KLA Raises WFE to $150B for the Fourth Time — and the Tape Sells the Entire AI-Infra Cluster

Record June quarter, a raised advanced-packaging guide, and a 20% second-half ramp meet a ~30% drawdown in the HBM/data-center tape KLA depends on

KLAC · Earnings Call · 2026-07-28

KLA's June quarter was the kind of print management dreams about: record revenue of $3.66B (+7% QoQ, +15% YoY), EPS at the upper end of guidance, operating margin of 43.7%, and yet another raise to an already-aggressive wafer equipment market outlook. The reward has been anything but typical, though. The company and the entire AI infrastructure process-control cluster are being de-risked hard — KLAC's tape is down roughly 30% on a split-adjusted basis since mid-June, and the 30-day global mover list is a wall of red for nearly every theme this call celebrated.

Four Raises, Now $150B of WFE

The headline from Bren Higgins was unambiguous: the 2026 wafer-equipment market is now expected at "approximately the low $150 billion range... up from our prior expectation of $140 billion plus," implying mid-20% growth off 2025's ~$120 billion. It's the fourth upward revision to this year's outlook, and it comes with rare precision. Rick Wallace reported that since March's Investor Day "“demand signals across AI infrastructure have strengthened materially, supported by accelerating hyperscaler data center investment, rising AI compute requirements” — Richard Wallace, Chief Executive Officer · 2026-07-28" — and Bren bracketed a second-half 2026 ramp of ~20% over the first half, a step-change from the "high teens" framing offered only last quarter, when Bren allowed "“getting the high teens... gets you into the 15-ish range.” — Bren Higgins, Chief Financial Officer (CFO) · 2026-04-29" The 2027 outlook is bolder still:

Given the unprecedented visibility from customers, we continue to plan for significant growth in calendar 2027 as broad-based investment across leading-edge logic, foundry, DRAM, both conventional and HBM, NAND and advanced packaging drives continued capacity expansion.
Bren Higgins, Chief Financial Officer · 2026-07-28

Management cited a consensus view of ~$190 billion of WFE in 2027 — a mid-20s growth year matching 2026 — and said it is already sizing supply to the more bullish scenarios. A quarter ago the company's own IR head declared "“there's no question '27 is going to be a massive buildup.” — Kevin Kessel, Vice President of Investor Relations and Market Analytics · 2026-04-29" The thesis hasn't so much changed as hardened.

Packaging, HBM, and a Memory-Price Drag

Three things are genuinely new in this call. First, advanced-packaging process-control revenue is now guided to ~$1.1 billion, up more than 70% — above the prior "high 50s" view and ~2x the packaging market — with hybrid bonding and SoIC (die-to-wafer bonding) finally inflecting as the "customers pulling us into that" story Rick described, putting front-end-class systems into the back end and central to KLA's process control intensity narrative. Second, high performance compute is now flowing into Specialty Process, PCB and Component Inspection — the Orbotech strategy paying off with those combined products up >25% this year. Third, the memory pricing headwind has worsened: Bren put it at "“somewhere around 100 basis points. It's probably a little bit more than that,” — Bren Higgins, Chief Financial Officer · 2026-07-28" and expects it to likely continue as they move through 2027. KLA won't reprice existing backlog — "it's pretty hard to go back to your customers after you've taken orders" — so the drag is absorbed into 2027, cushioned by new-product launches that reset pricing. Services delivered $820 million, +17% YoY, near the top of a new 13–15% long-term model, and Bren expects RPO (backlog) of ~$12.5 billion when the 10-Q lands — a "nice anchor" of forward visibility. On the fundamentals, gross margin sits at 61.1% with a 15-year uptrend of roughly +2pp, but the fundamentals file (10-Q dated 04/30) predates the June quarter's 62.4% print and the 62.5% guide — operating leverage is the stated plan to offset the drag.

The Tape Disagrees

Here's the tension: the 30-day global tape reads like a checklist of everything KLA just reaffirmed, all in red. HPC data centers are down 21.6%, high-bandwidth memory -17%, "AI data centers" -10.1%, data center AI -9.5%, and gig-per-lane optics down double digits. KLA's raw 90-day series shows a -89% headline off the June 11 split (a 10-for-1 split), but the underlying eight-week leg is a real ~29% drawdown from the June 9 peak — consistent with the cluster, not the call. The market is effectively pricing a peak in the AI-capEx cycle, or at least demanding proof of that ~$190B 2027 consensus before paying up. On pre-crash fundamentals the stock was far from cheap — price-to-free-cash-flow was ~52x — so the selloff reads as a brutal multiple conversion on a valuation that had raced ahead. Rick's closing quip — "We are really good at creating systems that are valuable to our customers... We are not very good at forecasting" — captured the moment: the company has never been more confident, and the tape has never been more skeptical. One of them will be spectacularly wrong.