Global Semiconductor Equipment Sales Forecast to Reach a Record $229 Billion in 2028, SEMI Reports

SEMI forecasts global semiconductor manufacturing equipment sales to hit a record $165.9B in 2026 (+23.2% YoY), then rise to a record $229.5B by 2028 over five straight years of growth. Growth is being driven by AI-fueled investment in leading-edge logic, advanced memory (notably HBM), and back-end test/packaging, with WFE expected at $143.9B in 2026 (+23.1%). The outlook revisions are particularly strong in test equipment (projected +31.0% YoY to $15.3B in 2026, after a +55.3% surge in 2025).

Analysis

This is bullish for the semicap complex, but the market mechanism is breadth, not just magnitude: the demand mix is shifting toward higher-content nodes where test, metrology, advanced packaging, and memory process steps carry more dollars of equipment per wafer. That should keep multiple dispersion wide — names with real exposure to HBM, inspection, and packaging intensity should outperform the “front-end-only” beneficiaries that already trade as AI proxies.

The cleaner second-order winner is not the chipmakers, but the tooling ecosystem that monetizes complexity. If HBM and 2nm ramps continue, the bottleneck moves from wafer starts to yield, test time, and heterogeneous integration, which tends to support revenue durability for Teradyne, KLA, Onto Innovation, and Amkor-like package content over a longer cycle. That said, the forecast also implies more capex, not necessarily better end-demand; the industry can overspend into 2027 and still create a later pricing air pocket, especially in memory.

Contrarian risk: this may be less incremental than it looks because AI capex optimism is already embedded in SMH/SOXX and in the highest-quality equipment names. The near-term falsifier is any commentary from TSMC, Micron, Samsung, or SK hynix implying that 2026 budgets are being pulled forward rather than expanded organically. Watch China exposure carefully: if China spending merely normalizes while the rest of the world offsets, headline growth can hold but order quality and margins may degrade, which would matter most over 6–18 months, not the next few sessions.

Key Decisions for Investors

• Buy dips in SMH or SOXX over the next 1-2 weeks; use a 3-6 month horizon and treat any post-news pop as confirmation rather than chase. Best risk/reward if semicap ETF pulls back 3-5% while AI capex headlines remain firm.
• Rotate inside semicap toward quality secondary beneficiaries: long TER/KLAC/NVMI versus a weaker basket of broad industrial cyclicals over 1-3 months. Thesis: yield, test intensity, and packaging content should compound faster than generic capital goods exposure.
• Use a pair trade: long TER, short MU, for 3-6 months if you want to express rising test/packaging intensity without paying for the full memory supply cycle. Falsify if memory ASPs and HBM lead times keep tightening into the next quarter.
• If already long ASML/AMAT/LRCX, trim 20-30% on any strength and re-add only after Q3 capex commentary from TSMC/Samsung/Micron confirms the 2026 budget path. The risk is multiple compression if the market decides the forecast is already fully discounted.
• No direct actionable read-through to AERA or KEP from this item; keep them neutral unless there is separate evidence of capex or electronics supply-chain linkage.