Is There Still Upside Left in Chip Equipment Stocks?

CapEx peak fears are being pushed out

The market has worried that Big Tech's AI infrastructure spending may soon peak. Recent signals are pushing that risk further into the future, although they do not eliminate it.

An internal $Meta Platforms (META.US)$ plan reportedly targets 14GW of computing capacity in 2027, following a 7GW deployment program in 2026. The exact definition of that capacity remains unclear.

Symbols

The broader Wall Street forecasts are also moving higher. BofA raised its combined CapEx estimates for $Alphabet-C (GOOG.US)$ , $Meta Platforms (META.US)$ and $Amazon (AMZN.US)$ AWS to $499 billion in 2026, $705 billion in 2027 and $816 billion in 2028.

That is about $2.02 trillion over three years, $232 billion above its previous estimates. These numbers are not official company guidance, but they show that analysts increasingly expect the AI buildout to remain larger for longer.

WFE forecasts are rising

More data centers require more GPUs, memory and networking chips. Chipmakers then need more equipment to produce that capacity.

Deutsche Bank expects global wafer fab equipment spending to rise from $110 billion in 2025 to $145 billion in 2026, $193 billion in 2027 and $219 billion in 2028.

The projected growth is concentrated in the most AI-sensitive areas. Leading-edge foundry and logic spending rises from $42 billion in 2025 to $94 billion in 2028. Combined DRAM and NAND spending increases from $27 billion to $83 billion. Lagging-edge spending remains broadly flat.

The message is constructive for equipment suppliers, but the rally has already priced in part of this growth. The next move depends on whether memory makers and advanced foundries convert expansion plans into real tool orders.

Which equipment companies benefit?

– $ASML Holding (ASML.US)$ offers the clearest exposure to advanced lithography. More leading-edge logic and next-generation memory capacity should support demand for EUV and advanced DUV systems.

– $Lam Research (LRCX.US)$ has greater sensitivity to memory through etch and deposition. More complex DRAM and 3D NAND structures require additional processing steps.

– $Applied Materials (AMAT.US)$ provides broad exposure across both logic and memory through deposition, materials engineering and wafer-processing tools.

– $KLA Corp (KLAC.US)$ benefits from inspection and metrology. Smaller and more complex structures require tighter process control to protect manufacturing yields.

Symbols

These companies do not have identical exposure. ASML is more tied to lithography, Lam to memory intensity, Applied Materials to broad process spending and KLA to defect control.

ASML is the first reality check

$ASML Holding (ASML.US)$ will report Q2 results on July 15. Investors should focus on bookings, EUV demand, memory orders and management's view of customer spending into 2027.

Strong bookings would suggest that higher CapEx and WFE forecasts are turning into real demand. Weak orders could show that equipment stocks have moved ahead of the actual investment cycle.

392K Views