ADMA Biologics [ADMA] — Research

Tier-2 memo · as of 2026-07-29 · price $8.79 (2026-07-28 close) · framework v1.7.0 Criteria

This memo issues no position verdict. It scores Criteria and outputs an analysis. The book decides.

0. The one-paragraph answer

The screen carried ADMA as a 49.0% compounder with a +44.7pp valuation margin. That growth rate does not exist any more and has not since Q2 2025. It is a three-year lookback (FY2022 $154.1m → FY2025 $510.2m) earned almost entirely by ASCENIV ramping off a near-zero base. The most recent reported quarter, Q1 2026, grew −0.3% year over year. On 6 May 2026 the company cut FY2026 revenue guidance from ">$635m" to $530–560m — a 12–17% cut delivered ten weeks after the same guidance was explicitly reiterated — and withdrew all long-term guidance, including the FY2029 ">$1.1bn revenue / >$700m Adjusted EBITDA" target. The mechanism behind the margin story is real but is a one-time step: an FDA-approved yield-enhancement process (PAS approved April 2025) that lifts IG output ~20% from the same plasma. FY2026 is its first full year; there is no second one. Meanwhile the company has reduced its plasma collection network from ten centres to seven and states it does not plan to build more — so the binding physical input to future volume has been externalised to third-party suppliers on escalating-price contracts. Alongside this, FY2025 cash conversion collapsed: net income $146.9m against cash from operations of $50.4m (0.34x), with receivables rising $108.4m on revenue growth of $83.7m and DSO going from 38.8 to 103.9 days, in a business where 73% of revenue and 87% of receivables sit with two distributors.

1. What the business actually is

ADMA Biologics is a US plasma-derived biologics manufacturer with two reporting segments:

Three FDA-approved products, all plasma-derived:

• ASCENIV (IVIG, approved April 2019) — the franchise. Made by blending normal source plasma with high-titer RSV plasma identified by ADMA's patented microneutralisation assay. The 10-K states this high-titer plasma "is typically found in less than 10% of the total donor collection samples we test." Positioned as a late-line therapy for patients refractive to standard IG; commands premium pricing.
• BIVIGAM (IVIG, approved May 2019) — a standard, undifferentiated IG. Competes on price.
• Nabi-HB (hepatitis B immune globulin) — immaterial, folded into "intermediates and other."

Manufacturing physics matter here and they are disclosed: cold-ethanol (Cohn) fractionation, and the 10-K puts "the time for collection, manufacturing and release of a batch of IG at 7 to 12 months." Production runs at a 4,400-litre scale approved in 2021. This is not a software scaling curve. Output is bounded by litres of qualifying plasma × yield per litre × approved throughput.

Product mix — the whole story in one table

Source: FY2025 10-K note 13 (disaggregated revenue) and Q1 2026 10-Q MD&A.

ASCENIV is now 85% of revenue. ADMA is, functionally, a single-product company selling one premium IVIG through two distributors.

2. Mechanism — named, evidenced, and finite

The margin mechanism is the Yield Enhancement. Quoting the FY2025 10-K directly:

"In April 2025, the FDA approved our Prior Approval Supplement (the "PAS") for our innovative yield enhancement production process… This innovative process has demonstrated an ability to increase ASCENIV and BIVIGAM production yields by 20% or more from the same starting source plasma volume. Fiscal year 2026 will be our first full year of yield-enhanced production, supporting anticipated sustained margin expansion."

It works, and it is visible in the accounts:

Gross margin rose 17.3pp year over year on flat revenue. That is the yield step plus the ASCENIV/BIVIGAM mix shift, and it is genuine.

But it is a level shift, not a growth rate. Three independent pieces of evidence:

1. The company says so. "FY2026 will be our first full year of yield-enhanced production." A 20% one-time uplift is fully in the run-rate by the end of FY2026. There is no disclosed second yield programme.
2. It cannot be repeated by adding plasma, because ADMA is shrinking that. From the 10-K: ADMA BioCenters operated ten centres at 31 December 2025; eight at the filing date; and will operate seven after the third divestiture closes. The company states plainly: "At the present time, we do not plan to build additional plasma collection facilities." A typical centre collects 30,000–50,000 litres a year, so seven owned centres bound internal collection at roughly 210,000–350,000 litres — of which under 10% qualifies as the high-titer plasma ASCENIV requires.
3. The incremental input is contracted, escalating and third-party. The Amended & Restated Grifols agreement (effective 1 October 2024, expiring September 2039) supplies a minimum of 35,000 litres of RSV plasma annually, "with an escalating price per liter depending on the volume supplied… with a minimum annual price increase every 12 months." ADMA also says it now has "access to an aggregate of approximately 280+ plasma collection centers both internally and through such third-party agreements."

Answer to the capacity question: growth is capacity-limited, and ADMA has deliberately externalised the binding constraint. That is a legitimate capital-efficiency choice — it converts capex into opex and lifts ROIC — but it means incremental ASCENIV volume arrives at a contractually rising input cost, and the company no longer owns the throttle. The 49% historical CAGR was produced by a period (FY2022–FY2024) in which ADMA was building centres and ramping a new product into an unpenetrated indication. Both of those engines are off.

3. Accounting quality — is the reported growth real?

This is the section the memo exists for. Five findings, all from primary filings.

3.1 FY2025 converted 34 cents of every reported dollar of net income into cash

*FY2023 operating income; net income line differs on tax items.

FY2025 was reported as a "record" year with +20% revenue growth and $231m of Adjusted EBITDA. It produced $27.8m of free cash flow. Adjusted EBITDA converted to FCF at 12%.

Sloan-style accruals for FY2025: (NI − CFO) / average total assets = (146.9 − 50.4) / 556.5 = 17.3%. That sits in the highest-risk accruals decile in the published literature. Accruals: FAIL.

3.2 The cash went into receivables, and the receivables sit with two customers

Receivables rose $108.4m across FY2025 while revenue rose $83.7m. DSO tripled in five quarters. (DIO is inflated by the fractionation cycle and by the Q1 2026 COGS collapse; the level is not comparable to a normal manufacturer, but the direction — inventory building 30% while revenue is flat — is.)

Concentration, verbatim from the FY2025 10-K:

"For the years ended December 31, 2025 and 2024, two customers, BioCare, Inc. and Priority Healthcare Distribution, Inc. d/b/a CuraScript SD represented an aggregate of approximately 73% and 72%, respectively, of our consolidated revenues. As of December 31, 2025, two customers, BioCare and CuraScript, represented an aggregate of approximately 87% of our consolidated accounts receivable."

Two distributors: 73% of revenue, 87% of receivables, DSO of 104 days at year end. This is structurally the Applied Optoelectronics pattern the brief names — with the important difference that ADMA discloses it clearly, in both the risk factors and the concentration footnote. It is not hidden. It is simply large.

3.3 The company's own short-seller rebuttal discloses the channel position

On 24 March 2026 Culper Research published a report alleging channel stuffing and undisclosed related-party transactions. On 27 March 2026 ADMA filed an 8-K rebuttal. In refuting the allegation, it disclosed distributor inventory data it had never previously published:

Source: 8-K EX-99.1 filed 2026-03-27. ADMA notes 5 January 2026 is the closest available data point to 31 December 2025, and that days are computed on each distributor's previous-month run rate.

Read as ADMA intends, this shows product pulling through the channel. Read arithmetically it also shows that at the close of the "record" FY2025, the channel held roughly a full extra quarter of ASCENIV above contractual minimums, and that ~30% of that coverage drained in the following eleven weeks.

I decline to convert coverage days into dollars, and the reason is itself a finding. Converting 84 excess ASCENIV days at ADMA's own FY2025 revenue-per-day implies ~$83m of excess ASCENIV in the channel; the same conversion for the January-to-March drawdown implies ~$40m of Q1 2026 sell-through that was not replaced by sell-in, which would put underlying ASCENIV demand up ~80% year over year. That does not reconcile with any other disclosure and I do not believe it. Either the coverage-day denominators (distributor-specific trailing-month run rates, across a customer set broader than the two 73% distributors) are not commensurable with ADMA's revenue per day, or one of the two series does not measure what it appears to. The disclosure is recorded here as the company published it; no dollar figure is asserted.

The Audit Committee subsequently completed a review with independent forensic accountants and external counsel and concluded there was no improper channel stuffing, no undisclosed related-party transactions and no evidence of illegal activity; the FY2025 financial statements were left unchanged. ADMA also notes unqualified audit opinions from a Big Four firm for FY2024 and FY2025, and that its prior accounting firm resigned in late 2024 "in advance of being acquired by a private equity firm."

Position taken here: no allegation of impropriety is made or implied. The findings above are about cash conversion and channel position, which are matters of arithmetic, not conduct.

3.4 Two one-offs flatter the reported P&L

• Q1 2026 operating income includes an $8.0m pre-tax gain on the sale of three plasma collection centres ($6.3m after tax), booked within operating expenses. Reported Q1 2026 operating income of $58.3m is $50.3m clean — a 43.9% operating margin, not 50.9%. GAAP net income of $45.3m "grew 68%"; Adjusted net income grew 22%, and the company's own reconciliation removes the gain.
• FY2024 gross profit included a $12.6m release of the Medicaid rebate accrual. ADMA discloses this itself: ex the release, FY2024 gross margin was 50.0%, not the 51.5% reported. The FY2024→FY2025 gross margin expansion is therefore 50.0% → 57.4% (+7.4pp), not +5.9pp — this one cuts in ADMA's favour, and is recorded for symmetry.

3.5 A $125m debt-funded buyback executed five days after guidance that was withdrawn ten weeks later

The dated sequence, all from filings:

Guidance was reiterated on 25 February with two of the three months of Q1 already elapsed, and cut 12–17% (Adj EBITDA 17–26%) on 6 May. Long-dated targets that had been the centrepiece of the equity story were withdrawn entirely. Five days after the reiteration, the company drew $125m of revolver debt to buy stock at $16.58; that stock is $8.79 today, a mark-to-market loss of roughly 47% on the repurchase, against a balance sheet that went from $72m to $197m of debt in a single quarter.

This is the item that most changes the underwriting. It is not an accounting error. It is a demonstrated failure of management's own forward visibility into its two largest customers' ordering, occurring at the exact moment the company was levering the balance sheet on the strength of that visibility. Any forward number in this memo that depends on management guidance inherits that track record, and is flagged where it does.

4. Validating the screen's inputs

The brief requires these be checked, not trusted. The screen's own record carries data_quality_ok: false and the name was advanced anyway.

Nothing in the screen is fabricated. Every number reproduces. The failure is that a trailing three-year CAGR and a growth-matched multiple derived from it were applied to a business whose growth had already stopped — and the screen's own data_quality_ok: false did not gate it.

5. Mention-frequency over time

Corpus substitution, disclosed. The Alpha Vantage EARNINGS_CALL_TRANSCRIPT quota (25/day, shared) was exhausted when tested on 2026-07-29. The corpus used instead is ADMA's own quarterly earnings releases, filed as 8-K Item 2.02 Exhibit 99.1 — 13 consecutive quarters, 2023Q1 through 2026Q1, pulled directly from EDGAR. These are management's own prepared written commentary for the same event, so they measure the same thing (where management chooses to spend words) from a primary source. They are not call transcripts and the Q&A is absent, which will suppress analyst-driven topics. Counts are stem-matched and normalised per 1,000 words because release length varies 1,483–2,644 words. Raw data: data/adma_mentions.json.

Run generatively — these are the anomalies, listed before interpretation.

First-ever appearances in 2026Q1, across a 13-quarter series

Escalating

Decaying

The open questions this raises

1. Why did capacity go to zero and stay there? The capacity narrative that justified the multiple through 2024–25 is gone from management's own language at exactly the point revenue stopped growing. Answer found in the 10-K: they stopped building centres and sold three. The language change preceded nothing — it reflected a disclosed strategic reversal that the screen's growth extrapolation did not price.
2. Why does yield-enhanc* collapse to 0.5/1k in the quarter its benefit is largest? Gross margin hit 70.5% in 26Q1 on the yield process, yet management stopped talking about it. Consistent with the mechanism being complete rather than ongoing — "first full year," no follow-on programme.
3. demand at an all-time high while revenue is flat. Management substituted a demand narrative (record utilisation, new patient starts, prescriber breadth, April run-rate) for a revenue result. That is a recognisable pattern and it is worth scoring against the Q2 2026 print.
4. label expansion / pediatric at zero for nine consecutive quarters — then the approval landed anyway (4 May 2026). This one is a false negative for the technique and is recorded as such: the catalyst was live in the 10-K throughout (sBLA filed June 2025, approval guided to 1H 2026) but absent from the earnings releases. Mention-frequency on a press-release corpus missed a real, dated, delivered catalyst.

Point 4 is the honest limitation of the substituted corpus and is the reason the transcript source should be re-run when quota allows.

6. Product-cycle intelligence

7. The industry condition — why BIVIGAM is collapsing

The Q1 2026 release describes it directly:

"During the first quarter of 2026, the U.S. IG market experienced a reset due to elevated raw material plasma supply, increased competitive PDT inventories across the distribution network and aggressive discounting and rebating in standard IG."

This is a supply-side glut in the fractionation industry. It is the reciprocal of ADMA's own advantage: when plasma is abundant, the scarcity premium in undifferentiated IG evaporates. ADMA argues ASCENIV is insulated by its late-line positioning and patented high-titer pooling, and the Q1 numbers support that — ASCENIV +27.7% while BIVIGAM −54%. But two things follow that management does not draw out:

1. Abundant plasma is only good for ADMA if it converts to volume, and ADMA has just capped its own collection at seven centres with contractually escalating third-party pricing.
2. The mix shift that is lifting gross margin is the same event that is killing revenue growth. Every point of BIVIGAM decline is a point of revenue lost at ~50% incremental margin and a point of blended gross margin gained. The 70.5% Q1 gross margin and the −0.3% revenue are one phenomenon, not two.

8. Criteria scoring

Per criteria.md (v1.7.0). Every Criteria returns PASS / FAIL / INDETERMINATE. A missing input is INDETERMINATE, never FAIL.

Archetype: COMPOUNDER — profitable, high margins, value in the visible forecast period. Confirmed, and the screen's classification is right.

9. What is unsupported, missing or uncertain — stated, not filled in

1. No Street consensus. Alpha Vantage quota exhausted. Every forward revenue and EBIT figure in this memo is either the company's own guidance or a stated house extrapolation, and is labelled as such. No consensus number is invented.
2. No FY2027 guidance exists — it was withdrawn. The FY2027 leg of the NTM base in the 12-month target is a house extrapolation at the FY2026 guided growth rate, flagged in ADMA_Valuation.md §4.
3. Grifols is excluded from the quantitative peer anchor. GRFS files IFRS with no reliably tagged share count or borrowings in its SEC XBRL; its EV/EBIT is dominated by several billion euro of net debt that I could not verify from a primary source. Its growth (FY2024 revenue €7,212m, +9.4% YoY, 3-year CAGR 13.5%, EBIT margin 16.5%) is used for the growth-bracket test only. No Grifols multiple is asserted.
4. CSL Ltd is not an SEC filer and is excluded entirely rather than sourced from an unverifiable secondary.
5. Channel-inventory dollars are not asserted — see §3.3. The reconciliation to reported revenue does not close and the failure is reported rather than papered over.
6. The Q2 2026 earnings date is not announced. The calendar records the three-year pattern (2023-08-09 / 2024-08-08 / 2025-08-06) and says "expected early August 2026, date not announced." No date is fabricated.
7. SG-001 is excluded from all valuation. Preclinical, IND not filed.
8. Mention-frequency is on a press-release corpus, not transcripts — §5 — and demonstrably missed the paediatric approval.
9. Realised-vol-based WACC. CAPM gives 8.7% (beta 0.88, correlation to SPY only 0.20). A 55%-vol single-product name whose risk is almost entirely idiosyncratic is poorly served by CAPM; base case uses 10.0% with 8.7% and 12.0% shown. The choice is stated, not buried.
10. The paediatric expansion is unsized. ADMA disclosed no patient-count or revenue estimate; none is invented.

Sources

All primary, all retrieved 2026-07-29.

• FY2025 Form 10-K, filed 2026-02-25 (accession 0001140361-26-006815)
• Q1 2026 Form 10-Q, filed 2026-05-06 (accession 0001140361-26-019336)
• 8-K EX-99.1 2026-05-06 — Q1 2026 results, guidance cut, LT guidance withdrawn (0001140361-26-019295)
• 8-K EX-99.1 2026-02-25 — FY2025 results, LT guidance reiterated (0001140361-26-006781)
• 8-K EX-99.1 2026-05-04 — ASCENIV paediatric approval (0001140361-26-018667)
• 8-K EX-99.1 2026-03-27 — Culper rebuttal, channel-inventory disclosure (0001140361-26-011569)
• SEC XBRL companyfacts, CIK 0001368514
• 13 quarterly 8-K Item 2.02 EX-99.1 releases, 2023Q1–2026Q1
• Alpaca Markets: daily bars (SIP) 2019-01-02 → 2026-07-29; options contracts and snapshots
• Peer XBRL companyfacts: KMDA 0001567529, EBS 0001367644, GRFS 0001438569