Nvidia and Apple Now Hold More of the S&P 500 Than Any Two Stocks in History — and the Index Has No Hedge Left

The arithmetic is brutal in its simplicity. At their September 24 trading levels, Nvidia carries a $5.46T market capitalization on roughly 24.15 billion shares outstanding and Apple carries $4.95T on roughly 14.69 billion shares — together about $10.4T of equity value inside an S&P 500 with a total market cap of roughly $50T to $55T. Both companies individually now exceed the entire weight of the energy or utilities sector in the benchmark, and their combined cap weight runs above 15%, the highest two-stock concentration on record according to Yahoo Finance's tally of Mallouk's data. To set the bar: the previous record, Microsoft plus General Electric near the March 2000 peak, topped out at 9.1%.

Apple crossed the $5 trillion threshold intraday on July 28, 2026, the second company to do so after Nvidia briefly crossed it on April 24. By September 23, Apple's market cap stood at $4.96T per stockanalysis.com's market-cap tracker, and Nvidia's at $5.45T per the same source. Their combined enterprise value exceeds the GDP of every country outside the United States and China. This is not a normal concentration episode — it's a structural compression of the cap-weighted index into a binary AI-capex bet, and the comparison to 2000 is no longer rhetorical.

Passive index ownership has compressed toward a small handful of mega-cap stocks because of a mechanical loop that historyofmarket.com flags directly: when seven stocks make up 33.5% of an index, an equal-weighted $5,000 allocation across the S&P 500 puts roughly $1,675 into those seven names — but a cap-weighted allocation puts far more than that, and a cap-weighted $5,000 allocation is precisely what flows into the largest passive ETFs. The Vanguard S&P 500 ETF (VOO) is now the world's largest ETF at $1.76T total net assets as of August 31, 2026, having passed $1.0 trillion on June 2. Every dollar that arrives at VOO automatically allocates roughly 7.50% to Nvidia and 6.58% to Apple, per the August 21 fund holdings.

BlackRock's iShares franchise crossed $6 trillion in ETF assets during the second quarter of 2026, per MarketsMedia, roughly doubling in three years. With State Street's SPDR family and Vanguard's VOO, the three passive vehicles above now hold about $3.46T of S&P 500 exposure in forms that mechanically propagate the two-stock weight every time a 401(k) contribution lands. The structural implication — one BlackRock acknowledged directly in its 2025 SEC comment letter on indexing — is that the index has effectively no way to redistribute weight away from Nvidia and Apple until cap-weighted flows reverse, which is precisely what Michael Burry has been warning about since he called the tape \"Trump's market\" in his August Substack post.

On September 16, 2026, the FOMC voted 12-0 to raise the federal funds target range by 25 basis points to 3.75%–4.00% — its first increase since 2023, per the Federal Reserve's own statement. New Chair Kevin Warsh, confirmed by the Senate on May 13, 2026 and promising a \"regime change\" to defeat inflation when he took over in January, delivered exactly that. The summary of economic projections showed headline PCE at 3.7% and core at 3.4% for 2026, both 0.1 percentage point higher than the June projections, with inflation not seen returning to target until 2029. The dot-plot showed 16 of 18 participants expecting another hike; four saw two more as possible.

Burry himself — who in May compared the tape to the last months of the 1999-2000 bubble and warned of \"a 1987-type fall\" on August 10 — quietly started trimming risk. A September 10 disclosure showed Scion exiting its December 2026 put options on Nvidia and Palantir without rolling them over. The bearish voice who most loudly tied late-cycle concentration to AI-capex euphoria is now reducing his short book at the same time the Fed is doing the same thing on rates.

The clearest evidence that the concentration risk is now being actively traded sits in a single ETF: the Invesco S&P 500 Equal Weight ETF crossed $100 billion in assets on September 1, 2026, per Invesco's announcement. RSP has absorbed more than $12 billion in net inflows year-to-date, and the asset growth has come alongside a 3-percentage-point year-to-date outperformance versus the cap-weighted iShares Core S&P 500 ETF through August 21. Investors voting with their wallets decided that the way to play a 15% two-stock index is to NOT be index-weighted.

Three binary events will move the two-stock concentration debate before the end of the year. First, Nvidia's next earnings — fiscal Q3 2027 results in late November — will be the single most important data point for the cap-weighted index, given that Q2 FY27 revenue came in at $96.22B with net income of $59.69B per the company's August 26, 2026 10-Q filing. A cap-weighted beat sustains the concentration; a cap-weighted miss is the cleanest path to broader rotation. Second, the October 28-29 FOMC will reveal whether Warsh's dot-plot signal of \"16 of 18 expecting another hike\" is genuine — Napier's detailed commentary on the Jackson Hole \"inflation is a choice\" framing suggests it is.

On a 1–3 year horizon, the structural question is whether the AI capex cycle that drove Nvidia's TTM revenue to $302.97B and net income to $192.88B — implied annualized growth of roughly 56% from the $130.5B FY25 base — can sustain a 30%-plus operating margin (FY27 TTM operating margin of 65.2%, per Nvidia's company overview) through the Warsh-era higher-rate regime. Nvidia trades at 28.4× trailing earnings and 23.8× book, and Apple at 38.5× earnings — multiples that price in continued earnings growth at materially above-market rates. A flattening of the AI capex curve in 2027, with Warsh still tightening, would mark the first time in this cycle that the cap-weighted S&P 500 faced a concentration unwind without a Fed put to lean against.