953 founding, the Hughes Aircraft tax-shelter history, and Patman's 1969 criticism: Wikipedia, "Howard Hughes Medical Institute", attributed.
- The 1985 sale of Hughes Aircraft to General Motors (closed December 20, 1985, roughly $5.2 billion as $2.7 billion cash plus 50 million GM Class H shares): Wikipedia, "Hughes Aircraft Company", attributed.
- The "people, not projects" model, seven-year renewable appointments, the roughly $11 million per term, and the more than 265 investigators / 16 Nobel laureates figures: HHMI's Investigators program page and the EurekAlert release; the investigator and institution counts appear on the 990's Part III as well.
- Janelia Research Campus (opened 2006, Ashburn, Virginia): Wikipedia, "Janelia Research Campus", attributed.
- AI@HHMI, the ten-year $500 million commitment (announced August 12, 2024): HHMI news release.
- The in-house investment office structure and perpetuity language: HHMI's page on CIO Donald Koch and its FY2024 audited financial statements.
- Erin O'Shea as president since 2016 and first woman to lead HHMI: HHMI leadership page and Harvard Gazette.
- The February 2025 end of the $60 million Inclusive Excellence program affecting 104 institutions, and the 2021 $2 billion inclusion commitment: STAT News and Nature; the fellowship-competition pause is stated in the 990's own Part III program narrative.
This post is informational and historical, not financial or investment advice. All figures are reproduced from the cited public filing and the Institute's own publications. Individuals are discussed from the public record as nominative fair use, with no affiliation implied and nothing endorsed by the Howard Hughes Medical Institute. Descriptions of investment allocation and strategy characterize the filing's schedules, not recommendations.
