KLA Just Printed Its Best Quarter Ever — and the Stock Is Down 47% From June
Backlog, clean rooms and DRAM intensity are all accelerating. The share price, and the gross margin, are not.
KLAC · Earnings Call · 2026-10-28
The quarter and the quote are telling different stories
On the tape of the business, KLA just had the best quarter in its history. Revenue of $4.14 billion was a record, operating margin crossed 46% for the first time ever, and December guidance of roughly $4.7 billion implies the second half of calendar 2026 grows about 25% over the first. Rick Wallace did not hedge: “we currently expect the wafer equipment market to exceed $200 billion in calendar 2027” — Rick Wallace, President & Chief Executive Officer · 2026-10-28.
On the tape of the stock, it has been a rout. The shares are down roughly 35% across the last 90 trading days — a single unbroken downtrend on the price tape — and about 47% below the late-June high. Same company. Same quarter. The gap between those two tapes is the entire story: the market is not disputing the revenue line, it is disputing what the revenue line costs to produce and how long the AI build-out can stay this vertical.
The bottleneck moved — and KLA stopped apologizing for it
The most striking thing about this quarter's keyword set is what climbed to the top. long lead time sits at rank two, backlog growth at three, and two brand-new company themes — clean rooms and new fab — cracked the top six. Those are not marketing words. They are the vocabulary of a company that has decided the constraint is no longer its own supply chain.
Wallace's answer on 2028 capacity was the tell. Instead of promising equipment, he promised coordination:
there's one customer right now that's desperate for more capacity and they can't take equipment because their cleanroom is delayed... the bottleneck moves around. Bottleneck could be construction. It could be the fact that they can't get certain raw materials or they can't get power.
That reframing — KLA as one leg of an ecosystem rather than the gating item — is new, and it echoes what the wider market is already pricing in industrial names: the AI data center complex globally has 45 advancers against 11 decliners on a 360-day view, but clean room space shows up among the global decliners over 90 days. The build-out is real; the timing of the physical shell is the wobble.
Brent Higgins put the supply side plainly: “the biggest long lead time aspect of our build of materials (is) optical components” — Rick Wallace, President & Chief Executive Officer · 2026-10-28. And Wallace was unusually candid about the year just spent: “I could absolutely have shipped more. My demand was ahead of my supply.” — Rick Wallace, President & Chief Executive Officer · 2026-10-28
DRAM is quietly becoming a logic market
The genuinely new claim on this call is DRAM. Memory is guided to roughly 60% of the business in 2027, DRAM about 84% of that, and Tim Arcuri calculated the December DRAM step-up at something like $450 million — a number he said he had never seen. Wallace's explanation: “In some products, we see intensity levels in DRAM that are similar to leading-edge logic. So the market is definitely shifting there.” — Rick Wallace, President & Chief Executive Officer · 2026-10-28
Before treating that as a fresh narrative, check the prior calls. This has been building for over a year — in July 2025 Higgins was already quantifying it: EUV insertion added about 100 basis points of process control intensity in DRAM, and HBM another 100 basis points or so on top. What changed is not the thesis but the magnitude, and management's willingness to say it out loud. That is a confirming escalation rather than a new story — and it is why the 2030 target model, built on $220 billion of WFE, now reads as sandbagging: “today it looks like a very conservative one” — Rick Wallace, President & Chief Executive Officer · 2026-10-28.
The margin that will not heal
Here is where the equity market's skepticism is rational. December gross margin is guided to 62.5% — still below the 63–64% target range laid out at Investor Day, and still below the mid-60s the process control business itself used to earn. The culprit is unchanged and has been for at least three quarters: memory costs inside KLA's own systems. In July, Higgins called it “somewhere around 100 basis points... probably a little bit more than that” — Bren Higgins, Chief Financial Officer · 2026-07-28, and added that it likely continues through 2027.
Asked directly when process control gets back to the mid-60s, Wallace gave a timeline that lands past the visible horizon: “I could see it trending there as we move into the second half of '27” — Rick Wallace, President & Chief Executive Officer · 2026-10-28. The structural reason is that KLA cannot reprice what is already ordered — “It's pretty hard to go back to your customers after you've taken orders and start to change prices on those orders” — Bren Higgins, Chief Financial Officer · 2026-07-28. With a gross margin stuck near 61%, a 25% half-over-half revenue ramp buys less incremental profit than the headline implies, and the 40–50% incremental operating margin framework is already being outrun rather than levered.
The balance sheet offers no offset: revenue of $3.7 billion in the last filed quarter sits well below the number just reported, and a price-to-earnings multiple above 80x leaves no room for a margin that arrives in 2027's back half.
What to watch
Two things would resolve the tension. First, does the memory-cost headwind actually turn into a tailwind as new product platforms carry new pricing — the path management has promised for four quarters running. Second, does the clean-room bottleneck stretch the cycle rather than truncate it. Wallace argues the latter: delays in shells and power mean "it's going to prolong the rate of growth." If he is right, the drawdown is an entry point into an elongating cycle. If he is wrong, KLA has guided to a $200 billion-plus market while telling you its own margin will not normalize until well after the market has already decided what that is worth.