Benchmark Sells 2030 Visibility While Its HPC Catalyst Slips to 2027

A record $811 million quarter and a raised guide collide with a 26% drawdown — and the next-gen HPC program has now moved right for the fourth call in a row

BHE · Earnings Call · 2026-10-28

Record Quarter, Broken Tape

Benchmark just delivered the best print in its history and was paid for none of it. Revenue of $811 million was a company record, up 19% year over year; non-GAAP EPS of $0.84 rose 35%, nearly twice the rate of sales; and management raised, guiding the fourth quarter to $815-855 million and calling for full-year revenue above $3 billion, roughly 16% growth. The equity, meanwhile, closed 25.7% below its June 30 peak of $98.67 and is down 19.3% over the last ninety sessions. That gap — an accelerating business inside a de-rating tape — is the whole story, and it has two halves: one macro, one company-specific. The June quarter printed $777 million of revenue, up 17% year over year, before September's $811 million. Operating leverage is real but still thin in GAAP terms: GAAP operating margin was 3.9% in the June quarter, still shy of the 2024 high — which is precisely why the non-GAAP 5.5% print carries so much narrative weight.

From Recovery to Durability — On a Customer Forecast

The most consequential new language on the call was about Semi-Cap duration. “Following this year's event, the message has shifted from recovery to durability. Industry participants are now discussing demand visibility extending into 2029 and 2030.” — David Moezidis, President and Chief Executive Officer · 2026-10-28 Semi-Cap revenue grew 30% year over year and 8% sequentially, and management expects that to accelerate again in Q4 and then keep growing through 2027. Advanced Computing & Communications grew 50% year over year on an on-prem AI ramp that ran hot enough to force a manufacturing footprint expansion mid-flight, with PT4 in Penang live since July and a Thailand build due late 2027. But the load-bearing word in this quarter's vocabulary is customer forecast — the top-ranked, highest-momentum keyword in Benchmark's own Q3 lexicon, and a phrase that did not exist in its prior-quarter list. It arrived because the very first analyst question went straight at the weak seam: how do you get confident about 2027 outside of your customers' own forecasts? The answer was forecasts, ramps already in flight, and conversations. “The durability comes straight from our customer base. Words like big and sustaining comes from our customer base.” — David Moezidis, President and Chief Executive Officer · 2026-10-28 That is a share-taker's answer from a company that has been winning Semi-Cap programs since the 2023 downturn, and it deserves credit — but it is also soft evidence dressed as visibility. The confirming evidence is external, not internal. In the current reporting window, the toolmakers themselves are making the same claim in their own words: KLA's process control franchise language, Teradyne on higher stack heights and AI accelerators, Amkor on advanced packaging capacity, ASM on high-performance DRAM for HBM. When the capital-equipment vendors and the outsourced manufacturer converge on the same durability story, it is a sector signal, not a Benchmark quirk. The trouble is the tape disagrees: AI data center is one of the heaviest decliners in the 90-day window, with six advancers against fifty decliners, and adjacent optics and memory themes like High bandwidth flash and PCIe Gen 6 are down roughly 7% over the same stretch. Benchmark's business is accelerating while the equity complex that owns its end market is de-rating. That is a multiple problem, not an earnings problem — and it explains the drawdown better than anything said on the call.

The HPC Slip That Never Lands

Here is the genuine company-specific change, and it is a deferral dressed as a housekeeping note. Next-generation HPC, the sovereign and government-scale program management has flagged for over a year, has moved again — and this time it has been removed from the numbers entirely.

Our current outlook assumes no contribution from next-generation HPC opportunities, which today appear more likely to begin contributing mid to late 2027.
David Moezidis, President and Chief Executive Officer · 2026-10-28

Set that against the same executive six months ago: “as we exit the year and enter 2027, HPC is going to actually start picking up on its own and contributing nicely as well” — David Moezidis, President and Chief Executive Officer · 2026-04-30 — and three months ago: “we expect HPC to start picking up very late in Q4 and into 2027” — David Moezidis, President and CEO · 2026-07-29. The program has now slipped across four consecutive calls, tracked in this company's own keyword history as next generation HPC opportunities. The bullish reading is the honest one: Benchmark ripped up the crutch and still raised. Q3 beat, Q4 guides to 19% growth, and AC&C is compounding without the marquee program. The bearish reading is that a promised catalyst has quietly become a 2028 event, and the market is pricing the difference.

What Management Stopped Saying

Aerospace & Defense is the tell. For several quarters, space and defense were the loudest enthusiasm in the deck — space applications led this company's keyword table in early 2026 with bookings momentum right behind it. This quarter A&D was down 1% year over year, and the explanation was annual growth delivered over the prior two years that now has to pause; growth is once again promised for 2027. That is the third consecutive call where the defense and space reacceleration lives in the next fiscal year. Record bookings are real, but the conversion keeps sliding. Medical, by contrast, is the delivered promise, and it should get more credit than it does: “we feel that we've turned the corner as the inventories were starting to clear up in our customers' channels” — David Moezidis, President and Chief Commercial Officer · 2025-11-04 was the claim a year ago, and medical is now on pace for double-digit growth with the largest single piece of new business awarded in Q3 going to a new medical device customer. Management also kept talking about supply chain, for the third straight quarter — memory first, now spilling into ASIC supply, with extended lead times being managed through joint inventory investments with customers rather than through price. The balance sheet gives them room to do it. Cash conversion ran at 54 days, twenty-three better year over year, on a cash conversion cycle that has been the most durable operational theme in this company's entire keyword history, and the company sits on roughly $133 million net of debt. Inventory rose 16% sequentially — faster than revenue — but that is ramps at eleven sites, not sloppiness, and turns are meant to settle back toward the five-to-five-and-a-half target. Two things temper the leverage narrative. Q4 guidance implies an operating margin of 5.6% at the midpoint versus 5.5% delivered — essentially flat — so the promised 1.5-to-2x earnings growth is coming from below the operating line and from a variable-compensation headwind that turns into a tailwind in 2027, not from gross margin. And the multiple that the tape is now unwinding was extraordinary: price to revenue stood at 1.2x at the June mark, up from 0.72x a year earlier. Rewind a quarter of that and you are still paying a growth multiple for a 10.5%-gross-margin assembler. So: a genuinely good business cycle, a management team that has earned credibility on medical and cash conversion, a Semi-Cap thesis confirmed by its own suppliers, and a stock that has spent four months arguing about a program that will not ship until 2027. The quarter was not the problem. The price was.