Thesis

COYA Therapeutics ($4.49, $108M mkt cap) is a single-asset clinical-stage biotech developing COYA 302 — a low-dose IL-2 + abatacept combination targeting regulatory T-cell (Treg) enhancement in ALS. The stock trades at base-rate Phase 2 failure odds (≈18% implied P(success)) while independent evidence has materially shifted the probability upward.

The MIROCALS trial, published in The Lancet in May 2025 (n=220, Phase 2b), showed that low-dose IL-2 monotherapy reduced death risk by over 40% in a pre-specified biomarker-stratified ALS subgroup (≈80% of patients). COYA 302 adds abatacept on top of IL-2 — a dual mechanism that could improve on monotherapy results.

The market hasn't connected MIROCALS to COYA's price. At $4.49, you're buying a program with independent mechanism validation, a $12B pharma partner (Dr. Reddy's) paying milestones, no competing approved ALS therapy on market, and cash through the Phase 2 readout — all at a price that implies nothing beyond base-rate odds.

The Setup

COYA's 10-K (FY2025, filed March 16, 2026) confirms clean execution:

• ALSTARS Phase 2 commenced dosing December 2025. Enrollment completion targeted H2 2026 — first time this date has appeared in a filing. Topline before H2 2027.
• $57.9M effective cash ($46.8M year-end + $11.1M January PIPE). No going concern. Runway stated to H2 2027, past the readout.
• Dr. Reddy's has paid $15.9M in milestones ($7.5M upfront + $4.2M IND + $4.2M first dosing). Up to $40M in development milestones and $677M in sales milestones still pending.
• FTD IND accepted January 2026 — second indication in 4 months. COYA retains all FTD commercial rights (outside the DRL license).
• Shimon Sakaguchi, Treg discoverer and COYA SAB member, won the 2025 Nobel Prize in Physiology or Medicine.

No going concern. No ATM facility. No trial delays. Burn is accelerating (+42% YoY to $21.2M net loss) as ALSTARS scales, but the cash math works through the answer.

Where the Edge Is

This is a single-factor binary bet. ALSTARS Phase 2 outcome (F1) drives ≈70% of thesis value. Everything else — DRL partnership, competitive landscape, pipeline optionality — is derivative. If ALSTARS hits, DRL milestones trigger and the platform is validated. If it misses, the stock goes to cash value.

The edge is in F1, and it comes from MIROCALS.

Base rate: Phase 2 neurology success is ≈15-20%. The market prices COYA at ≈18% implied probability (backing out from $4.49 vs $15 success / $2.25 failure targets).

Our estimate: 25-35%, central 30%.

The adjustment:

Net: +8-10% above base rate. Edge = ≈12 percentage points.

Moderate, not extreme. But 12 points of edge on a 3.3:1 asymmetry pays.

The Bear Case (Taken Seriously)

MIROCALS primary endpoint failed. The full ITT population showed no survival benefit. Only the pre-specified biomarker subgroup (lower CSF-pNFH, ≈80% of patients) showed the >40% reduction in death. Pre-specified subgroups are better than post-hoc, but they're not primary endpoints. The biotech graveyard is full of "it worked in the subgroup" stories that didn't replicate.

Critical open question: Does ALSTARS use CSF-pNFH biomarker stratification? The 10-K doesn't disclose this. If ALSTARS doesn't stratify, it may repeat MIROCALS' full-population miss. If it does stratify, the commercial population narrows but P(success) increases materially. This is thesis-critical and currently unanswered.

DRL is not excited. They mentioned COYA exactly once across six quarters of earnings calls — a single line in prepared remarks, no analyst follow-up. Not discussed in their 20-F annual filing. Total COYA spend ($25.9M) is less than 1% of DRL's annual revenue. The Immutep deal ($20M upfront + $350M milestones) received more prominence in one quarter than COYA has across the entire relationship. DRL is maintaining a cheap option, not declaring a strategic priority.

Will need to raise again. Net loss jumped 42% YoY. As ALSTARS enrollment scales, 2026 clinical expenses will grow from $4.9M toward $8-12M. Estimated 75% probability of additional raise before topline. If enrollment goes well, they raise at $6-7 on reasonable terms. If it stalls, they raise at $3-4 with warrants. No ATM means each raise is a discrete event — visible but dilutive.

Binary concentration risk. Eight employees. One clinical program that matters. If ALSTARS fails, the stock goes to cash value or below. FTD provides minimal floor ($0.25-0.50/share optionality at best). This is not a diversified thesis with multiple shots on goal.

Options market split personality. Near-term puts are heavy (P/C 35:1 on March 20 expiry), but these are expiring ITM contracts — legacy positioning, not fresh signal. The real story is May and August: 24:1 calls-to-puts in May with 218 OI at the $8 strike. Someone is positioned for the enrollment window. This cuts both ways — smart money sees the same catalyst, but the near-term bearish overhang is real.

Competitive Landscape

The ALS market is structurally favorable for COYA right now.

Amylyx's Relyvrio — the only ALS-specific therapy approved since riluzole and Radicava — was formally withdrawn by the FDA in August 2025 after the Phase 3 PHOENIX trial failed. There is a commercial void and regulatory pressure to approve something.

The active Phase 2-3 pipeline targets different mechanisms: Ionis/Biogen (ASO for FUS-ALS), Prilenia (sigma-1 receptor), BrainStorm Cell (stem cells), Zydus (NLRP3 inflammasome). None compete directly on Treg enhancement. The closest Treg competitor is Cellenkos (private, Phase 1, allogeneic cord blood Tregs) — completing 2026 but well behind COYA.

The broader Treg field is gaining validation: Nektar (NKTR) is Phase 3 ready with a Treg-selective IL-2 agonist in atopic dermatitis. Sonoma Biotherapeutics has positive Phase 1 CAR-Treg data in rheumatoid arthritis. But COYA's positioning in neurodegeneration is unique among public companies.

Factor Decomposition

Beta 0.23, idio variance ≈80-85%. This stock moves on its own news — market and sector exposure is negligible. COYA was down 34.5% over the past year while XBI was up 41.1%. Completely decoupled.

Edge is concentrated in one factor. This is a single-catalyst bet — size accordingly.

Forward EV

Market-implied P(success): ≈18%. Backed out from $4.49 current, $15 success target, $2.25 failure floor.

Our P(success): 30%. Based on MIROCALS mechanism validation, POC data, sound biology, net of subgroup and sample size caveats.

Three scenarios, probability-weighted:

Edge-weighted alpha (adjusting for edge quality by factor): ≈32% annualized.

Kelly optimal is ≈15%, half-Kelly ≈7-8%, but single-factor binary risk warrants conservative sizing: 1-2% of portfolio.

Timing

The catalyst map has a clear structure:

Q2 2026 (Apr-Jun): Biomarker publication targeted H1 2026 — if it shows Treg enhancement correlating with functional improvement in COYA 302 specifically, that's mechanism confirmation worth a re-rate. Also the window for enrollment pace updates and conference presentations (AAN, ENCALS).

H2 2026 (Jul-Dec): Enrollment completion — management committed this date in the 10-K. If enrollment completes on schedule, it materially de-risks the topline timeline and is the key add trigger. If it slips, trim.

H1-H2 2027: ALSTARS topline readout. The binary event. Everything before this is positioning; everything after is reaction.

Entry: Current range ($4.00-4.50) is the level. The stock is at base-rate pricing. Tranche in: 0.5% now, add 0.5-1.0% on enrollment pace confirmation (Q2-Q3 2026), max 2%.

Thesis kill: ALSTARS enrollment suspended, DRL terminates, going concern in next filing, FDA clinical hold, or CEO departure without replacement.

What I Don't Know

Two open questions that could materially shift the thesis:

1. Does ALSTARS use biomarker stratification? MIROCALS worked in the CSF-pNFH-low subgroup and failed the full population. If ALSTARS doesn't stratify, our 30% estimate may be too high. If they do, it may be too low. The 10-K doesn't disclose trial design at this level of detail.

2. Enrollment pace. H2 2026 completion is management guidance, not a regulatory commitment. ALS patients are notoriously difficult to enroll (rapid disease progression, narrow treatment windows). The first concrete enrollment update — likely Q1 2026 earnings call or a conference in May-June — will tell us whether the timeline is real.

Both are resolvable. Neither is resolvable today. Size for not knowing.

Evidence