MannKind Corporation [MNKD] — Financial Model Notes
Every derived figure in this memo, with its arithmetic, so any number can be checked without re-running anything. No Excel workbook was built — the analysis is driven by segment disclosure and a reverse DCF, neither of which needs one, and the time-box directed effort to the accounting-quality and valuation work per the brief's priority order. That omission is stated rather than concealed.
Underlying data files are in data/:
- mentions.json — raw mention counts and source URLs for all 30 earnings press releases
- evsales.json — the daily EV/Sales series (1,309 sessions) behind the percentile table
- bsdata.py — the 21-quarter point-in-time balance-sheet table used to build it
1. Revenue: quarterly, derived, reconciled
Q4 is not separately tagged by US registrants. Every Q4 below is derived as FY − (Q1+Q2+Q3) and the FY figure is the tagged 10-K number.
| Quarter | Revenue $000 | Source |
|---|---|---|
| 2023 Q1 | 40,626 | tagged |
| 2023 Q2 | 48,611 | tagged |
| 2023 Q3 | 51,253 | tagged |
| 2023 Q4 | 58,472 | 198,962 − 140,490 |
| 2024 Q1 | 66,263 | tagged |
| 2024 Q2 | 72,386 | tagged |
| 2024 Q3 | 70,079 | tagged |
| 2024 Q4 | 76,776 | 285,504 − 208,728 |
| 2025 Q1 | 78,354 | tagged |
| 2025 Q2 | 76,527 | tagged |
| 2025 Q3 | 82,130 | tagged |
| 2025 Q4 | 111,955 | 348,966 − 237,011 |
| 2026 Q1 | 90,171 | tagged |
TTM (Q2 2025 – Q1 2026) = 76,527 + 82,130 + 111,955 + 90,171 = $360,783k. Equivalently FY2025 − Q1'25 + Q1'26 = 348,966 − 78,354 + 90,171 = $360,783k. Ties to the screen exactly. Not a 15-month window — the four periods are consecutive and were checked to be so.
Prior-year TTM (Q2 2024 – Q1 2025) = 72,386 + 70,079 + 76,776 + 78,354 = $297,595k.
2. The revenue split, by line, TTM
TTM line = FY2025 − Q1 2025 + Q1 2026, from the face of the income statement in each filing.
| $000 | FY2025 | − Q1 2025 | + Q1 2026 | TTM | % of TTM |
|---|---|---|---|---|---|
| Commercial product sales | 114,137 | (18,973) | 33,907 | 129,071 | 35.8% |
| Collaborations and services | 106,713 | (29,376) | 23,515 | 100,852 | 28.0% |
| Royalties | 128,116 | (30,005) | 32,749 | 130,860 | 36.3% |
| Total | 348,966 | (78,354) | 90,171 | 360,783 | 100% |
United Therapeutics-derived = 100,852 + 130,860 = $231,712k = 64.2% of TTM revenue.
3. Product-level reconstruction (derived — carries ±$1m of rounding)
Individual products are not separately XBRL-tagged. Reconstructed from MD&A gross revenue, gross-to-net
percentage and stated dollar changes, then cross-checked to sum to the tagged Commercial product sales
line.
Q1 2026 (10-Q MD&A): - Furoscix: gross $20.9m, gross-to-net 26% ($5.4m) → net $15.5m (stated directly) - Afrezza: gross-to-net 31% = $6.8m; net revenue "+$0.4m, or 3%" on Q1'25's ~$14.8m → net ≈ $15.2m - V-Go: gross-to-net 36% = $1.8m; net "−$0.9m, or −23%" on Q1'25's ~$4.1m → net ≈ $3.2m - Check: 15.5 + 15.2 + 3.2 = 33.9 ≈ tagged $33,907k ✓
FY2025 (10-K MD&A): - Furoscix (7 Oct – 31 Dec stub): gross $32.4m, gross-to-net 28% → net $23.2m (stated directly) - V-Go: gross-to-net 38% = $9.8m → gross $25.8m → net ≈ $16.0m - Afrezza: residual → 114,137 − 23,200 − 16,000 = net ≈ $74.9m (MD&A says net rose "$10.5m, or 16%", implying ~$76.1m from a ~$65.6m base — the two routes differ by ~$1m of rounding; $75m is used) - Check: 23.2 + 16.0 + 74.9 = 114.1 = tagged $114,137k ✓
TTM by product: Afrezza ≈ 75.0 − 14.8 + 15.2 = $75.4m · V-Go ≈ 16.0 − 4.1 + 3.2 = $15.1m · Furoscix = 23.2 + 15.5 = $38.7m. Sum $129.2m vs tagged $129.071m — a $0.13m rounding gap. ✓
4. The organic calculation
Q1 2026: 90,171 − 15,500 = 74,671 vs Q1 2025's 78,354 → −4.70%.
TTM: 360,783 − 38,700 = 322,083 vs prior-year TTM 297,595 → +8.23%. (Prior-year TTM contains no Furoscix — MannKind did not own it until 7 Oct 2025 — so the comparison is clean on both sides.)
5. Contribution to the 51.8% CAGR
FY2022 → FY2025, total increase $249,196k:
| Line | FY2022 | FY2025 | Δ | % of total Δ | line CAGR |
|---|---|---|---|---|---|
| Royalties | 15,599 | 128,116 | +112,517 | 45.16% | (128,116/15,599)^⅓ − 1 = 102.0% |
| Collaborations & services | 27,924 | 106,713 | +78,789 | 31.62% | (106,713/27,924)^⅓ − 1 = 56.4% |
| Commercial product sales | 56,247 | 114,137 | +57,890 | 23.23% | (114,137/56,247)^⅓ − 1 = 26.5% |
| — of which acquired Furoscix | — | 23,200 | +23,200 | 9.31% | n/a |
| — of which organic | 56,247 | 90,937 | +34,690 | 13.92% | 17.3% |
| Total | 99,770 | 348,966 | +249,196 | 100% | (348,966/99,770)^⅓ − 1 = 51.8% ✓ |
United Therapeutics = 45.16 + 31.62 = 76.78% of the three-year increase. Screen's 51.8% reproduced exactly, confirming FY2022 → FY2025 is the window it used.
6. scPharmaceuticals: isolating the acquired revenue
From the 10-K FY2025 supplemental pro-forma disclosure (Note 3):
| FY2025 | FY2024 | |
|---|---|---|
| Pro-forma combined revenue | $396,150k | $321,836k |
| MannKind as reported | $348,966k | $285,504k |
| Difference = scPharma pre-acquisition period | $47,184k (1 Jan – 6 Oct 2025, 279 days) | $36,332k (full FY2024, 366 days) |
Per-day run rates: - scPharma Jan–Oct 2025: 47,184 / 279 = $169k/day - Q4 2025 stub under MannKind: 23,200 / 85 = $273k/day - Q1 2026: 15,500 / 90 = $172k/day - scPharma FY2024: 36,332 / 366 = $99k/day
Underlying growth: $172k/day vs $99k/day = +73% year-on-year. The $273k/day Q4 stub is the outlier (+61% above both the quarter before it and the quarter after it) and is most plausibly channel stocking around the acquisition plus Q4 deductible seasonality — flagged as an open question in the Research doc, not asserted.
7. Margins and returns
| $000 | FY2024 | FY2025 | TTM |
|---|---|---|---|
| Revenue | 285,504 | 348,966 | 360,783 |
| Operating income (EBIT) | 72,587 | 38,801 | 14,841 |
| Operating margin | 25.4% | 11.1% | 4.1% |
| Net income (loss) | 27,588 | 5,863 | (23,914) |
| Cash from operations | 42,511 | 18,255 | 19,267 |
| Stock-based compensation | 21,358 | 24,195 | ~25,265 |
TTM EBIT = 38,801 − 22,293 + (1,667) = $14,841k. TTM net income = 5,863 − 13,158 + (16,619) = $(23,914)k. TTM operating cash flow = 18,255 − (6,377) + (5,365) = $19,267k.
The screen's
op_margin_pct: 11.1is the FY2025 figure applied against TTM revenue. That period mismatch is what producesev_ebit: 28.7: $1,152m ÷ (11.1% × $360.8m = $40.0m) = 28.8x. Using the corrected EV of $1,440m and the actual TTM EBIT of $14.8m gives 97.0x.
Gross margin (not tagged; the screen returned null):
| $000 TTM | |
|---|---|
| Cost of goods sold — commercial | 26,800 − 3,768 + 7,509 = 30,541 |
| Cost of revenue — collaborations & services | 61,160 − 13,748 + 9,964 = 57,376 |
| Amortisation of acquired intangibles | 3,973 − 0 + 4,367 = 8,340 |
| Gross profit ex-amortisation | 360,783 − 87,917 = 272,866 → 75.6% |
| Gross profit incl. amortisation | 360,783 − 96,257 = 264,526 → 73.3% |
Flattered by mix: royalties are 36.3% of revenue and carry no cost of revenue at all. Collaborations and services gross margin is (100,852 − 57,376) / 100,852 = 43.1%.
ROIC: NOPAT ≈ TTM EBIT $14.8m (full valuation allowance, no cash tax). Invested capital = term loan $318.7m + Sagard royalty liability $150.6m + Danbury financing liability $103.2m + stockholders' deficit $(59.2)m = $513.3m. ROIC ≈ 2.9% vs WACC ≈ 11.7%.
8. Net debt — the $246.5m correction
| $000, 31 Mar 2026 | |
|---|---|
| Cash and cash equivalents | 52,834 |
| Short-term investments | 81,027 |
| Long-term investments | — |
| Liquid resources | 133,861 |
| Blackstone term loan (principal $325,000; carrying) | 318,722 |
| Net debt (narrow) | (191,139) |
| Sagard liability for sale of future royalties (14,010 + 136,561) | 150,571 |
| Danbury financing liability (10,407 + 92,784) | 103,191 |
| Net debt (broad) | (444,901) |
Cross-checked to the 10-Q's own MD&A sentence: "As of March 31, 2026, we had capital resources comprised of cash, cash equivalents and investments totaling $133.9 million, and total principal amount of outstanding borrowings of $325.0 million."
Balance sheet cross-foot (all in $000): current assets 52,834 + 81,027 + 28,137 + 49,166 + 39,996 = 251,160 ✓ tagged. Total assets 251,160 + 747 + 0 + 82,554 + 67,595 + 185,708 + 129,600 + 5,024 + 22,015 = 744,403 ✓ tagged — which is how "Long-term investments $5,012k" was assigned to the December column rather than March. Total liabilities sum to 803,598 ✓. Equity (59,195) ✓.
Pro forma for the two known July events: 133,861 + ~47,000 (net placement proceeds after expenses) − 45,000 (CVR payment) = ~$135,861k. Q2 2026 operating burn is not included; Q2 2026 is unfiled.
9. Shares
| Outstanding, 31 Mar 2026 (10-Q balance-sheet parenthetical) | 308,907,331 |
| + private placement shares (8-K 2026-07-24) | 10,440,838 |
| + pre-funded warrants, $0.01 strike, no expiry (same 8-K) | 2,412,632 |
| Economic share count | 321,760,801 |
| Other potentially dilutive at 31 Mar 2026 (10-Q EPS note): 3.0m RSUs + 3.5m options/PNQs | 6,500,000 |
| 12-month assumption used in the target | 325,800,000 |
EPS cross-check. Q1 2026: $(16,619)k ÷ 308,267k weighted basic = $(0.054) vs filed $(0.05) ✓. FY2025: $5,863k ÷ 305,639k = $0.019 vs filed $0.02 ✓. Single share class — the dual-class XBRL-aggregation caveat does not apply.
10. The EV/Sales history series
data/evsales.json, 1,309 trading days, 2021-05-12 → 2026-07-29. Construction:
- Daily closes, Alpaca SIP feed, split- and dividend-adjusted.
- TTM revenue as known at that date — stepped in at the actual 10-K / 10-Q filing date from the SEC submissions index, never at the period end. The series is therefore not forward-looking.
- Point-in-time share count and net debt from XBRL instant facts, taken at the most recent balance-sheet
date at or before the revenue period end (
data/bsdata.py, 21 quarters). This deliberately departs from the framework helper's approach of holding shares and net cash at today's verified values. For MannKind that would be wrong: net debt moved from +$50m net cash to −$191m and shares from 249m to 309m inside the window, and freezing them would put the entire balance-sheet change into the multiple. EV = close × shares_pit + net_debt_pit;EV/Sales = EV ÷ TTM_as_known.
Percentiles are in MNKD_Valuation.md §3.3.
11. Reverse DCF parameterisation
assets/reverse_dcf.py, solving for cagr, 5-year horizon, EV computed as spot × shares − net_cash.
| Run | --shares |
--net-cash |
--terminal-margin |
--exit-multiple |
--wacc |
Result |
|---|---|---|---|---|---|---|
| Screen reproduction | 308.950 | +57.415 | 0.162 | 22.6 | 0.10 | 7.03% |
| Corrected, base | 321.761 | −182.8 | 0.162 | 22.6 | 0.10 | 11.82% |
| Corrected, name WACC | 321.761 | −182.8 | 0.162 | 22.6 | 0.117 | 13.50% |
| Corrected, broad EV | 321.761 | −436.6 | 0.162 | 22.6 | 0.10 | 15.52% |
(−182.8 rather than −189.1 because the reverse-DCF runs use the term loan's carrying amount $318.722m rather than the $325.0m principal, for consistency with the balance sheet. The difference is $6.3m of unamortised issuance costs and moves the required CAGR by under 0.1pp.)
Full sensitivity grids over exit multiple, WACC and terminal margin are in MNKD_Valuation.md §2.4.
12. What was NOT built, and why
- No Excel workbook. Per the brief's stated priority order (accounting quality and mechanism first, valuation and target second, publication third) under a time-box. The reverse DCF and the segment bridge are the two instruments the Criteria actually require and both are fully specified above.
- No forward DCF.
criteria.md: where terminal value exceeds 60% of EV the reverse DCF is mandatory as the primary output and the forward DCF becomes supporting evidence. On MannKind, with TTM EBIT of $14.8m against a $1,440m EV, effectively all value is terminal. A forward DCF would have added false precision to the parameter that determines the answer. - No consensus overlay. Alpha Vantage quota exhausted 2026-07-29; Consensus Criteria is INDETERMINATE. The NTM revenue base in the target is an explicitly-labelled house build, not consensus.