Phase Space AI

03 Trade Construction

BillionToOne [BLLN]

BillionToOne, Inc. [BLLN] — Trade Construction & Risk Management (Task 5)

⚠️ SUPERSEDED IN PART — 2026-07-29

The position verdict in this document is retired. Under the current framework (references/criteria.md, 2026-07-29) the memo outputs an analysis, not a position. Whether an analysis justifies a position is a question about a particular book, and two books answer it differently.

The Gate block and the Gate 4 expected-return arithmetic below are also superseded, by the named Criteria (each with a type: BINDING or MEASURED, returning PASS / FAIL / INDETERMINATE), the reverse-DCF implied-path test, and a separate 12-month target.

→ Current analysis: BLLN_Criteria_and_Valuation_2026-07-29.md

Everything else here — the research, the evidence, the mechanism work — stands. Residual references to "Watchlist" in the prose below are the historical record of the 2026-07-28 assessment and are left intact deliberately.

The four conclusions, kept separate

1. Fundamental conclusion — POSITIVE and genuinely so. Revenue tripled in two years; gross margin went 24% → 73%; the company crossed into GAAP profitability and positive operating cash flow in FY2025 and accelerated in Q1-2026. Net cash $447m. Payor coverage reaches ~90% of the US population. This is a high-quality business and the memo says so without hedging.
2. Expectations conclusion — the market already knows. Consensus FY2026 revenue $454m against company guidance of $450–465m; the house Base case is $462m, +1.8%. On FY2027 the house is +4.2%. The stock trades 8.6% above the average Street target of $122.14. There is no material near-term expectations gap.
3. Valuation conclusion — stretched on cash flow, in line on comps. DCF base $53.35 (Gordon) to $79.90 (4.5x exit) versus spot $133.64. Yet EV/FY2027E sales of 10.8x is below NTRA (~11.9x) and GH (~14.4x) at roughly triple their growth. These two facts genuinely conflict, and the conflict is the honest centre of this memo rather than something to resolve by picking the convenient one.
4. Portfolio conclusion — no position. Set out below.

Hard gates

Gate 1 — Causal mechanism: PASS (long-side)

A specific, evidenced, currently-operating mechanism that improves fundamentals: payor in-network conversion driving ASP, from >250m contracted lives (FY2025 10-K) to ~300m after the Anthem contract (Q1-26), with realised ASP +28% YoY to $571 and gross margin +9pp to 73%. This is not a hypothetical — it is disclosed, dated and quantified. Quantitative corroboration is strong and in the same direction: accruals −1.14% (clean), gross profitability 0.356 (top-decile), F-score 7/8.

Gate 2 — Variant vs consensus: FAIL — the binding gate

2A (estimate variant): fails on magnitude. House FY2026 $462m vs consensus $454m = +1.8%; FY2027 +4.2%. That is inside estimate dispersion and inside a single quarter's true-up noise. Agreeing with the Street to within 2% is not a variant.

2B (duration/optionality variant): fails on evidence. The candidate vector is real and was found generatively — management stated on the Q1 call that health-system-driven adoption is explicitly excluded from guidance, and the tumour-naive MRD launch plus Northstar Response Medicare coverage are both guided for year-end 2026. Tested against all four mandatory legs:

A duration variant supported by fewer than all four legs is narrative, not evidence. Leg 1 fails outright. The single hard, filed piece of independent evidence is the 2026-06-23 8-K lease for 62,659 sq ft in Union City more than tripling oncology-dedicated lab space on a 12-year term — a genuine capital commitment, and the best evidence in the file. It is not enough on its own to carry leg 1 when the clinical record is one publication and zero registered MRD trials.

Gate 3 — Catalyst: PASS

Q2-2026 results (~early August 2026) directly test whether organic sequential growth re-accelerates from +5.4% to the ~8–10% the full-year guide requires. Northstar Response MolDX/Medicare decision and the tumour-naive MRD launch are both guided to year-end 2026.

Gate 4 — Expected return: FAIL, but marginally and by judgement, not by analysis

Twelve-month targets from FY2027E revenue × an EV/sales multiple, plus net cash, over 53.028m diluted shares (computed live in the Excel Scenarios tab):

E[R] = +0.64% against the 4.7% cash hurdle from portfolio_book.json → fails by 4.06pp.

Reported as a range, per framework item C4:

• Full E[R] range across plausible weightings: −9.01% to +7.76%.
• The 4.7% hurdle sits INSIDE that range. Per the framework, the honest statement is that this decision is not determined by the analysis — it is determined by a judgement input (the scenario probabilities), and it is reported that way rather than as a confident FAIL.
• Flip point on probability: holding the bull weight at 25%, E[R] crosses below the hurdle at a bear weight of 22%. The house uses 30%. A reasonable analyst using 20% bear would pass this gate.
• Flip point on price: E[R] clears 4.7% at approximately $127 and below — only ~5% under spot.

This is a materially closer call than the recent healthcare comparables (TXG and TWST failed by roughly 50pp even at a 0% bear weight). See chart 13.

Gate 5 — Implementation feasibility: PASS for equity, FAIL for options

• Equity: clean. Alpaca asset record returns shortable: true, easy_to_borrow: true, borrow_status: "easy_to_borrow", 30% margin. 30-day ADV 703,753 shares / ~$83.0m notional. Short interest 10.51% with 3.6 days to cover and declining — not crowded.
• Options: exist, but are not usable at size. 232 live contracts across four expirations (2026-08-21, 2026-09-18, 2026-12-18, 2027-03-19), strikes $40–190. Total chain volume on 2026-07-27 was 1,585 contracts (~$21m notional). Median at-the-money bid/ask spreads are 32.3% of mid (Aug), 26.4% (Sep), 16.4% (Dec), 10.8% (Mar-27).
• IV was recomputed from mid prices by Black-Scholes rather than trusted from the API, per the standing gotcha. On this name Alpaca's impliedVolatility field proved reliable — my recomputation matched to within ~0.01 vol points across the Sep chain (e.g. $130 call: API 0.687, recomputed 0.683). Recording that the field checked out here, since it demonstrably did not on another name.
• Put-call parity: a consistent ~$1.40–1.81 violation across the Sep-18 chain (calls cheap to puts). This sits inside the $2.84–4.05 bid/ask spreads, so it is quote noise and mild put demand rather than arbitrage — but it confirms the mids are not reliable enough to price a spread tightly.
• ATM IV 72.6–76.8% versus realised volatility of 85.3% since IPO and 79.0% over 60 days. Implied is below realised — unusually, options are not systematically overpriced here. That would ordinarily favour buying premium; the 16–32% bid/ask spreads destroy the edge before it can be captured. See chart 20.

Gate 6 — Momentum / tape: PASS (with the tape, for a long)

BLLN closed at $133.64 on 2026-07-28 — its all-time high, 100.0% of its 52-week high, +22.5% above the 50-day mean and +122.7% above the $60 IPO price. A long is squarely with the strongest possible momentum, so the gate passes on one line. A short would be fighting the tape at its most extreme, and no catalyst here is strong enough to break that trend within the horizon. Note the limitation: fewer than 200 trading sessions exist, so there is no true 200-day moving average and no 12-1 momentum figure.

---

POSITION VERDICT: RETIRED 2026-07-29 — the memo outputs an analysis, not a position.
                  See BLLN_Criteria_and_Valuation_2026-07-29.md for the current Criteria block.
The gate lines below are the 2026-07-28 record and are superseded.
GATES: 1(Mechanism): PASS   2(Variant vs Consensus): FAIL  <- BINDING   3(Catalyst): PASS
       4(Expected Return): FAIL (+0.64% vs 4.7%; hurdle INSIDE the -9.01%/+7.76% range)
       5(Feasibility): PASS equity / FAIL options   6(Momentum): PASS - with the tape
ENTRY: $105.00  |  TARGET: $136.01 (base)  |  INVALIDATION: see triggers
TIME HORIZON: to Q2-2026 results (~early Aug 2026), then the year-end MRD / MolDX cluster
SCENARIO-WEIGHTED E[R]: +0.64% net at spot; +28.1% at the $105 entry
       Simple: risking ~44% to make ~52%
SIZING: Conviction: LOW (composite 0.44)  |  Volatility: HIGH (85.3% realised, beta 1.35 house)
        Resulting size: 1.0% of book ON CONVERSION ONLY - zero today
VEHICLE: Outright equity only. Options are excluded by Gate 5 on spreads, not by direction.

Why not a Short? Because this is precisely the NET case the framework was built to catch. The fundamentals are strong, the factor scorecard is clean in every dimension a short thesis would need to be dirty, the house forecast is above consensus, and the stock is at an all-time high. "Expensive" is the entire bear case, and expensive is not a short. Stated explicitly because references/trade-construction.md requires it to be named whenever it applies — and it applies here.

Why not a Long? Gate 2 fails on evidence: there is no variant. Buying a name whose value rests on an oncology franchise with one publication and no registered trial, at 8.6% above the Street's target, with E[R] of +0.64% against 4.7% cash, is paying for optionality that is not yet underwritable.

Position sizing logic (applies on conversion only)

Gate 2 and Gate 4 near zero cap conviction at LOW regardless of the strong balance sheet and trajectory. Low conviction × High volatility ⇒ the grid returns "below minimum size — round down to Watchlist." On conversion at $105 the payoff score rises materially and the grid supports 1.0%, ceiling 1.5%.

Consensus positioning: a long here would be with a Buy consensus but against its price target (spot is 8.6% above the $122.14 average). Consensus has been strengthening — FY2026 revenue estimates up ~12% since March. Going long is not contrarian; going short would be, and would also be against a strengthening consensus and a top-percentile tape.

Entry / exit and invalidation

• Conversion entry: $105.00. Derived from the payoff structure, not an arbitrary discount: at $105 the scenario-weighted E[R] is +28.1%, giving genuine margin over the 4.7% hurdle rather than the ~5% sliver available at $127. It also sits between the base-case exit-multiple DCF ($79.90) and the base 12-month target ($136.01).
• Upgrade to Long also if the variant route completes (price-independent): all three of (i) tumour-naive MRD launches on time with peer-reviewed validation data published, (ii) Northstar Response secures MolDX/Medicare coverage, and (iii) a health-system contract is quantified in guidance. That would satisfy Gate 2B leg 1 and convert the thesis from narrative to evidence.
• Move to Avoid if: organic sequential growth (ex prior-period true-ups) prints below +4% for two consecutive quarters; or FY2026 revenue guidance is cut; or ICFR material weaknesses are still unremediated at the FY2026 10-K; or the MRD launch slips beyond H1-2027.

Thesis-invalidation triggers (specific, dated, checkable)

1. Q2-2026 (~early Aug): revenue below $114m, or organic sequential growth (ex true-up) below +5%.
2. Q2-2026: any FY2026 guidance reduction from $450–465m.
3. Q2/Q3-2026: prior-period true-up revenue again exceeding ~5% of quarterly revenue — would confirm the growth rate is being carried by retroactive re-pricing rather than by current-period business.
4. Year-end 2026: tumour-naive MRD launch slips, or launches without peer-reviewed validation data.
5. Year-end 2026: Northstar Response MolDX coverage denied or deferred.
6. FY2026 10-K (~Mar 2027): ICFR material weaknesses still unremediated.
7. Any quarter: overall ASP declines sequentially — the ASP engine is the whole margin story.
8. Any quarter: contracted lives fail to advance beyond ~300m and ASP growth falls below +10% YoY.

Portfolio context (portfolio_book.json, as of 2026-07-27)

Book is 100% cash, zero positions, 13 watchlist names. Correlation of BLLN daily returns against every relevant name, measured over the 179 overlapping sessions since IPO:

No pair approaches the book's 0.60 disclosure threshold — the maximum is +0.356. Notably BLLN correlates only +0.338 with Natera despite being a direct prenatal competitor, and +0.261 with Guardant. BLLN would open a genuinely idiosyncratic sleeve rather than double an existing bet; it is the least correlated healthcare name in coverage. It would sit in the life-science tools & diagnostics cluster with GH, NTRA, TXG and TWST for the 25% sector-concentration limit, but on correlation evidence it does not duplicate any of them.

Capital-competition test: at spot, E[R] of +0.64% loses to cash at 4.7%. Cash wins. No position.