Soros Fund Made Intel Its Largest Holding One Quarter Before the Stock Halved

Soros Fund Management reported 2,830,000 shares of Intel (NASDAQ:INTC) as its largest common-stock holding on September 30 — a stake worth $340,250,900 at quarter-end that has since lost roughly half its value. The fund's third-quarter 13F landed at 4:05 p.m. ET on Friday, November 13, 2026. Intel closed that same day at $59.66, down 2.52% and 50.4% below its September 30 close of $120.23.

The story matters beyond one fund's misfortune. It shows how quarterly disclosure works — and fails — as a signal: a 13F (the quarterly list of U.S. stock holdings that large managers must file) reveals September positions only in mid-November, long after the market has repriced them. It also captures the larger fault line in AI hardware investing, where server demand is real but capital spending is swallowing the cash it generates.

What did the filing actually disclose?

The Intel position grew from 100,000 shares at June 30 to 2.83 million at September 30 — roughly 28 times larger. It represented 4.0% of an $8.51 billion portfolio spread across 263 positions.

Two lines in the filing were larger than the Intel stake, and both were put options — contracts that gain value when prices fall:

• $670.2 million of puts on the Energy Select Sector SPDR ETF (NYSEARCA:XLE)
• $327.5 million of puts on the SPDR S&P Oil & Gas Exploration & Production ETF (NYSEARCA:XOP)

The fund, run by CEO and chief investment officer Dawn Fitzpatrick, also cut its NVIDIA (NASDAQ:NVDA) stake by 19% to 858,822 shares, worth $196.1 million. It held puts covering 350,000 NVIDIA shares — a figure that counts the shares the contracts control, not the cash the fund committed. NVIDIA fell 2.01% in the same November 13 session.

There is an important caveat. The filing values the Intel stake at the September 30 closing price, which can differ sharply from what the fund actually paid. The document does not show when the shares were bought, what price the fund paid, or whether it still holds them at all. The reported loss is therefore an assumption, not a confirmed fact.

How large is the damage?

If the fund still holds every share, the stake is now worth about $168.8 million — a decline of roughly $171 million in six weeks. That loss hurts, but it is small next to Intel's $317.09 billion market value, and smaller still against the fund's overall book.

Other large holders sat in the stock at similar prices. Primecap Management reported 77,261,208 Intel shares worth $9.29 billion, its largest single holding. NVIDIA's own connection to the story is its $5.0 billion equity stake in Intel — a modest position for a company worth $3.932 trillion.

The takeaway in plain words: one fund's bad quarter, but a cluster of large holders caught at the same prices.

Why did the thesis break?

The fund was, in all likelihood, buying into AI-driven server demand and hopes for Intel's foundry business — the bet that the chipmaker's manufacturing arm would ride the same wave as its rivals' chips. The third quarter initially validated that idea. Revenue rose 21% to $16.5 billion, and adjusted EPS (earnings per share) of $0.52 beat estimates.

The guidance broke the trade, not the results. Intel forecast fourth-quarter revenue of $15 billion to $16 billion against analyst expectations of roughly $17.07 billion. Its adjusted EPS forecast of $0.25 missed the $0.43 analysts expected.

The CFO added a harder warning: Intel probably will not generate positive free cash flow next year, because it is paying for three manufacturing processes at once. On October 30, the stock fell 28.6% to $65.76 — its worst day in more than 50 years.

Consider the price context. Every Intel close during the third quarter fell between $81.88 and $127.39. The stock now trades below that entire range, meaning even the fund's best-case entry price sits underwater.

What should Intel investors track next?

The bull case and the bear case now rest on the same facts. Server demand is real, and the third quarter proved it. But Intel faces a year of heavy spending with no free cash flow, management expects PC units to fall roughly 10% in 2027, and the foundry lost $1.9 billion last quarter alone.

One signal cuts in Intel's favor: the company's chief executive bought shares after the record drop, which is a real show of confidence from someone with inside knowledge of the order book. But insider buying is a sentiment indicator, not a guarantee — executives have been wrong before at turning points.

The practical warning for anyone tempted to copy the trade: 13F mimicry is structurally risky because you learn about September positions only in November. Soros Fund Management may have already exited, trimmed, or hedged its stake; the filing cannot tell you.

Where does NVIDIA fit?

NVIDIA sits on materially stronger financial footing. It trades at about 33 times trailing earnings with a 55.60% net margin — it already generates the cash Intel is still hoping to produce. That contrast is the quiet argument behind the Soros fund's own moves: it trimmed NVIDIA by 19% yet held puts on 350,000 NVIDIA shares, hedging exposure rather than abandoning the sector.

Two dates and levels now frame the story. NVIDIA is scheduled to report earnings on Tuesday, November 17. And an Intel close back above $65.76 — the October 30 panic level — would be the first evidence that the selloff overshot the fundamentals rather than accurately pricing them.