A $10,000 stake in Goldman Sachs Future Tech Leaders Equity ETF (NASDAQ:GTEK) on the last trading day of 2025 was worth about $15,200 at Monday’s close, with the fund up 52% year to date through June 1, 2026. The same $10,000 dropped into the S&P 500 via the SPDR S&P 500 ETF Trust (NYSEARCA:SPY) over the identical window grew to about $11,100, with the benchmark up 11%. The headline ratio rounds to roughly 4.6 to 1 in GTEK’s favor, and that is on a price basis with no help from cherry-picked entry dates.
For an active tech ETF that intentionally skips the Magnificent Seven, that is the kind of run that gets noticed. It is also the kind of run that demands an honest look at what produced it, because the mechanism here is unusually specific and the conditions sustaining it are unusually fragile.
The Arithmetic, Without the Mag 7 Crutch
GTEK closed at $61 on June 1, 2026, up from $40 on December 31, 2025. Over the trailing twelve months the fund is up 83%, and over the trailing month alone it gained 16%. Most of the YTD gap opened in the last six weeks, which matters when you try to project the run forward.
ETF Database flagged GTEK as "up just under 26% YTD" on May 1, 2026, and the fund has roughly doubled that figure in the five weeks since. The acceleration is what matters here, because as recently as September 2025 the fund was up 19% YTD versus 15.9% for the Vanguard Information Technology Index Fund, a respectable but unremarkable spread for an active tech sleeve charging 75 basis points.
What Actually Did The Work
GTEK’s prospectus mandate is to invest in U.S. and non-U.S. technology companies that the Investment Adviser believes are driving technological innovation or benefitting from the enablement of technology. In practice, the portfolio managers translate that into a hard ceiling. The fund actively invests in innovative global tech firms with market caps under $100 billion and deliberately avoids megacap names to reduce concentration risk. Where most tech ETFs are 30% to 50% NVIDIA, Microsoft, Apple, and Alphabet by weight, GTEK is structurally none of those.
That design choice is the entire mechanism. In the first half of 2026, capital that had been crowded into the Mag 7 for two years began rotating into the second tier of AI beneficiaries, the so-called picks-and-shovels names that sell into hyperscaler capex without being hyperscalers themselves. Goldman Sachs Asset Management’s own 2026 outlook flagged this trade directly, writing that in the small and mid-cap space the firm sees potential opportunities among enablers, so-called "picks and shovels" of the AI boom, including companies on the front line of AI innovation. GTEK has been long that thesis since well before it worked. Known holdings include Cadence Design Systems and Snowflake, with earlier positioning in Palo Alto Networks, Zscaler, and Workday.
The S&P 500 is up 11% YTD precisely because the Mag 7 has stalled. GTEK is up 52% because the capital that would have gone into Mag 7 marginal-buyer demand went somewhere, and a chunk of it landed in exactly the names GTEK was already holding. The fund was built to benefit from exactly this kind of rotation whenever it arrived, and in 2026 it arrived.
The Capex Question Hanging Over Everything
The Goldman outlook is also where the forward look gets uncomfortable. The firm asks plainly whether AI-fueled growth can compensate for underlying economic and labor market weakness, noting that slowing consumer spending and a stagnant housing market have been counterbalanced by AI capex driving business and investment activity. Their own framing is that this is an uneasy equilibrium, and that a marked reversal and broad unwind of AI-related investments could precede a hard landing.
GTEK is the cleanest equity expression of that equilibrium holding. A diversified tech megacap can absorb a capex pause by leaning on cloud contracts already booked, ads already sold, and devices already shipped. A sub-$100 billion enabler often cannot. Cadence sells design tools into a customer base that is currently building chips as fast as fabs allow. Snowflake monetizes data workloads that scale with AI deployment. If hyperscaler capex flattens in 2027, the GTEK book reprices first and hardest, in the same direction and with the same speed it repriced upward in May.
What To Watch Going Forward
Three observable indicators tell you whether the GTEK setup is intact or breaking. The first is hyperscaler capex guidance, reported quarterly by Microsoft, Alphabet, Amazon, and Meta. As long as the four largest buyers of AI infrastructure are still raising the number, the demand pipeline for GTEK’s holdings is still expanding. The moment that number flattens for two consecutive quarters, the rotation thesis loses its fuel source. The second is the spread between an equal-weighted S&P tech index and the cap-weighted version. The Mag 7 unwind that powered GTEK in 2026 shows up there before it shows up in the headline ETF prints. The third is credit spreads on AI-oriented issuers, which Goldman highlighted as more active in the corporate bond market in 2025. When the cost of financing AI buildouts widens, the marginal project gets cut, and the marginal project is what GTEK’s holdings sell into.
Institutional positioning has been adding to the run. Black Swift Group increased holdings by 13.6% in Q3, now owning 156,655 shares valued at $6.11 million, with Jane Street, Raymond James, and Goldman Sachs Group also on the holder list. Short interest jumped 81.1% in February to 21,489 shares, which is a tiny absolute number but a directional tell that some traders are starting to position for the unwind rather than the continuation.
The honest read is this. GTEK’s 4-to-1 beat over the S&P 500 in 2026 was earned by being structurally on the right side of a specific rotation, and the rotation is real, and the holdings are good businesses. The fund is also now a much more expensive way to own that thesis than it was on January 1, and the conditions that produced the run are the same conditions that would produce a sharp reversal if AI capex guidance softens. Watch the hyperscaler capex prints. That is the indicator that matters.