The VanEck Semiconductor ETF (SMH) started 2026 at $360 and closed at $638 on June 3, 2026 — a 77.13% gain in just over five months (247wallst.com). A move like that normally takes years. Then the Philadelphia Semiconductor Index (SOX), a widely watched basket of chip stocks, dropped nearly 20% from its June high amid what's been called AI-related anxiety (Seeking Alpha). The swing from euphoric buying to a near-bear-market decline took weeks, not months.

Here's the setup. An ETF, or exchange-traded fund, is a single investment that holds many stocks at once. SMH and SOXX are both ETFs that hold semiconductor (computer chip) companies. You'd expect them to perform about the same. They didn't. Through mid-2026, the iShares Semiconductor ETF (SOXX) returned 169.68% compared to SMH's 135.91% over the same period (ETF.com). That's a 33.77-percentage-point gap between two funds that most people would assume track the same thing.

The reason for the gap comes down to how each fund is built. SOXX holds a wider, more evenly spread mix of chip companies. SMH puts more weight on its biggest holdings. When the largest chip stocks lead the market, SMH wins. When a broader range of companies matters more, SOXX pulls ahead. The 2026 spread suggests mid-size and smaller chip firms contributed a lot, while the giant companies that dominate SMH had a rougher ride.

The June 3 close at $638, followed by the near-20% drop, frames the core tension. The same story about AI driving demand for chips powered the surge. Now that same story is being cited as the reason for the selloff. The Seeking Alpha report points to "AI-related anxiety" — worry that chip stock prices had risen faster than even the most optimistic earnings outlooks could justify.

The bigger picture is what this means for anyone holding these funds. The question is whether the drop is a healthy pause that lets the market catch its breath, or the beginning of a longer decline where investors permanently lower what they're willing to pay for these stocks. A pause suggests the uptrend could resume. A longer decline suggests something has fundamentally changed in how investors see these companies.

The SOXX-versus-SMH gap through mid-2026 is the detail worth holding onto. A 169.68% versus 135.91% difference forces a question: what does "semiconductor exposure" actually mean in your portfolio? The two ETFs carry very different levels of risk. SMH's concentration makes the highs higher and the lows lower when its biggest stocks swing. SOXX's wider mix softens that risk but can fall behind when the giant chip stocks are carrying the sector. Picking one over the other is a choice about concentration versus spread, even though both are labeled "semiconductor."

The 77% surge and the 20% drop are a reminder that when a sector ETF shoots up that fast, it rarely comes back down slowly. The speed of the round-trip tells you something: too many investors were holding the same trade, and when they all headed for the exit at once, it got messy. For anyone holding these funds as part of a diversified portfolio, the episode shows that a sector ETF tied to a hot theme can fall just as hard as a single stock when sentiment turns.

What we know from the data: SMH hit $638 on June 3, SOX fell nearly 20% from its June high, and SOXX beat SMH by nearly 34 percentage points through mid-2026. What we don't know is whether the AI demand story will keep driving the earnings growth that investors have already priced in, or whether the June drop is the first sign that story is cracking.

In my view, the gap between the two ETFs suggests the market is already sorting among individual chip companies rather than treating them all the same. That sorting matters more than where the headline index number lands. When two funds labeled "semiconductor" diverge by 34 points, the label has stopped telling you much.