Phase Space AI

Research

BillionToOne [BLLN]

BillionToOne, Inc. [BLLN] — Company Research (Task 1)

Ticker: BLLN (Nasdaq, Class A) | CIK: 2070849 | SIC: 8071 Services–Medical Laboratories Price: $127.98 (2026-07-27, Alpaca SIP daily close) | Fully diluted market cap: $7.00bn | EV: $6.42bn Memo framework version: investment-memo v1.4.2, run as written. No proposed amendment applied. Scope: Tasks 1, 2, 3, 5 only. Tasks 4 (charts), 6 (calendar), 7 (assembly/publication) and 9 (ledger) are out of scope for this run.

---

0. THE CENTRAL CONSTRAINT ON THIS MEMO: BLLN HAS BEEN PUBLIC FOR 8.6 MONTHS

BLLN priced its IPO at $60.00/share on 2025-11-06 and first traded 2025-11-06. Everything downstream is bounded by that:

Every one of these is stated explicitly rather than silently omitted, per the parent instruction and the skill's honesty rules.

---

1. MENTION-FREQUENCY / MANAGEMENT-ATTENTION PASS — RUN FIRST, GENERATIVELY

Per references/mention-frequency.md, this was run before any thesis existed, as hypothesis generation. Its outputs below are recorded as open questions with no interpretation attached, then chased against independent corpora. Provenance of every hypothesis in this memo is tagged [MF-generated] or [prior].

1.1 Corpus — one source, named, never mixed

Source: SEC EDGAR 8-K Exhibit 99.1 quarterly earnings press releases, first-party, complete, near-constant genre. Alpha Vantage EARNINGS_CALL_TRANSCRIPT was not used: this run has a hard budget of one Alpha Vantage call (shared 25/day cap across concurrently running agents), which was spent on EARNINGS_ESTIMATES for the mandatory consensus bridge. The reference file explicitly endorses 8-K Ex-99.1 releases as a first-party substitute where the primary source is exhausted.

Boilerplate ("Forward-Looking Statements" onward) is identical quarter to quarter and was truncated before counting; counts are reported per 10,000 words as the reference requires after the ISRG length artifact.

1.2 The table

History window: three quarters (2025Q3 – 2026Q1). This is the company's entire public history and it is far short of the reference file's 8-quarter minimum. A three-point series on ~1,000-word documents cannot distinguish trend from noise, and no claim below rests on the series alone.

1.3 Open questions generated (before any view existed), and what each returned

Q1. oncology rises monotonically 38.2 → 47.7 → 61.5 per 10k while prenatal peaks at 25Q4. Is attention rotating to oncology? → Independent corpus: YES, decisively. An 8-K filed 2026-06-29 (event date 2026-06-23) discloses a signed 12-year lease on 62,659 sq ft at 3260 Whipple Road, Union City CA, $46.7m aggregate base rent, $12.5m tenant-improvement allowance, commencing ~2027-08-01. The filing states the space "will be dedicated to the Company's oncology products, more than tripling the current oncology-dedicated laboratory space", and — unprompted — that "total test volumes have exceeded the Company's plan over the last year." This is a landlord contract, not a press release: capital is committed, the counterparty is external, and the disclosure is compelled. [MF-generated] — this hypothesis came from the language series, not from a prior.

Q2. profitability appears for the first time in 26Q1 (0, 0, 2) while operating income goes to zero (4, 5, 0). Is a hard GAAP metric being replaced by a soft one — the TWST disclosure-retirement pattern? → The finding dissolved on inspection, and is reported as dissolved. The Q1-26 release says "Income from operations of $17.8 million" and "Operating margin was 16%". My term was operating income; the company switched to income from operations. Word-boundary artifact, not a disclosure degradation. Further, the guidance language strengthened: 25Q3/25Q4 said "expects positive operating income"; 26Q1 says "expects to operate the business such that it will continue to generate profitability approaching current levels, even with significant continued investments" — a 16%-margin commitment versus a merely-positive one. This is exactly the failure mode references/mention-frequency.md warns about, caught by opening the record.

Q3. contracted lives and payor appear for the first time in 26Q1 with a specific number. → Q1-26 release, CEO quote: "significantly expanded our payor contracts, reaching $300 million contracted lives in the U.S." (the "$" is the company's typo for a count). New KPI, one quarter old. Durability unverified — flagged as a metric to check on 2026-08-05.

Q4. Fetal Cell / circulating fetal cell appear from nothing in 26Q1. → Unity Confirm, launched May 2026: a circulating-fetal-cell confirmatory assay using "Fetal Cell Capture" technology, positioned to confirm a high-risk NIPT result without CVS or amniocentesis. → Independent corpus: ClinicalTrials.gov. NCT07643896 — the ADVANCE study ("Assay Development and Validation for Pre-Natal and Obstetric Conditions"), RECRUITING, start 2026-01-10, n=1,000, conditions: Aneuploidy, Trisomy 21/13/18, 22q11.2 deletion, sex-chromosome abnormalities, primary completion 2027-12. Sponsor field verified as BillionToOne; conditions read directly from the registry, not inferred. This is a real, dated, externally-registered prenatal assay-validation programme consistent with the new language.

Q5. Northstar, PGx, clonal hematopoiesis, fetal antigen, HDFN, FNAIT all spike in 25Q4 and vanish in 26Q1. → Genre artifact, not signal. All sit inside the "Recent Operating Highlights" bullet block, which by construction names only that quarter's launches. Discarded.

Q6. MRD / minimal residual disease is absent from all three releases. → Confirmed against the FY2025 10-K, which uses the future tense: "As we expand our oncology offerings into applications such as MRD testing, as well as potentially testing for early detection in the future…" (10-K, Competition). BLLN has not launched MRD. This matters: MRD is the single largest value driver in Natera's franchise (Signatera). Any BLLN valuation that credits MRD is crediting unlaunched optionality. This memo credits none.

Q7. Natera / competition never appears in any earnings release. → Not itself a signal (companies rarely name rivals in releases). The 10-K names them fully — see §5.

1.4 Honest verdict on the metric for this name

Mention-frequency is not meaningfully computable on BLLN. Three ~1,000-word documents cannot support a trend claim. Its value on this name was generative: it produced Q1, which led to an 8-K that no screen would surface, and Q6, which is the single most important thing this memo declines to underwrite. Both are recorded as MF-generated for future ledger scoring of generative-vs-prior hypotheses.

---

2. THE BUSINESS

BillionToOne is a Menlo Park molecular-diagnostics laboratory built on one proprietary chemistry: Quantitative Counting Templates (QCT) — synthetic molecular counters spiked into a sample so that DNA molecules can be counted rather than merely detected. The company describes the resulting platform as smNGS (single-molecule next-generation sequencing). One platform, two commercial franchises.

2.1 Revenue by product (disclosed — not an estimate)

Source: 8-K Ex-99.1, 2026-03-04 and 2026-05-06. Product mix is genuinely disclosed here — unusual, and worth noting since the skill's discipline rules assume it usually is not.

Revenue by geography: not disclosed. The 10-K reports one operating segment and one reportable segment (NumberOfOperatingSegments = 1) and gives no geographic split. Alternative corpora attempted: 10-K segment note (nothing), 424B4 business section (US-centric language, no split), payor discussion (Medicare/MolDX/US commercial only). Conclusion: the business is effectively all-US; the absence of a split is consistent with that, and is stated as a limitation rather than estimated.

2.2 The revenue engine — volume × ASP

Two things follow and both are load-bearing for Task 3: 1. BLLN has never printed sequential revenue growth below +12.8% as a public company. 2. The growth mix has shifted. FY2025 revenue doubled on +51% volume and +35% ASP. Q1-26 was +44% volume and +28% ASP, but sequentially ASP added only +1.8% while volume added +10.6%. The ASP tailwind is maturing; volume is now the engine. Any forecast that keeps extrapolating ASP is wrong.

2.3 Products

Prenatal (89% of Q1-26 revenue) - UNITY Fetal Risk Screen — the differentiated asset. A single maternal blood draw that simultaneously screens for aneuploidy (T21/18/13, SCAs, 22q) and directly assesses fetal risk for recessive single-gene conditions (cystic fibrosis, SMA, sickle cell, thalassemias) without needing a paternal sample. Conventional carrier screening tells a couple they are carriers; UNITY tells them whether the fetus is affected. That is the QCT chemistry doing something competitors' read-depth-based methods cannot do cleanly. - UNITY Fetal Antigen NIPT — expanded Feb 2026 to red-blood-cell and platelet antigens; per the company the first and only US non-invasive tests covering both, for HDFN and FNAIT. Small today, genuinely uncontested. - Unity Confirm (launched May 2026) — circulating-fetal-cell confirmatory assay. See §1.3 Q4.

Oncology (10% of Q1-26 revenue, +392% YoY) - Northstar Select — liquid-biopsy comprehensive genomic profiling for therapy selection in stage III/IV cancer. The 10-K claims a 2–5× lower limit of detection than comparators, yielding ">50% more actionable mutations." Medicare coverage secured in 2025 via the MolDX programme (Palmetto GBA) under LCD L38043 — a hard, checkable regulatory fact, not a claim. - Northstar Response — ctDNA therapy-response monitoring. - Northstar PGx and Northstar Select CH — added Q1-26 (pharmacogenomics for chemotherapy safety; clonal-hematopoiesis filtering). - Not launched: MRD. See §1.3 Q6.

2.4 Clinical evidence base — thin, and this is a real weakness

ClinicalTrials.gov, sponsor = BillionToOne, complete list (registry read directly, three studies total):

Three trials, two of them n=20 and n=50. For contrast, the NTRA memo in this coverage recorded ~70 PubMed publications behind Signatera. BLLN's oncology growth is commercial, not evidence-led — it is winning on assay performance and MolDX coverage, not on a guideline-changing trial base. That is a materially different competitive posture and it caps how fast oncology can move into guideline-driven, reimbursed indications like MRD.

---

3. FINANCIAL PROFILE — THE THING THAT ACTUALLY DISTINGUISHES THIS COMPANY

All figures SEC EDGAR XBRL / filed statements. The model ties to each of these to the dollar (see BLLN_Model.xlsx, IncomeStatement rows 15–22, all reading 0).

Balance sheet (2026-03-31): cash $537,450k · AR $61,305k · inventory $19,217k · total assets $700,727k · total liabilities $195,107k · long-term debt (fair value) $90,005k · equity $505,620k · accumulated deficit $(256,777)k.

3.1 What is genuinely remarkable here

BLLN is the only high-growth molecular-diagnostics company in its peer set that is GAAP profitable. In Q1-26: NTRA −13.4% operating margin, GH −40.2%, TEM −24.3%, NVCR −38.7%, FLGT −48.7% — against BLLN at +16.5%, while growing 84%. Gross margin went 24.1% → 53.0% → 68.3% → 73.0% across three years. The company reached breakeven and then kept accelerating.

3.2 Two things I do not like, stated plainly

(a) DSO jumped 11.5 days in one quarter. AR rose +47.3% QoQ ($41.6m → $61.3m) against revenue +12.8%.

Not discussed in the release or the 10-Q MD&A. In diagnostics this is usually new-payor onboarding with slower first-cycle collection — consistent with the "$300 million contracted lives" expansion — but it is also the classic place revenue quality erodes first. Mitigant: operating cash flow was still +$15.4m in the same quarter, and TTM accruals are −1.20% of assets (i.e. cash exceeds earnings). It is a flag to watch on 2026-08-05, not evidence of a problem.

(b) They drew $30.0m of Oberland Capital structured debt in Q1-26 while sitting on $500m of cash. Cash-flow statement: "Proceeds from issuance of debt 30,000". The 2024 Notes are tranche-based and carried at fair value under ASC 825, which is why a −$4,261k "change in fair value of term loan" runs through Q1-26 other income and will grow as the equity rises. Drawing expensive structured capital against a $537m cash balance is, at best, unexplained. Most likely a use-it-or-lose-it tranche deadline. Flagged as a capital-allocation question for the Q2 call; not underwritten as a negative.

---

4. TAM — BOTTOM-UP, WITH THE REQUIRED IMPLIED-PENETRATION STATEMENT

Per references/tam-sizing.md. Units build, US only (see §2.1 on geography).

Prenatal SAM - ~3.6m US live births/yr; ~4.0m pregnancies reaching the gestational window for screening (assumption, flagged). - BLLN's realised blended ASP is $571 (disclosed, Q1-26). UNITY is multi-analyte, so a fully-penetrated ASP is at or above this. - 4.0m × $571 ≈ $2.3bn; allowing for multi-product attach and antigen testing, call the served prenatal market $2.5–3.0bn. - BLLN FY2026E prenatal ≈ $410m → ~15% penetration. Room to roughly double before saturation.

Oncology therapy-selection SAM - ~1.8m new US cancer cases/yr; ~700k advanced/metastatic eligible for comprehensive genomic profiling; tested 1–3× over the course of care → ~1.2m billable liquid-biopsy CGP tests/yr (assumption, flagged). - Medicare CGP reimbursement ~$3,000/test → ~$3.6bn, split today among Guardant, Foundation Medicine (Roche), Tempus, Caris, NeoGenomics. - BLLN FY2026E oncology ≈ $55–70m → under 2% penetration. This is the real runway.

Combined served market today ≈ $6.1–6.6bn. BLLN FY2026E $480.6m ≈ 7.5% penetration.

REQUIRED IMPLIED-PENETRATION STATEMENT. The reverse DCF (BLLN_Valuation.md §4) shows that $127.98 requires either 2035 revenue of ~$6.8bn — more than today's entire combined US prenatal + oncology-CGP served market — or a 7.97× terminal EV/Sales multiple on the house base case. Mature diagnostics trade at 2.0–4.3× and the sector's largest recent strategic transaction cleared at ~7.1× trailing. The market is not pricing a share-gain path inside the served market; it is pricing entry into MRD and/or early detection, neither of which BLLN has launched and neither of which this memo credits. That is the whole valuation argument, stated as a falsifiable claim.

---

5. COMPETITION

From the 10-K (Item 1, Competition), verbatim scope:

The 10-K states directly: "we know of competitors, in particular Natera and Myriad, that have been working on creating competing products to UNITY that also assess fetal risk for recessive conditions non-invasively." BLLN's core differentiation has a named, funded, well-capitalised assault underway. Management's answer is that smNGS "will allow us to maintain clear clinical and technical superiority" — an assertion, not evidence.

Sector M&A is live and repricing the group. Abbott completed its acquisition of Exact Sciences on 2026-03-23 at $105.00/share in cash, ~$21bn — roughly 7.1× EV/FY2025A revenue for a slower-growing, loss-making asset. This is (a) the best available anchor for what a strategic will actually pay in this sector, and (b) a live takeout-risk consideration for anyone short the group.

---

6. MANAGEMENT, OWNERSHIP AND INSIDER ACTIVITY

Management. Co-founder, Chairman & CEO Dr. Oguzhan Atay; co-founder and CTO David Tsao; CPO Shan Riku Sakakibara; SVP Laboratory Operations John Roderick ten Bosch. Dual-class: 4,552,650 Class B shares held by the founders (10:1 voting per the charter; converts on death/incapacity or departure). Board strengthened post-IPO with Anthony Pagano (CFO of Genmab) as Audit Committee Chair effective 2026-01-01 — a credible appointment for a company that had never filed a 10-K.

Register (primary sources — 13D/13G, not an aggregator):

Institutional ownership is disputed and I am reporting the dispute rather than picking a number. stockanalysis.com reports 25.96% institutional; wallstreetzen reports 0.04%. The reconciliation is almost certainly definitional: stockanalysis counts 13D/13G venture holders as institutions, wallstreetzen counts only 13F filers. The primary-source picture is unambiguous either way: the register is venture capital + founders + retail, with essentially no traditional active institutional base. Named 13F holders that did surface are small (Emerald Advisers ~132k shares; Wells Fargo, Smartleaf, Rossby all sub-$100k new positions). Per references/consensus-bridge.md, "held by institutions" is a much weaker signal than it sounds — here it barely applies at all.

Insider transactions (all 39 Form 4s parsed from EDGAR, transaction-level):

Applying the empirical asymmetry (Cohen, Malloy & Pomorski 2012) honestly: there is no insider signal here in either direction. - The "purchases" are subscriptions at the offer price through the directed-share programme, not opportunistic open-market buys. Clustered and multi-insider they may be, but they carry none of the informational content the literature credits, and it would be misleading to present them as a bullish signal. They are not counted as evidence. - The sales are textbook routine post-lock-up 10b5-1 diversification, spread across five insiders and priced $94.76–$129.78. CEO Atay sold ~90,000 shares (~$10.7m) against a ~3.2m-share position — under 3%. Per the reference, sales are noise unless extreme in size or unusually timed. These are neither.

---

7. RISKS

1. Valuation is the dominant risk and it is not a hypothetical — see BLLN_Valuation.md. Every method except trading comps sits far below spot.
2. Natera and Myriad are explicitly building UNITY competitors (10-K). BLLN's moat is one chemistry, and the incumbents have vastly larger salesforces and payor relationships.
3. Reimbursement is the whole business. ASP is set by payor mix. MolDX/LCD coverage can be revised; the FY2026 ASP assumption of ~$580 has no contractual floor.
4. DSO +11.5 days in one quarter, undiscussed (§3.2a).
5. LDT regulation. The 10-K discusses FDA oversight of laboratory-developed tests; a change in that regime would hit fixed costs and time-to-market for every new Northstar/UNITY variant.
6. Float and register. 24.09m-share float against a $5.9bn basic cap; a venture register that came unlocked on 2026-05-05 and has been distributing since. Any large holder's exit is a material share of daily volume ($75m/day).
7. Concentration in a single technology platform. QCT underpins 100% of revenue. There is no second leg.
8. Structured debt marked to the equity. As the stock rises, the fair-value charge on the Oberland notes rises, mechanically depressing GAAP net income (−$4.3m in Q1-26 alone).
9. Key-person. Founder-CEO with a dual-class structure and a 19.6% venture holder; governance leverage sits with a very small group.

---

8. WHAT IS NOT PUBLISHED — SCOPING RESULT

Per references/unpublished-scoping.md, "not disclosed" is a research task. What was attempted and what it returned:

---

Prepared 2026-07-27. Sources: SEC EDGAR (submissions, XBRL company facts, 424B4, 10-K FY2025, 10-Q Q1-26, 8-K Ex-99.1 ×4, 39 Form 4s, SC 13D/13G ×4); Alpaca Markets (equity bars, options snapshots, asset/borrow status); Alpha Vantage EARNINGS_ESTIMATES (one call, cached to data/av_estimates.json); ClinicalTrials.gov API v2; stockanalysis.com; marketbeat.com; investing.com; wallstreetzen.com.