Adicet Bio trades at $7.21 with $158.5M in cash. Market cap is $69M. Enterprise value is negative $89.5M. You are buying dollars for 44 cents and getting an allogeneic gamma delta CAR-T platform with three FDA Fast Track designations for free.

The market is pricing 81% failure. We think 35%. The entire thesis is one question: does gamma delta T-cell biology work in autoimmune disease? If yes, 7-14x upside. If no, -72%. The H1 2026 clinical update collapses the superposition within months.

The Setup

ACET is a pre-revenue clinical-stage biotech developing prula-cel, an off-the-shelf (allogeneic) gamma delta T-cell therapy for autoimmune diseases. The platform uses Vdelta1 gamma delta T cells -- a fundamentally different biology from every competitor in the space, all of whom use conventional alpha-beta CD19 CAR-T.

Seven patients with lupus nephritis and SLE showed positive preliminary results in an October 2025 data cut. No quantitative endpoints were disclosed in the 10-K. But the safety profile is the cleanest in the field: zero CRS, zero ICANS, across all clinical experience. The FDA aligned on outpatient dosing in November 2025 -- extraordinary for CAR-T, which normally requires hospitalization.

The financials are simple. Zero revenue. $95M/year burn. $614.7M accumulated deficit. $158.5M cash gives ≈20 months of runway. No debt. Deloitte has audited since 2020, no going concern. A 1-for-16 reverse stock split in December 2025 signals the prior year's distress. The company will need another $75-100M raise by 2027.

Why This Might Be Mispriced

Three observations that don't fit an 81% failure probability:

The cash floor is real and deep. $16.51 cash per share vs $7.21 stock price. Even in liquidation, the recovery is $14-16. The market is simultaneously pricing clinical failure AND value-destructive capital allocation -- management burning through cash rather than returning it. Both need to be true for the current price to be fair. Cash erodes at $8M/month, so this floor has a 12-month expiration, but right now it's 2.3x the stock price.

The sector is on fire. $6.35B in autoimmune CAR-T M&A in nine months: AbbVie acquired Capstan ($2.1B), BMS acquired Orbital ($1.5B), Lilly acquired Orna ($2.4B), Gilead acquired Interius ($350M). BMS CEO Christopher Boerner called autologous CAR-T data in lupus "spectacular" and Orbital's in vivo platform "transformative" (Q3 2025 earnings call). Kyverna is approaching the first-ever autoimmune CAR-T BLA filing. A $69M company with a differentiated platform in a space attracting multi-billion-dollar pharma checks is worth more than 81% failure implies.

Gamma delta biology is genuinely differentiated. Every competitor -- Kyverna, Cabaletta, Fate, Allogene, AstraZeneca/Gracell -- uses alpha-beta T cells targeting CD19 or BCMA. ACET uses Vdelta1 gamma delta T cells, which are MHC-independent (no GvHD risk), have natural tissue-homing properties (relevant for autoimmune disease in tissues), and have produced the cleanest safety profile in the field. The discontinued oncology program (ADI-270) showed 100% disease control rate and 20% ORR in renal cell carcinoma with zero CRS/ICANS in 5 patients -- killed for strategic reasons, not safety. The platform works. The question is whether it works better than CD19 in autoimmune.

Why the Market Might Be Right

Seven patients is a postage stamp. The 10-K reports "positive preliminary results" without quantitative endpoints. No SLEDAI scores, no BICLA response rates, no durability data. Compare to Fate Therapeutics' FT819: 13 patients, 70-78% SLEDAI reduction, 5/10 clinical remission, 2/2 complete renal response in lupus nephritis. Cabaletta has 76 patients enrolled. ACET's dataset is the thinnest in the field.

ACET is 4th-5th in the race. Fate FT819 leads the allogeneic category with an RMAT designation (stronger than ACET's Fast Track), more patients, quantitative efficacy data, and a registrational trial planned for 2026. Kyverna is approaching BLA filing for stiff person syndrome. Cabaletta has 10x the enrolled patients. ACET's competitive position is weak by development stage.

Big pharma is buying in vivo, not ex vivo. All $6.35B in autoimmune CAR-T M&A went to in vivo platforms -- RNA-delivered or LNP-delivered CAR-T that doesn't require manufacturing cells outside the body. Zero went to ex vivo allogeneic approaches like ACET's. BMS terminated its only ex vivo allo collaboration (Century Therapeutics) in March 2025. Pharma views ex vivo allo as transitional technology, not the endgame. This undermines the M&A optionality argument.

Dilution is inevitable and potentially destructive. The company has raised $111M (2024) and $75M (2025). Another $75-100M raise is coming. If clinical data is weak, that raise happens at $3-5/share, massively diluting existing holders. October 2025 offering investors bought at $16/share (adjusted) and are down 55%. The reverse split is the scar tissue of this pattern.

Factor Decomposition

Statistical regression (250 days): 83% idiosyncratic variance, 18% XBI (biotech sector), negative SPY beta. Passes the 75% idio threshold -- this is genuinely stock-specific.

But the economic decomposition reveals a one-factor bet:

Blended edge: ≈29%. Below our 45% threshold for sizing.

The biology, cash floor, and dilution factors are causally linked -- not independent. Positive biology holds the cash floor and enables a favorable raise. Negative biology erodes the floor and forces destructive dilution. Strip the dependencies and this is a single-variable bet: does gamma delta work in autoimmune?

The market can't price "gamma delta superiority" because there's no data to price. This is a latent factor -- it doesn't exist in any regression yet. If it materializes (H1 2026 data shows quantitative superiority over CD19 benchmarks), the factor appears for the first time and drives a massive re-rating. If it doesn't, ACET is just another 4th-place biotech with $95M/year burn.

Forward EV

Scenario analysis at 12 months:

EV-weighted target: $19.90 (176% upside).

Market-implied probabilities (backing out from $7.21 using our scenario targets):

The divergence is enormous. If we're right that the bear case is 35% not 81%, this is deeply mispriced. The question is whether we have any basis for that confidence. With 7 patients and no quantitative data, honestly -- we don't have much.

Edge-adjusted alpha (Paleologo): 156% idio return x 50% conviction x 29% edge = 22.6% annualized. Attractive on paper, fragile in practice.

Catalysts

Dense calendar creates multiple resolution points:

• H1 2026: Clinical update with more patients, longer follow-up in LN/SLE/SSc. THE catalyst.
• Q2 2026: FDA meeting on pivotal trial design. Single-arm vs randomized controlled trial.
• H2 2026: Pivotal study initiation.
• Q3 2026: ADI-212 (prostate cancer) IND filing.
• External: Kyverna BLA filing (H1 2026) will signal FDA's appetite for autoimmune CAR-T broadly.

Cash floor trajectory (shares held constant at 9.6M):

The floor disappears in 12 months. The catalysts arrive in 3-6 months. The race between data and cash depletion is the timing thesis.

Verdict

This is a doorway state. Two patterns fit the evidence. Neither can be eliminated until the H1 2026 data arrives.

The setup is compelling: negative EV, differentiated biology, near-term catalyst, sector tailwind. But we have low informational advantage on the single variable that determines the outcome. The market sees the same setup -- 7 analyst Buys proves it. Recognizing asymmetry isn't edge.

Full position: no. Edge below threshold, methodology is clear. A starter (0.5-1%) is defensible as an option-like payoff -- lose 0.72% of portfolio if wrong, gain 1-3% if right. Optimal play: wait for H1 2026 data, enter at $12-15 with dramatically higher edge (60%+) even though the cost basis is higher.

What creates edge from here: quantitative Phase 1 data (SLEDAI/BICLA scores to benchmark against Fate), the undisclosed CAR target identity (CD19 or novel?), and the Q2 FDA meeting outcome. The answer to "does gamma delta work?" isn't in the 10-K. It's in the data that hasn't been released yet.

Watch. Don't size. Let the catalyst come to you.

Evidence