Phase Space AI

Trade Construction

FuelCell Energy [FCEL]

FuelCell Energy [FCEL] — Trade Construction

Spot $18.065 (2026-07-29) · realised 252-day vol 127.6% · Jan-2027 ATM implied vol ~140% 12-1 momentum +510.9% · −49.8% from the 52-week high of $36.01, +362.0% above the 52-week low of $3.91

The memo issues no position verdict.


1. Liquidity verification — done before any structure was proposed

Chain pulled live 2026-07-29 via Alpaca options snapshots, expiry 2027-01-15, 110 contracts.

Contract OI Bid Ask Bid size Ask size IV Delta
Call 20.0 5,002 5.80 6.36 175 512 1.413 +0.649
Call 25.0 2,597 4.48 5.56 224 862 1.440 +0.566
Call 15.0 1,918 7.33 8.05 117 702 1.410 +0.752
Call 23.0 1,099 4.79 6.53 1,509 1,066 1.485 +0.608
Put 21.0 333 7.33 8.67 1,494 83 1.287 −0.392
Put 23.0 285 9.37 10.09 242 10 1.344 −0.420
Put 40.0 146 23.40 24.78 1,152 144

Maximum open interest across the expiry: 5,002. Total open interest: 47,821. Two-sided quoted size runs from tens to over a thousand contracts depending on strike.

Liquidity Criteria: PASS — this is not the HCA failure (18 contracts of maximum OI across a whole expiry). But note the asymmetry: open interest is concentrated in calls, and the put strikes are thin (OI in the tens to low hundreds, ask sizes as low as 10 contracts). The chain is fillable for a long-call expression and awkward for a put expression, which is exactly what one expects on a name up 511% in twelve months.

Cash equity. IEX-only 63-day ADV $8.29m on ~394k shares/day; scaled to consolidated on the FSLR-calibrated 21.7x IEX share, approximately $180m/day. The IEX figure and the scaling basis are both stated because the multiplier is an estimate.


2. Why no position — on either side

Long: Quality Criteria (BINDING) fails on the archetype's core test. criteria.md on INFLECTION: gross margin as a LEVEL "proves the unit economics work." FCEL's is −16.7% (FY2025) and −36.3% (Q2 FY2026), and every segment except externally funded R&D is negative. Valuation Criteria (BINDING) fails: at the demonstrated revenue CAGR the price requires a 29.2% terminal operating margin, above the best gross margin any segment has ever recorded. Cheap cannot rescue a Quality failure, and this is not cheap — 6.10x EV/Sales at the 83rd percentile of its own usable history.

Short: the analysis is negative and the vehicle is the problem, not the view.

  1. Implied vol is ~140%. Buying a put costs 140 vol against 127.6 realised — paying up for realised volatility that is already extreme. The Jan-2027 21-strike put costs $8.67 on an $18.07 stock: 48% of spot for a 5½-month option. Breakeven $12.33, −31.7%. The bear case reaches it, but the premium consumes most of the move.
  2. The put strikes are thin. OI 333 on the 21, 285 on the 23, ask size 83 and 10 respectively. Size is limited to a few tens of contracts at the touch.
  3. 12-1 momentum is +510.9%. Momentum Criteria is MEASURED and governs timing, never whether — but the timing signal here is unambiguously hostile to a short. A company that raised $213.4m of equity at $21.00 three weeks ago, signed a 380 MW headline agreement five weeks ago and received US government financing four weeks ago is not a name to be short into the next press release.
  4. The catalyst set is asymmetric. Fit's phase-1 election is a binary the company does not control and the market cannot date. A single 100 MW election would be worth several years of revenue at the headline.
  5. Cash short is worse. Borrow and days-to-cover were not pulled (they are not part of Liquidity Criteria on this book), but 127.6% realised vol on a name +362% off its low, with a nearly-exhausted ATM and a fresh automatic shelf, is a poor cash-short profile.

Correct action on a long-only absolute-return book: no position. Correct action on the RV fork: none either — the vehicle economics do not support the view. This is a name to watch, not to trade, and the watch item is named in §4.


3. The structure that would express the view, if the vehicle improved

Recorded for the ledger, not proposed.

Jan-2027 21/13 put spread. Long 21 put (OI 333, ask 8.67, IV 1.287, delta −0.392); short 13 put (OI 12, ask 10.94/bid 9.02 — too thin to trade). The short leg does not exist in size, so the spread cannot be constructed. That is a real finding and it is the reason no structure is proposed: the defined-risk expression is unavailable at any size on this chain.

Outright Jan-2027 21 put, $8.67 debit, breakeven $12.33 (−31.7%), maximum loss 100% of premium at $18.07 or above. At a 0.5%-of-book maximum loss the position is 5.8 contracts per $1m of book — below any sensible minimum clip. Not investable.


4. What to watch, and what would change the analysis

Trigger Level / event Meaning
The single hard invalidation Company-level gross margin turns positive for two consecutive quarters The INFLECTION gross-margin level test flips. Everything in this memo is downstream of a −36.3% gross margin
Fit phase-1 election Fit Energy USA LP elects phase 1 (100 MW) with the non-refundable deposit paid 30 MW contracted becomes 130 MW. Warrant tranche 1 (4.0m shares at $26.44) begins vesting. Re-underwrite immediately
Product backlog Reverses the −63.2% year-on-year decline ($36.1m at 2026-04-30 vs $98.2m) The near-term manufacturing order book is the only backlog line that measures demand
Advanced Technologies backlog Reverses the −48.0% decline ($15.4m vs $29.6m) The only positive-margin segment's forward book
Generation segment gross margin Improves from −154.1% (Q2 FY2026) / −33.2% (FY2025) $928.5m of contracted backlog converts through this line. At a negative margin it is a liability
Torrington funding A committed, sized financing for the 500 MW expansion that is not an equity raise The expansion is explicitly unfunded and the stated route is government assistance programmes
Equity issuance Any raise below ~$12 Confirms the going-concern formulation: the equity is a call option on continued access at elevated prices
Price, up $36.01 (52-week high) The market has priced something the analysis does not contain
Impairments A third consecutive impairment on project assets $65.8m (FY2025) + $42.6m (Q2 FY2026) already taken against a fleet being simultaneously built

5. Sizing

Not applicable — no position. For the record, inverse-volatility sizing at a 20% reference volatility gives a scalar of 0.157x a unit position on 127.6% realised vol. FCEL would be the smallest position in this cluster on that rule, and the second-smallest thing about it would be the conviction.


6. Recommendation-ledger entry

No trade recommendation is issued, so no line is appended to trade_recommendations.jsonl. The Downside Criteria scenario set in FCEL_Valuation.md §5.2 (bull 20% / base 35% / bear 45%, with a ~15% going-concern subset) is logged for ledger_scorer.py to Brier-score if the book later takes a position.

A specific record for the calibration file: the framework's defined-risk default — a vertical spread — was unconstructable here because the short leg had 12 contracts of open interest. That is the HCA lesson repeating on a different name, and it is only visible because the chain was pulled before the structure was proposed. The rule works.