Robotics
PrivateA distressed sub-$50M-cap #3 in a radiotherapy duopoly — operating losses, a 15% RIF, withdrawn guidance, ~23M warrants of dilution overhang and a stock at $0.26 — where the entire equity is a levered option on a China-JV-and-EMEA recovery the company admits it cannot time; structurally cheap (0.4x EV/sales) but for cause, and not investable until book-to-bill and FCF inflect.
Research
The verdict
A distressed sub-$50M-cap #3 in a radiotherapy duopoly — operating losses, a 15% RIF, withdrawn guidance, ~23M warrants of dilution overhang and a stock at $0.26 — where the entire equity is a levered option on a China-JV-and-EMEA recovery the company admits it cannot time; structurally cheap (0.4x EV/sales) but for cause, and not investable until book-to-bill and FCF inflect.
Primary sources
Source documents — open to read in full
Accuray designs, manufactures, services and sells radiation-oncology systems — high-precision machines that deliver targeted radiation to destroy tumors while sparing healthy tissue. Two platform families, both robotic/image-guided:
The model is capital equipment + a long services annuity. FY2025 split: Products $237.6M (52%), Services $220.9M (48%). Services is the higher-quality line — it grows with the installed base and recurs (service contracts + spare parts), giving the business a floor even when system sales stall. Key payment characteristic: systems carry long order-to-revenue cycles (multi-quarter installation timelines), so revenue is a lagging, lumpy function of orders booked 4–8 quarters earlier — backlog conversion is the swing variable. Customers are hospitals, cancer centers and (heavily) international distributors — 81% of FY2025 gross orders came through distribution partners in international markets, up from 74%. The single most important customer relationship is the China JV (see Lens 2), which Accuray sells systems into — $101.6M of FY2025 product revenue (43% of product revenue) went to the JV.
Headquarters: Madison, Wisconsin (relocated from Sunnyvale, CA legacy). Nasdaq: ARAY. One reporting segment ("oncology systems group"). Smaller reporting company; accelerated filer. Auditor: Grant Thornton (since 2006), unqualified opinion on FY2025.
Upstream inputs → Accuray → end customer. Names where the filings give them; the rest flagged.
n/a — not disclosed in 10-K. The chokepoint is real and management-confirmed: "difficulties in obtaining a sufficient supply of component materials and increased component costs have adversely affected our gross margins" through at least FY2026. NEW in FY2026: tariffs are now a direct input-cost chokepoint — $2.6M of tariff cost hit product COGS in Q3 alone, ~530bps of product gross margin.Chokepoints: (1) single-/limited-source components → margin; (2) the China state-owned JV → 27% of revenue by geography and 43% of product revenue; (3) MENA distribution corridor → currently "delayed indefinitely." Three concentrated dependencies, all currently under stress.
Honest read: a narrow, eroding moat held by a distant #3. New-system sales are "primarily dominated by two companies" — Varian (now Siemens Healthineers) and Elekta — with Accuray, RefleXion, and ZAP Surgical as the chasing pack. Varian's oncology segment alone did €3.9B of revenue in 2024; Elekta ~$2.0B LTM. Accuray's total revenue is ~$0.45B — roughly one-tenth of the leader.
What moat exists:
Bargaining power: weak on both sides. Versus suppliers — limited-source components and now tariffs mean Accuray is a price-taker on inputs. Versus customers — international demand flows through distributors (81% of orders) who hold the end relationships, and the buyers are budget-constrained hospitals deferring capex. The 10-Q explicitly blames "reduced budgets and lower capital deployment priority for radiotherapy equipment" in the US since FY2024. Net: a real clinical niche and a sticky service annuity, wrapped around structurally subscale share and weak pricing power. A moat, but a shallow one — and the duopoly is widening it.
Accuray reports one operating/reporting segment (oncology systems group); the CODM does not assess product-line P&L, only revenue and long-lived assets by geography. So the meaningful cuts are by sales classification and by geography.
By classification (FY2025 vs FY2024):
| Line | FY2025 | FY2024 | Δ |
|---|---|---|---|
| Products | $237.6M | $234.2M | +1% |
| Services | $220.9M | $212.4M | +4% |
| Total | $458.5M | $446.6M | +3% |
| Of which sales to the JV: products $101.6M (FY25) vs $77.5M (FY24); services $18.5M vs $15.0M. Read: flat-to-up top line in FY25 was carried almost entirely by services (+$8.5M) and JV product shipments; ex-JV product revenue actually fell. |
By geography (FY2025 vs FY2024):
| Region | FY2025 | FY2024 | Δ |
|---|---|---|---|
| Americas | $88.8M | $90.2M | −2% |
| EIMEA | $144.3M | $168.6M | −14% |
| China | $124.5M | $103.4M | +20% |
| Japan | $53.6M | $55.7M | −4% |
| Asia Pacific | $47.4M | $28.7M | +65% |
| Read: the FY25 mix shift is the whole story — China (+20%) and APAC (+65%) masked a collapsing EIMEA (−14%) and a stagnant Americas. Accuray became, in effect, an Asia-centric company (China alone = 27% of revenue) — which is precisely why FY2026's China macro softness + MENA shipment freeze is so damaging. |
The Q3 FY2026 deterioration:
This is where the filings and the tape converge into one ugly number.
Unusual vs. own history: the FY2025 10-K showed a company near operating breakeven (+$7.8M op income, ~zero net). Two quarters later it is running a $32M nine-month operating loss on tariffs + restructuring + interest + a frozen MENA corridor. The speed of the reversal is the most important fact in this dossier.
Transcripts were absent from the research layer (transcripts/ empty); the read below is `` across the last 3–4 calls.
Peer set: the radiotherapy hardware field. Multiples are `` with source/date or n/a. Do not read across — these are at radically different points of distress.
| Company | Ticker | Mkt cap (USD) | EV (USD) | EV/Sales | EV/EBITDA | P/E | Div yld | 5-yr avg ROE |
|---|---|---|---|---|---|---|---|---|
| Accuray | ARAY | ~$30–41M | ~$176M | 0.41x | n/a — negative EBITDA | n/a — loss-making | 0% | negative (FY25 ROE ~−2%; LTM ROE ~−101% ) |
| Elekta | EKTA-B (STO) | ~$2.0B | ~$3.0B | 1.6x | 9.9x | 25.3x | yes (Elekta pays a dividend; rate n/a) | n/a (recent net margin ~1%) |
| Siemens Healthineers (Varian) | SHL (ETR) | n/a at group level for this run | n/a | n/a (group blends imaging+dx+Varian) | n/a | n/a | yes | n/a |
| RefleXion | private | n/a — private | n/a — private | n/a — private | — | — | — | |
| ZAP Surgical | private | n/a — private | n/a — private | — | — | — | — |
Read: ARAY at 0.41x EV/sales vs Elekta's 1.6x looks ~75% cheaper on revenue — but that gap is earned: Elekta makes 19% EBITDA margins and a profit; Accuray makes operating losses and an LTM ROE near −100%. The cheapness is a distress discount, not a mispricing, until the P&L inflects. The EV ($176M) being >4x the equity (~$30–41M) is the tell — this is a debt-dominated capital structure where the equity is the thin residual tranche.
Pattern over the last ~2 years, `` throughout:
insider-transactions.csv absent from the research layer; insider ownership % n/a this run. Worth noting the lenders hold ~23.4M warrants ($1.68 and $0.01 strikes) — the most economically significant non-management equity claim.Acting as a forensic analyst. Every figure labeled.
Regulatory findings (required sub-section).
quote n/a — no material specific proceeding surfaced in the reviewed text.Built bottom-up from FY2025 actuals + the Q3 FY26 nine-month run-rate. Every line labeled; outputs ``. Note: this is a loss-making, sub-$0.30 stock — EPS precision is meaningless at this share price; the real question is survival/inflection, so treat these as scenario sketches, not point forecasts. Per --watchlist rules, no forecast.ts log created.
Anchors: FY25 revenue $458.5M, ~zero net. 9-mo FY26 revenue $301.0M (−9%) → implied FY26 revenue ~$400–405M. Share count ~119M and rising toward ~135–142M as the ~23.4M warrants exercise.
| Scenario | FY2026E | FY2027E | FY2028E | Logic |
|---|---|---|---|---|
| Bear | rev ~$395M; EPS ~$(0.45) | rev ~$385M; EPS ~$(0.30) | rev ~$390M; EPS ~$(0.20) | MENA stays frozen, China soft, tariffs persist; cost-out only partly offsets; persistent op losses + interest. Aligns with sell-side's ~$17M 2028 loss. |
| Base | rev ~$402M; EPS ~$(0.40) | rev ~$415M; EPS ~$(0.12) | rev ~$435M; EPS ~$(0.02) → ~breakeven | $25M cost program lands; MENA partially reopens FY27; China stabilizes; services annuity grows; interest still ~$30M/yr caps net income. |
| Bull | rev ~$408M; EPS ~$(0.35) | rev ~$435M; EPS ~$(0.02) | rev ~$470M; EPS ~$+0.08 | MENA backlog ships, China JV equity income re-accelerates to >$5M, book-to-bill back >1.2, full $25M cost-out + operating leverage → return to net profit by FY28. |
Base call (qualitative, the only one that matters): ARAY does not reach sustained GAAP net profit before FY2028 at the earliest, and only if the MENA corridor reopens AND the $25M cost-out fully sticks AND interest doesn't rise. The binary that actually scores the equity is not an EPS line — it is "does FCF turn positive and book-to-bill hold ≥1.1 before liquidity forces another dilutive raise?".
Bull case. A sub-$50M-cap company with $458M of revenue, $356–427M of backlog, a 48%-of-revenue recurring service annuity, and a 49% stake in a growing China JV is priced for liquidation, not for the cash-generative niche medtech it structurally is. Book-to-bill is still 1.2 trailing — orders exceed product revenue, so the backlog is growing even now. New management is executing a credible $25M cost-out ahead of plan, the worst margin pressures (tariffs, MENA freeze) are exogenous and mean-reverting, and the maturity wall is pushed to 2030. If MENA reopens and the cost-out lands, FY28 returns to profit and the equity — as the thin tranche on a ~$176M EV — re-rates violently off a $0.26 base. The contrarian view: the market has confused a cyclical/geopolitical air-pocket with a structural failure, and the service annuity + backlog put a floor under the enterprise that the equity price ignores.
Bear case. Three things can permanently impair this equity: (1) Liquidity — $38M cash, ~$19M nine-month burn, inventory-bloated working capital, $18M converts due June 2026, term-loan covenants, Z-score <3. One more frozen-corridor quarter forces a dilutive raise into a $0.26 stock, crushing the residual equity. (2) The China JV is a single point of failure that is also a geopolitical hostage — 43% of product revenue and the only real growth, running through a Chinese state nuclear counterparty amid escalating US-China tech/tariff conflict; a tariff/export-control or JV-governance shock is existential, not marginal. (3) Structural subscale — at ~1/10th Elekta's revenue and a distant #3, Accuray can't out-R&D the duopoly (R&D ~$48M vs Varian/Elekta's multiples of that), so share erosion to RefleXion/ZAP at the low end and Varian/Elekta at the high end is the base rate. Pre-mortem (18 months out, thesis broke): MENA never reopened, a 2026 cash crunch forced an equity raise at $0.15 that doubled the share count, China JV income kept fading on tariffs, and the $25M cost-out was eaten by revenue decline — the stock is a sub-$0.10 perpetual restructuring. Multiples too high? On EV/sales, no — 0.41x is cheap; but on equity value the right question is solvency, and there the margin of safety is thin. Most likely truth: the enterprise is worth more than $176M to a strategic acquirer; the public equity may not survive to capture it without dilution.
Dismantling the bull. (1) Revenue concentration that bulls hand-wave: "diversified geography" is a fiction — China + APAC are 37% of revenue and China alone 27%, routed through a 49% JV with a state nuclear entity you cannot control, audit independently, or repatriate freely; if Beijing or Washington moves, 27% of revenue and the entire growth narrative reprice overnight. (2) The moat is the service base, and the service base only exists because old systems were sold — if product orders keep falling (products −13% last quarter), the annuity stops growing and eventually shrinks with the aging fleet. (3) The most dangerous competitor bulls underestimate is not Varian — it's RefleXion and ZAP at the low end, who can take the budget-constrained international hospital that is exactly Accuray's distributor-driven customer. (4) Worst capital allocation: the 2024 refi traded a cheap balance sheet for a $150M term loan that now eats $24M/yr of interest and handed lenders 23.4M warrants — management mortgaged the equity to survive, and the interest line is the net loss. (5) What must hold for $0.26: that the company refinances/repays the June-2026 converts without a death-spiral raise, that MENA reopens on an unknowable timeline, and that the $25M cost-out outruns revenue decline — three things, at least two outside management's control. (6) Growth disappoints 20–30%: at $400M→$300M revenue the operating loss widens, covenants tighten, and a dilutive recap becomes near-certain — equity → near-zero. (7) Single scenario that permanently impairs: a China tariff/export-control escalation or JV unwind, which is plausible and rising, not tail. Short thesis: the equity is an out-of-the-money option on a geopolitically-hostage turnaround, priced like one for good reason.
A near-breakeven Chinese smart-EV OEM whose margin (GM 18.9% FY25, ~20% Q1'26) and a high-margin VW software-licensing annuity are real — but FY26 volume has rolled over (-22.6% YTD), and the IRON/eVTOL/robotaxi "embodied-AI" optionality the bulls pay for is unproven cash-burn; long the software+margin inflection at a 52-week-low multiple, but only if the GX/new-model cycle re-accelerates deliveries by 2H26.
A 30-year never-profitable robotic-navigation company that finally owns its full stack (GenesisX robot + MAGiC catheter) and is one H2 manufacturing ramp away from the razor-and-blade flywheel turning — a binary execution call where the Street's $4+ targets price the ramp as a near-certainty the $6.3M Q1 print does not yet support.
A $5M-revenue, revenue-SHRINKING robot company trading at ~90x sales is really a $270M cash-box wrapped in a serial-dilution, AI-partnership-hype machine — now under a 10b-5 class action for a Microsoft "partnership" Microsoft denied; the only real value is the cash, and management is burning it while printing stock.