Open in interactive viewer → charts, metric popovers & call review

Hyliion's $18 Million Year Meets Its First Reliability Test

Revenue guidance rises again, but a thermal-fatigue crack and a six-week Navy slip expose the gap between demand and proof.
HYLN · Earnings Call · 2026-11-11

The Beat Nobody Bought

Hyliion's third quarter was, in CEO Thomas Healy's words, “The third quarter was a very strong 1 for Hyliion.” — Thomas J. Healy, Chief Executive Officer · 2026-11-11 Revenue came in at $7.5 million, versus $759,000 a year earlier and $4.9 million in Q2. Management said it was “once again raising our full year revenue outlook this time to $18 million” — Thomas J. Healy, Chief Executive Officer · 2026-11-11, up from $15 million last quarter and $10 million at the start of the year. On the surface, this is a company whose Power Module business is inflecting. The tape tells a different story. HYLN is down 31.4% over the last 90 trading days and 38.6% from its October 2 high of $4.53. The full-history chart is even harsher: a 95% drawdown from the 2020 peak. At a roughly $496 million market cap, the market is not disputing that Hyliion has revenue; it is disputing the slope and durability of the ramp. The last filed quarter showed $5 million of revenue, up 226% year over year, before the Q3 step-up. The gap between the income statement and the stock chart is the story: Hyliion must prove that its early adopter units become a repeatable product, not just a series of R&D milestones.

Demand Is Still Not the Problem

The military pipeline is the clearest source of visibility. Hyliion signed a $41.7 million Navy contract in July, expects two more awards worth about $1 million and $7 million before year-end, and says total 2026 military awards could approach $50 million. Add the $45.5 million of remaining authorized funding under existing contracts, and the company is pointing to roughly $53.5 million of contract backlog. That is real money for a company this size. The keyword military contract has been the top company theme before, and it remains the load-bearing wall. The newer emphasis is on the appropriations process and the shift from R&D services to actual power-system deliveries. Healy argued that the military operates numerous tens of thousands of generators, and that different branches have different pain points — fuel flexibility for forward bases, low detectability for another branch, low maintenance for the Navy. That is a credible wedge. But it is still a pipeline, not a production run. Commercial demand is similarly loud. Hyliion has “signed LOIs for about 750 units worth of interest” — Thomas J. Healy, Chief Executive Officer · 2026-11-11, which management previously sized at around $400 million of potential revenue. The prior call made the same point: “about 750, KARNO Cores worth of LOIs ... around $400 million of potential revenue opportunity” — Thomas J. Healy, Chief Executive Officer · 2026-08-12. The difference this quarter is that the conversation has moved from LOIs to on-site trials. A large data center customer spent three days at Hyliion, ran the genset, and collected its own data. That is progress, but it is still proof-of-concept, not purchase orders. The company's own prior language still applies: “backlog interest, that's not our issue. Our issue right now is how do we get these units out into the field” — Thomas Healy, Chief Executive Officer · 2026-05-13. That has not changed. The data center pitch is also getting sharper. Hyliion is emphasizing transient load handling — the ability to absorb the spiky power draw of AI workloads — and native 800-volt DC output. Those are differentiated attributes versus turbines, and they explain why data center interest has grown. But the global tape has turned against the AI data center complex over the last month, and Hyliion is not immune to that de-rating even if its military backlog gives it a second leg.

The Crack in the Reactor

The most important new disclosure was not the guidance raise. It was a reliability issue in the reactor. Hyliion found that a printed component was susceptible to thermal fatigue after extended high-power operation. A micro-crack formed during printing, grew under thermal stress, and allowed gas to pass between cavities inside the part. Management reverted to an earlier revision to keep systems moving and redesigned the print file to eliminate the crack at the source.

We identified a printed component that was susceptible to thermal fatigue following extended high power operation. We have temporarily reverted back to an earlier revision of that component ... Meanwhile, we have completed a redesign of that part that we believe addresses the root cause.

Thomas J. Healy, Chief Executive Officer · 2026-11-11
Healy later called it “a pretty contained issue that we saw with the component” — Thomas J. Healy, Chief Executive Officer · 2026-11-11, and the fix sounds technically sound. But the timing matters. The 800-kilowatt system for the Navy's USX DEFIANCE is now expected early in Q1 2027 rather than by year-end — about a six-week slip. Management says the shift does not affect the 2027 scale-up. The market may be less forgiving. Hyliion's entire value proposition rests on low maintenance and high reliability; a thermal-fatigue crack in the reactor design is exactly the failure mode that customers will test for. The fact that the company caught it, contained it, and re-engineered the print file is a good sign. The fact that it happened at all is a reminder that this is still a first-generation manufacturing process. This is not the first component learning. Prior calls dealt with linear electric motor production, regenerator mesh, depowdering, and other manufacturing issues. The pattern is consistent: Hyliion solves one process problem, then finds the next. That is normal for a hardware scale-up, but it is also why the stock does not trade on LOI math alone.

Capacity, Cash, and the 2027 Proof Point

The manufacturing math is improving. Print-speed work has demonstrated roughly 2.5x improvement on at least one component, with a path to up to 3x. Current installed capacity is 5 to 10 megawatts per year, and management thinks the existing printer base can reach 15 megawatts with hardware and software upgrades. The Austin facility, once built out, is designed for about 250 megawatts of annual capacity. The capital-efficiency framework remains roughly $1.5 million of equipment per megawatt of capacity, which at current pricing supports $2.5 million to $3 million of annual revenue per megawatt. That is a compelling model if the demand converts. But the company is deliberately not buying printers yet. Capital spending was just $212,000 in the last filed quarter, down 95% year over year, and management expects printer purchases to restart in 2027. The prior call set the same expectation: “these print speed improvements are going to enable us to start spending again on printer acquisitions” — Jon T. Panzer, Chief Financial Officer · 2026-08-12. Until those orders go in, the 250-megawatt vision is a blueprint, not a capacity number. The balance sheet buys time. Hyliion ended Q3 with $121.4 million of cash and investments and expects to finish the year at the low end of $100 million to $120 million. It did not close the equipment financing it had hoped for, and it did not use its ATM. The quarterly cash spend was $11 million, down from $20.6 million a year earlier. Operating expenses fell 15% to $13 million, and R&D was $8.7 million, down 14%. That is a healthier cost structure, but it is still a company burning roughly $11 million a quarter before it has a commercial order book.

What Changed, What Didn't

What changed: revenue guidance moved up again, the military contract total is approaching $50 million, the first commercial site deployment is weeks away, and the company disclosed a real reliability defect along with a fix. What did not change: Hyliion still has no meaningful commercial order book, the 200-kilowatt power target is still a year-end goal, printer purchases are still a 2027 event, and data center demand is still in the trial phase. That is why the stock's drawdown and the operating beat can coexist. The market is not asking whether Hyliion can generate demand; the prime power and military use cases are real. It is asking whether the company can convert 750 LOIs into repeat orders, prove reliability in the field, and scale production capacity without another six-week slip becoming a six-month one. The Q3 call answered the demand question again. The 2027 call has to answer the execution question.