Xtract One Doubled Revenue and Still Has a Bookings Problem
Xtract One's fiscal 2026 was, by the numbers management chose to lead with, a breakout: revenue of $28.8 million against $13.9 million a year earlier, a 108% increase, delivered while operating expenses rose only 13%. Adjusted EBITDA improved $5.1 million to negative $3.5 million. The operating-leverage argument the company has been making for four quarters finally has arithmetic behind it.
Revenue grew by $14.9 million this year while operating expenses only grew by $2.8 million driving a $5.1 million improvement in adjusted EBITDA. This means that about 1/3 of every incremental revenue dollar reached adjusted EBITDA.
Cash rose to $17.3 million from $8.2 million with no financing activity, and the fourth quarter generated roughly $1.9 million of operating cash flow. Strip working capital and the quarter was within $500 thousand of operating cash breakeven. For a $130 million market-cap hardware-software company, that is a real inflection.
The record is in revenue. The problem is in bookings.
Total contract value of new bookings was $32.2 million, down from $38.5 million — a 16% decline — and the fourth quarter contributed only $6 million against a year-ago fourth quarter of $16.1 million. Management's framing is lumpiness, and there is a legitimate base effect: fiscal 2025's fourth quarter carried the first wave of XtractOne Gateway orders plus several large education awards landing simultaneously. The average deal is also getting larger and multi-phased, which mechanically widens the variance in any single quarter.
But the more useful number is buried in the backlog disclosure. Contractual backlog ended the year at $25.6 million, up 66%, and $10.5 million of it is scheduled for recognition in the next twelve months. Set that against a $28.8 million revenue base and the arithmetic is uncomfortable: only about a third of fiscal 2027's revenue is already under contract before any new bookings. The revenue record was manufactured by converting the backlog — agreements pending installation fell from $34 million to $18.4 million as deployments accelerated, and for the first time the installed-and-contracted value exceeds what is waiting to be installed. That is a genuine milestone, and it is also a warning: the fuel that produced this year's growth has been substantially spent. Replacing it requires bookings to re-accelerate.
Here the bull case has one strong data point: a qualified opportunity pipeline Peter Evans put at roughly “US$122 million of qualified opportunities in our pipeline” — Peter Evans, CEO · 2026-10-22 — engaged buyers with intent, not leads, and increasingly outside North America.
A margin trough that keeps receding
Gross margin was 57% in the quarter versus 71% a year earlier; 58% for the full year versus 66%. The explanation is product mix — a mature SmartGateway against a first-year XtractOne Gateway still absorbing early production and installation cost.
The tension is that this was supposed to be resolved. On the March 2026 call, Karen Hersh described the two products as sitting in the “higher 60s this quarter, whereas the Xtract One Gateway are in the very low 50s” — Karen Hersh, CFO · 2026-03-05, and when Amr Ezzat of Ventum asked whether the reported quarter was the gross-margin trough, she answered "Absolutely" and Evans followed with "100%." Six months later the fourth-quarter margin is 57%, below the full-year average, with management offering that “it is really a question of mix” — Karen Hersh, CFO · 2026-10-22 and no systemic issue. That may well be true — but the company has now had several quarters in which the trough was declared and then declined to stay put.
The mechanism matters more than the excuse. Fourth-quarter installations were similar in count to the third quarter, yet “revenue for the quarter was lower than the third quarter primarily due to the installation contract mix” — Karen Hersh, CFO · 2026-10-22, with more subscription deals. Subscription contract value lands in backlog rather than revenue, so both the revenue line and the margin line are functions of the same contract mix variable. Every quarter the sales team tilts toward subscriptions, margins and reported revenue compress together; every quarter it tilts to upfront, they flatter. The product-margin gap is narrowing, but the mix problem is structural to the business model and will outlive it.
What management stopped saying
The most informative signal this quarter is what disappeared. Through fiscal 2025 the calls were dominated by capacity: doubling manufacturing lines, safety stock of long-lead components, “the demand is outstripping the manufacturing capacity today” — Peter Evans, Chief Executive Officer and Director · 2025-12-04. This quarter, capacity is barely mentioned and appears only as a completed fact — manufacturing improvements in the first half let the company push deployments through the second half. The bottleneck rhetoric of 2025 was real, and it has now been solved rather than talked about.
The vocabulary also rotated. The weapon detection framing that topped the keyword list in 2024 has been replaced by segment-specific product language and, notably, by procurement language: total cost of ownership, secondary searches, staffing. The "laptop problem" of prior calls has been rebranded the "backpack problem." The proof point this quarter was operational rather than technological — alert rates of 4% to mid-teens on fully loaded backpacks, against a displaced competitive install that required eight staff doing secondary bag searches.
The end-market mix is genuinely different too. Commercial and manufacturing were 37% of bookings, healthcare 20%, education 15%; sports and entertainment, which was 24% of bookings a year ago, no longer rates a mention in the top three. That is a real diversification away from the arena base — into schools, hospitals, convention centers, and corporate offices of insurers, healthcare systems and data-center owners. It is also the segment mix that skews to upfront purchase rather than subscription, which is the mix that helps margins.
The tailwind is legislative. The data is a phone call away.
The demand driver management is underwriting is regulatory: California's AB 75 screening requirement for healthcare arrives in March 2027 with, on Evans's estimate, “probably under 10%-15% penetrated or deployed across all the health care organizations” — Peter Evans, CEO · 2026-10-22. Texas has a roughly $430 million school safety funding program; Pennsylvania and Virginia are moving. In the UK, Martyn's Law is pulling the same thread, with the British Museum — nearly seven million annual visitors, tested every alternative, now an active reference — functioning as the company's most valuable sales asset.
What is missing is the retention evidence behind the land-and-expand claim that underwrites all of it. Asked directly how much of fiscal 2026 bookings came from existing customers, the company did not have the number. Evans volunteered that the industry churn standard is around 8% and Xtract One runs well below it, then added:
I am actually interested and we will get those numbers for you, Steve, around how many customers are repeat customers and expanding customers and what that looks like.
A company whose entire growth narrative rests on reference-selling not tracking its repeat-customer cohorts is a gap worth watching, even if the anecdotal evidence — a provincial hospital network leading to the next health authority, one Texas district advocating to another — is strong. Evans is candid about the limits of the technology too: “any organization that says the systems are self learning is lying” — Peter Evans, CEO · 2026-10-22 — data has to be curated and categorized by humans before it trains anything.
The contrast
Nothing about this story is a market wave. The editor-curated global keywords for the quarter are dominated by tariff refunds, Batch Zero, and commercial readiness; the tape's dominant theme over twelve months is AI data center, which has been both the best and the worst performing cluster depending on the window. Xtract One is company-unique — an idiosyncratic physical-security cycle with legislative clocks attached. Its only incidental contact with the mania is that enterprise data-center operators are now among its commercial customers.
So the question is narrow and answerable. Fiscal 2027 already has $10.5 million of contracted revenue against a $28.8 million base, a $122 million qualified pipeline, a near-breakeven cash model, and a product that just spent a year proving it converts. The bear case is equally simple: bookings fell, the margin trough was mis-called once already, and dependence on state legislatures and a March 2027 California deadline is not the same as demand in hand.