FuelCell Energy [FCEL]
Spot $18.065 (2026-07-29) · EV $1,024.9m (pro-forma) · TTM revenue $167.879m · EV/Sales 6.10x
| Parameter | Value | Source |
|---|---|---|
| Spot | $18.065 | Alpaca 2026-07-29 |
| Shares outstanding | 78,322,459 | post the 2026-07-09 offering (10,714,286 shares at $21.00); the 1,607,143-share greenshoe is not assumed exercised |
| Market capitalisation | $1,414.9m | |
| Unrestricted cash 2026-04-30 | $373.167m | balance sheet |
| Restricted cash | $67.685m | excluded — company states $67.7m is pledged as performance security and LC collateral through October 2029 |
| Total debt and finance obligations | $134.180m | debt note, twelve instruments |
| Redeemable Series B preferred | $59.857m | liquidation preference $64.020m, $3.2m annual dividends, senior to common |
| Net cash, last filed | $179.130m | |
| Post-quarter adjustments | +$213.400m equity (2026-07-09), +$22.0m EXIM tranche 1 cash (2026-06-30), −$24.5m new EXIM debt | disclosed, dated |
| Pro-forma net cash | $390.030m | excludes ~3 months of burn (~$40m); realistic ~$350m |
| Enterprise value | $1,024.9m | |
| Base revenue | $167.879m TTM to 2026-04-30 | four filed quarters |
| WACC | 10.0% | house standard, held constant across the cluster |
| Horizon | 5 years | |
| Terminal value as % of EV | 100% | reverse DCF is mandatory as the primary output |
valuation.md requires the exit multiple to come from a comparator set whose growth brackets the subject's
growth at the exit year, or to be declared UNIDENTIFIED. It requires the terminal margin to be judgeable
against what the business has done. Neither input exists for FuelCell Energy.
A terminal operating margin is therefore not derivable from any observed period. The reverse DCF is run anyway, because the sensitivity is the finding.
assets/reverse_dcf.py, --solve cagr, all §1 parameters fixed. Demonstrated revenue CAGR is +13.2%
(FY2023 $123.394m → FY2025 $158.162m), and that itself overstates the trend because FY2024 revenue of
$112.132m was below FY2023 — the series fell then rose, so a two-point CAGR is not a growth rate. The
latest quarter is −4.9% year on year.
| Terminal operating margin | Basis | Exit EV/EBIT | Required revenue CAGR | Margin (13.2% − req) |
|---|---|---|---|---|
| 9.1% | the screen's own: "industry median of mature profitable peers (pre-profit subject)" | 15.0x | +48.4% | −35.2pp |
| 9.1% | " | 18.1x (the screen's own) | +42.9% | −29.7pp |
| 9.1% | " | 25.0x | +34.0% | −20.8pp |
| 26.7% | best-ever gross margin of the best segment, used as an absurd upper bound for an operating margin | 15.0x | +19.7% | −6.5pp |
| 26.7% | " | 18.1x | +15.3% | −2.1pp |
| 26.7% | " | 25.0x | +8.1% | +5.1pp |
Reference result: −29.7pp at the screen's own inputs. FAIL.
Note what the last two rows do. To get the margin near zero you must assume that the whole company, in perpetuity, earns an operating margin equal to the best gross margin ever achieved by a cost-plus-contract segment that is 13% of revenue and whose backlog just fell 48%. That is not a bull case; it is a category error, shown here only to establish that even the most abusive assumption available does not produce a pass with room to spare.
--solve terminal_margin, exit 18.1x (the screen's own multiple), all else fixed:
| Assumed revenue CAGR | Required terminal operating margin |
|---|---|
| +13.2% (the demonstrated two-point CAGR) | 29.2% |
| +30% | 14.6% |
| +50% | 7.2% |
At the company's own demonstrated revenue growth, today's price requires a 29.2% terminal operating margin. That is:
Following the NuScale precedent — where the flip point sat at 631x EV/EBIT, outside the solver's bounds, and "the price cannot be justified anywhere in the plausible range" was recorded as the correct output — the finding here is stated the same way:
There is no combination of revenue growth and terminal margin within FuelCell Energy's demonstrated range that justifies $18.065. The solver returns numbers, but every solution requires a margin the company has never approached in any segment in any period, or a growth rate (+43% to +48% compounded for five years) against a two-point demonstrated CAGR of +13.2% built on a revenue series that fell before it rose, and a latest quarter of −4.9%.
Implied compression, stated as a number: today's EV/Sales is 6.10x. The company's own post-reverse-split median EV/Sales is 1.97x (see §3). Holding revenue flat, mean reversion to that median alone is 4.13 turns of compression, −67.7%.
valuation.md: "If the history is too short or spans a regime change, declare it UNIDENTIFIED rather than
substituting a peer median." This is that case, and the reason is arithmetic rather than editorial.
FCEL effected a 1-for-30 reverse stock split on 2024-11-08 (611,278,662 shares → ~20,375,932). Any EV/Sales series built from split-adjusted prices and as-reported share counts is corrupted before that date — the pre-split portion of the full series returns values up to 4,511x, which is a data artefact, not a multiple. The usable window is 2024-11-08 → 2026-07-29, n = 429 trading days.
Within that window, point-in-time EV/Sales (TTM revenue as known at each date, share count and net cash from the most recently filed balance sheet at each date):
| Percentile | EV/Sales | Implied price¹ |
|---|---|---|
| p10 | −0.25x | negative EV — the stock traded below net cash |
| p25 | 0.77x | $6.63 |
| p50 | 1.97x | $9.20 |
| p75 | 3.80x | $13.12 |
| p90 | 9.58x | $25.51 |
| min / max | −0.47x / 47.55x | — / $106.79 |
| today | 6.73x | — the 83rd percentile of its own usable history |
¹ (TTM revenue $167.879m × multiple + pro-forma net cash $390.030m) ÷ 78.322m shares.
The interquartile range alone implies a twelve-month price of $6.63 to $13.12; the p10-to-p90 range implies a negative enterprise value to $25.51; the full range implies $0 to $106.79. A distribution with a 16-fold interdecile spread over 21 months carries no information about where the stock trades in twelve months. And the p10 being negative is not a computational error — it records that in early 2025 the market valued the operating business at less than nothing, which is the same market that now values it at 6.7x sales.
The twelve-month target is UNIDENTIFIED. No number is substituted, and no peer median is imported. The scenario set in §5 is what replaces it.
The obvious comparator is Bloom Energy [BE], the same end-market with a different platform (solid oxide rather than molten carbonate) and — decisively — a positive gross margin. A multiple derived from a company whose gross margin has the opposite sign cannot be applied to this one. That is the growth-matched-anchoring rule applied to margin sign rather than growth rate, and it is the same defect: extrapolating a relationship beyond the support of the sample.
Two full years of post-split trading history without a >300% twelve-month move, or a positive company-level gross margin that gives EV/EBIT a defined sign. Neither is available in the next twelve months. This name cannot be given a twelve-month target by this framework until one of them exists — and that is a statement about the name, not a gap in the process.
Named peer in the same end-market: Bloom Energy [BE]. Spread versus its own history is not computable against a comparator with a structurally different margin sign, and no percentile is asserted. Declared UNIDENTIFIED rather than defaulted to a sector median.
Type: MEASURED. Logged; blocks nothing.
Named cause: the Fit Energy option is never exercised, product and Advanced Technologies backlog continue to halve, and equity-market access reprices.
criteria.md: "A going-concern bear case must be argued explicitly and flagged." Flagged.
Arguments that there is no near-term going-concern event:
Arguments that the going-concern risk is nonetheless material on a three-to-five-year view:
Conclusion: FCEL equity is, structurally, a call option on continued equity-market access at elevated prices. That is the going-concern formulation and it is the sentence that should carry into any position decision.
| Case | Prob. | Path | Implied price | vs spot |
|---|---|---|---|---|
| Bull | 20% | Fit exercises phase 1 (100 MW) with non-refundable deposits, vesting a warrant tranche; product backlog rebuilds; revenue to ~$250m. Multiple to own p90 9.58x; net cash ~$300m; ~90m shares as warrants vest | $29.94 | +65.8% |
| Base | 35% | Backlog converts at $170–200m of revenue; gross margin stays negative; one more equity raise. Own p50 1.97x on $185m; net cash ~$300m; ~85m shares | $7.81 | −56.8% |
| Bear | 45% | Fit option lapses; product backlog (−63% y/y) and Advanced Technologies backlog (−48% y/y) keep falling; Torrington capex commits into a −20% to −36% gross margin; further dilution. Own p25 0.77x on $160m; net cash ~$250m; ~95m shares | $3.93 | −78.2% |
| Going concern (subset of Bear) | ~15% | The above plus equity-market access closing at a materially lower price, forcing a restructuring in which the $64.0m preferred and $134m+ of debt rank ahead of common | near zero | −100% |
Probability-weighted, the scenario set is deeply negative, which is the same conclusion the implied-path test reaches by a different route. Neither output is a twelve-month target and neither is presented as one.
Revenue fell 4.9% year on year in the latest quarter. Revenue per weighted share fell 61.9% ($1.721 → $0.656). Weighted shares went +149.4%. In this name, dilution is not a modelling adjustment — it is the dominant driver of per-share outcomes, and it points the wrong way in every scenario except the bull case, where warrant vesting adds 12.0m shares precisely when the thesis works.
net loss attributable to common ÷ weighted shares ≈ filed EPS: Q2 FY2026 $(78,707)k ÷ 54,224,428 =
$(1.4515) vs filed $(1.45). ✓gross_margin_pct: -35.0 at margin_period: 2025-10-31 could not be reconciled to any filed
annual or quarterly figure and is reported as unreconciled rather than explained away.