Harvard University’s endowment returned 27.1 percent in fiscal 2026, growing to about $68 billion, or roughly NT$2.2 trillion. Its disclosed U.S.-listed stock portfolio is heavily concentrated in AI-related names, led by SpaceX and TSMC.
According to the Liberty Times and U.S. media, Harvard’s endowment returned 27.1 percent in the fiscal year that ended June 30, 2026. The fund reached a record of about $68 billion, up roughly $11 billion in a year. The median return for large university endowments over the same period was about 18.9 percent.
광고 문의 · 300×250Holdings that Harvard Management Company (HMC) reported to the U.S. Securities and Exchange Commission as of the end of June (Form 13F) showed 18 listed positions worth about $4.26 billion. The Harvard Crimson said the total was up about 135 percent from the previous quarter, the largest since at least 2006.
One stock is more than half
The largest holding is SpaceX, which listed on June 12. Harvard holds 12,935,100 SpaceX shares worth about $2.21 billion, roughly 52 percent of the disclosed stock portfolio and about 3 percent of the whole endowment. When and at what price the shares were acquired was not disclosed.
Second is Taiwan’s TSMC, at about $350 million or just over 8 percent; HMC added to the position in the second quarter. Newly added AI chipmaker Cerebras (about $239 million) ranked third, followed by Amazon and Alphabet. U.S. investment outlet The Motley Fool calculated that these five stocks make up 75 percent of the disclosed portfolio.
The pros and cons of concentration
HMC also increased its Nvidia stake while exiting positions including Union Pacific and switching between gold exchange-traded funds. In crypto, it reportedly sold its Ethereum holdings and cut its Bitcoin stake.
Still, the disclosed stocks are only a small part of the $68 billion endowment, most of which sits in private equity, hedge funds and other non-public assets. SpaceX shares have swung widely since listing, so the value of Harvard’s stake could fluctuate sharply. The Motley Fool framed the AI concentration as a question: genius move or dangerous trap?