© phongphan5922 / Getty Images
XSD is down about 8% over the past month, yet up roughly 52% over the past year. That gap tells the whole story of what semiconductor investors are wrestling with right now: a structural AI demand boom running headlong into near-term supply friction and macro uncertainty.
SPDR S&P Semiconductor ETF (NYSEARCA:XSD) takes a different approach than most semiconductor funds. Rather than concentrating exposure in the largest names, it uses an equal-weight methodology across 43 holdings, giving smaller and mid-cap semiconductor companies the same portfolio weight as giants. That design choice is both the fund’s appeal and its primary source of risk, because it means a struggling turnaround story carries just as much influence as a high-growth AI pure-play.
The Macro Force That Will Define the Next 12 Months
The single biggest macro factor for XSD is the pace and durability of AI infrastructure spending. Hyperscalers are committing enormous capital to data center buildouts, and memory and logic chips sit at the center of that investment cycle. Micron’s numbers make this concrete: Q1 FY2026 revenue reached $13.64 billion, up 57% year-over-year, with Cloud Memory Business Unit gross margins hitting 66%. Management has guided Q2 revenue to $18.70 billion, and order books are reportedly stretching into 2027.
Marvell’s data center revenue tells a similar story. Data center revenue reached $1.52 billion in Q3 FY2026, representing 73% of total revenue and growing 38% year-over-year. CEO Matt Murphy stated that “our data center revenue growth forecast for next year is now higher than prior expectations.”
Cycle risk is real. Qualcomm already flagged it: industry-wide memory supply constraints are weighing on near-term handset demand, a ripple effect from the same AI-driven memory demand surge. If hyperscaler capex commitments slow or memory supply tightens further across end markets, XSD’s broad semiconductor exposure amplifies the downside.
Capital expenditure disclosures from Amazon, Microsoft, Google, and Meta are the clearest leading indicators for XSD’s AI-exposed holdings. The BLS monthly jobs report also matters as a proxy for consumer electronics demand, which feeds the non-AI side of many XSD holdings.
The ETF Mechanic That Could Surprise You
XSD’s equal-weight structure is the most important fund-specific factor to understand. Every holding receives roughly the same allocation regardless of market cap or growth profile. That means Intel, which posted a $2.51 billion Intel Foundry operating loss in Q4 2025 and guided Q1 2026 to non-GAAP EPS of $0.00, carries nearly the same weight as Marvell, whose custom AI silicon pipeline is at an all-time high.
Currently, Micron is the largest single position at 4.21% of the portfolio, while Marvell sits at 2.25% and Qualcomm at 2.07%. That spread reflects recent performance divergence, but the equal-weight methodology means the fund rebalances periodically, trimming winners and adding to laggards. When Micron outperforms sharply, as it has over the past year, XSD automatically reduces that exposure at the next rebalance.
This creates a structural drag in trending markets: the fund sells what is working and buys what is not. Intel’s ongoing foundry losses represent a real cost to the portfolio that a market-cap-weighted fund would naturally minimize. SPDR’s quarterly holdings files at the State Street ETF website and index reconstitution notices signal when names are added or removed entirely, which can shift the fund’s sector tilt meaningfully.
Intel’s Foundry Losses Are the Swing Factor
If hyperscaler capital spending holds at current levels through mid-2026, XSD’s AI-exposed holdings should continue generating strong earnings, supporting the fund’s valuation. The more immediate risk is Intel’s foundry drag under equal weighting: if the Intel 18A ramp delivers meaningful external customer wins and operating losses begin narrowing, that single holding could shift from headwind to tailwind for the entire fund.
