US Semiconductor Equipment: $250B Reasons to Stay Long; Lifting PTs on AMAT, LRCX, & KLAC

Research evidence excerpt

US Semiconductor Equipment: $250B Reasons to Stay Long; Lifting PTs on AMAT, LRCX, & KLAC

Action |

17 Jun 2026 05:00:00 ET │ 20 pages

US Semiconductor Equipment

$250B Reasons to Stay Long; Lifting PTs on AMAT, LRCX, & KLAC

CITI'S TAKE

Atif Malik AC

We update our bull case WFE to $145B/$200B/$250B in 2026/27/28, align +1-415-951-1892

AMAT/LRCX/KLAC estimates to our revised WFE model, and lift PTs to atif.malik@citi.com

$710/$450/$290 respectively as we roll forward our price targets based on

CY28 earnings powers. Incrementally, we are bullish on NAND equipment Elizabeth Sun, CFA

demand as a widening gap between required DRAM and available supply +1-212-816-3308

helps adoption of complementary solutions such as KV cache off-loading. elizabeth.sun@citi.com

Introducing 2028 WFE of $250B — We update our top-down WFE analysis for

2026/2027, and introduce 2028 WFE forecasts based on Citi’s updated hyperscaler

capex model of 84%/56%/38% growth in CY26/27/28. We now see bull case WFE of

~$145Bn/$200Bn/$250Bn in 2026/27/28, implying another solid 25% growth in

2028. We are more constructive on 2028 WFE given continued capacity constraints

and expansion at both TSMC and memory makers, as well as recent progress at Intel

and Samsung foundries.

DRAM Bottleneck is Good for NAND Demand — The rise of agentic AI is driving a

structural increase in NAND demand as memory requirements surge and DRAM

supply tightens. Multi-step inference workflows are dramatically expanding KV

cache footprints, pushing total memory requirements well beyond what high-cost

HBM and DRAM can efficiently support, especially in an environment of constrained

DRAM supply and elevated pricing. This pressure is already driving architectural

trade-offs, as evidenced by report that Nvidia has reduced SoCAMM2 DRAM