The thesis
“Semiconductor stocks will keep beating the S&P 500 because of AI demand.”
How buffet read it
$10,000 invested in SMH (VanEck Semiconductor ETF) on 1 January 2015.
Held through 30 August 2026 with no changes.
Dividends reinvested.
Benchmark: S&P 500 (SPY).
SMH · 2,931 trading days · benchmark SPY
$10,000 became
$0
The same money in the S&P 500 would have been $45,379 over 2015–2026 — the thesis came out ahead by $182,307.
The worst stretch
−45.3%, December 2021 to October 2022 — you would have sat through ten months of that, and it took until July 2023 to get back to level.
the S&P 500 fell 33.7% at its worst over the same period. The thesis moved 32.1% a year against 17.6% for the index.
Where the thesis was wrong
SMH beat SPY by $182,307 over 2015–2026, finishing at $227,686 against SPY's $45,379. The cost was volatility nearly double the index at 32.1% annual versus 17.6%, and a −45.3% drawdown from December 2021 to October 2022 that took nine months to recover. In 2022 the thesis lost −33.5% while the index lost −18.2%, and in 2018 it lost −9.1% against the index's −4.6%. The entire result depended on holding through those losses without selling, and on concentration in a single sector that happened to be the decade's standout.
Counter-test
Same money, same schedule, parked in 1–3 month T-bills instead: $12,507.
Year by year
| Year | The thesis | SPY |
|---|---|---|
| 2015 | −0.1% | +1.3% |
| 2016 | +35.5% | +12.0% |
| 2017 | +38.5% | +21.7% |
| 2018 | −9.1% | −4.6% |
| 2019 | +64.5% | +31.2% |
| 2020 | +55.5% | +18.4% |
| 2021 | +42.1% | +28.7% |
| 2022 | −33.5% | −18.2% |
| 2023 | +73.4% | +26.2% |
| 2024 | +39.1% | +24.9% |
| 2025 | +49.2% | +17.7% |
| 2026 | +53.6% | +13.4% |
What this verdict does not include
- US-listed stocks and ETFs only, priced on end-of-day closes.
- No tax, no trading fees, no slippage, no bid-ask spread.
- Dividends are reinvested, because closes are adjusted for them.
- Survivorship is not corrected for: only instruments that exist today are held.
- Only 12 years of history exist for this. Read it accordingly.
- The forward-looking claim ('will keep beating') cannot be tested; this backtest shows only what already happened.
- SMH is used as the single broadest semiconductor ETF in the catalogue; SOXX is a close alternative with nearly identical holdings.
- No timing rule tied to AI demand signals was applied — this is a straight buy-and-hold.
Keep this thesis
buffet re-runs it once a month and mails you the new number. Nothing else is ever sent.
Another thesis
Every verdict names something it could not test. That gap is usually the next thesis.
Run another thesis — $5