Harmony Biosciences (HRMY): Full Memo
Data date: price as of the 2026-10-16 close (Yahoo Finance). Financials from SEC XBRL company facts, fiscal years 2021–2025 (FY2025 10-K). Guidance and Q2 2026 results from Harmony's Q2 2026 preannouncement and related coverage (July–Aug. 2026). Q3 2026 results were not yet out at the data date.
Labels: Reported = from a filing or price feed. Estimate = third-party consensus or company guidance. Assumption = my model input.
1. Investment thesis
What it does: Harmony is a commercial-stage biopharmaceutical company focused on sleep disorders. Its main product is WAKIX (pitolisant), a treatment for narcolepsy, with other products and programs in development. Reported in company filings; product detail is general knowledge and not checked against the current 10-K.
How it makes money: Prescription sales to patients through specialty pharmacies, with reimbursement from insurers. Net product revenue was $868M in FY2025 (reported).
Why customers choose it: Physicians prescribe WAKIX for narcolepsy patients who need a treatment that is not a stimulant, and insurance coverage and patient support matter for uptake. General knowledge.
Competitive advantages: Patent protection on WAKIX, an established sales force, and a prescriber base. The advantage is temporary: it lasts only as long as patents or exclusivity hold. Assessment.
Growth drivers: WAKIX growth in narcolepsy, expanded indications and new products in development. Q2 2026 WAKIX net product revenue was about $261M, up 30% year over year (reported in Q2 2026 preannouncement coverage).
Major threats: Loss of exclusivity (generic or competing product entry), competing treatments in development, reimbursement changes, and dependence on one or two products. Assessment. The exact patent expiry dates and settlement terms were not verified in this pass; they are the central unknown.
AI opportunities and disruption: Little direct AI exposure. AI could speed drug discovery at the margin, but it does not change the main risk here, which is patent-protected revenue. Assessment.
What must be true:
- WAKIX keeps growing at double digits through at least 2028, with 2026 net product revenue at or above the $1.0–1.04B guidance.
- Patent protection or competing-product timing lets the franchise run for the next 5–10 years.
- Pipeline or new indications offset any erosion after exclusivity ends.
- Management stability holds after the CFO's July 2026 departure.
Bull thesis: Revenue grew 21% in 2025 and is guided up 15–20% in 2026. Operating margin is about 24% on a GAAP basis and about 34% in the feed. The company generates about $300M a year of SBC-adjusted free cash flow and holds about $0.5B of net cash. The stock trades at a low multiple of cash flow.
Bear thesis: A single product drives most of the value. If generic or competing entry comes earlier than the market expects, revenue could fall sharply and the stock's cash-flow multiple would not protect the price. Growth is already slowing from 43% (2022) to 21% (2025).
2. Business quality
| Factor | Assessment | Basis |
|---|---|---|
| Revenue model | Product sales, recurring through prescriptions | Reported (product revenue) |
| Recurring share | High in practice: chronic-condition treatment | Assessment |
| Customer concentration | Specialty pharmacies and wholesalers concentrate distribution; exact percentages not pulled | Not verified |
| Geographic exposure | US | General knowledge |
| Pricing power | Moderate to strong for a branded product with no generic yet | Assessment |
| Market position | Leading branded product in its narrow category | Assessment |
| Moat | Narrow, with a defined end date tied to patent life | Assessment |
| Switching costs | Low to moderate for patients; prescriber habit matters | Assessment |
| Network effects | None | Assessment |
| Brand | Moderate within prescribers | Assessment |
| Cost advantages | Not material | Assessment |
| Regulatory advantage | Patents and FDA exclusivity; time-limited | General knowledge |
| Industry structure | Concentrated niche, with competing sleep-disorder therapies in development | General knowledge |
| Long-term runway | Depends on patent life and pipeline; not verified | Assessment |
Moat rating: Narrow. Protection comes from patents and exclusivity that expire on a known schedule. Its length is the most important unknown in the memo.
Business quality rating: Strong for the present. Operating margin is about 24% on GAAP, revenue is growing fast, and cash conversion is high. The rating falls if exclusivity ends soon.
3. Financial quality (SEC 10-K, fiscal years ended Dec. 31)
| Metric | FY21 | FY22 | FY23 | FY24 | FY25 |
|---|---|---|---|---|---|
| Revenue ($M) | 305 | 438 | 582 | 715 | 868 |
| Revenue growth | – | 43% | 33% | 23% | 21% |
| GAAP net income ($M) | 35 | 181 | 129 | 145 | 159 |
| GAAP diluted EPS | 0.58 | 2.97 | 2.13 | 2.51 | 2.71 |
| Operating income ($M) | 88 | 120 | 192 | 191 | 208 |
| Operating cash flow ($M) | 99 | 144 | 219 | 220 | 348 |
| Capex ($M) | 0.3 | 0.2 | 0.3 | 1.2 | 0.3 |
| Stock-based comp ($M) | 16 | 26 | 32 | 43 | 45 |
| Share repurchases ($M) | – | – | 100 | – | – |
| Cash ($M) | 234 | 244 | 312 | 453 | 753 |
| Long-term debt ($M) | 190 | 190 | 179 | 163 | 144 |
| Stockholders' equity ($M) | 187 | 403 | 467 | 659 | 870 |
| Diluted shares (M, approx.) | 57 | 59 | 60 | 57 | 57 |
Source: SEC XBRL company facts. Reported unless marked derived. Share counts are cover-page values and approximate.
Derived:
- Revenue CAGR FY21–25: about 30% (derived).
- EPS CAGR FY21–25: about 46%, but starting from a low base in FY21 (derived; FY21 EPS was depressed).
- Operating margin: 29% (FY21), 27% (FY22), 33% (FY23), 27% (FY24), 24% (FY25) (derived).
- SBC-adjusted FCF (OCF minus capex minus SBC): $303M in FY25, or 35% of revenue (derived).
- FCF conversion: OCF of $348M against net income of $159M, about 220% (derived).
- ROE FY25: about 21% on average equity (derived; feed shows 20%).
- ROIC and interest coverage were not computed in this pass. Interest expense was not captured.
Cash and debt: Feed shows cash of $668M and total debt of $160M (including leases), so net cash of about $0.5B. SEC data shows cash of $753M and long-term debt of $144M at FY25, implying net cash of about $0.6B. The model uses $0.5B (feed, conservative).
Share count and dilution: Shares have been roughly flat at 57–60M since FY21. There was a single $100M buyback in FY23. SBC of $45M a year is offset by no meaningful repurchases, so dilution is modest. Reported.
Dividends: none.
Buybacks versus SBC: Not applicable in any meaningful way. Harmony has not used buybacks to reduce its share count, so there is no evidence either way on whether buybacks created value.
Acquisitions: not captured in this pass.
Improving trends: Cash grew from $234M to $753M over four years, and operating cash flow rose from $99M to $348M. The company is now self-funding and has no need to raise capital.
Deteriorating trends: Revenue growth has slowed each year, from 43% in FY22 to 21% in FY25. Operating margin fell from 33% in FY23 to 24% in FY25 as spending rose.
Accounting concerns: FY22 net income of $181M was well above operating income of $120M; the reason was not identified in this pass. The 2026 non-GAAP EPS forecast in the feed (implied around $7) is far above GAAP EPS of $2.71 in FY25 and the $1.28 reported for Q2 2026. Check the non-GAAP definition before relying on any forward multiple.
Red flags: Loss of the CFO in July 2026, and dependence on one product. Neither is a financial red flag on its own.
4. Fundamental scorecard
| Metric | Current | 5-year trend | Peer comparison (JAZZ / AXSM) | Assessment |
|---|---|---|---|---|
| P/E, GAAP trailing | 14.0x (on FY25 EPS of $2.71) | 73x (FY21) → 14x (FY25) | JAZZ 14.4x (feed) | Low for its growth |
| Forward P/E | 5.4x (feed, implies ~$7 EPS, unreliable) | — | JAZZ 8.1x / AXSM 38.9x | Not usable until EPS basis is confirmed |
| EV / EBITDA | 6.7x (feed) | — | JAZZ 8.2x / AXSM negative | Low |
| Price / SBC-adjusted FCF | 7.3x | — | — | Low |
| FCF yield (SBC-adjusted) | About 14% | — | — | High, if the product holds |
| PEG | Not meaningful: FY21 base distorts the 5-year EPS CAGR | — | — | Not used |
| Debt metrics | Net cash ~$0.5B; D/E 0.16x (feed) | Deleveraged | JAZZ D/E 0.92x; AXSM 2.7x | Strong |
| ROE | About 21% | 18–21% recent | JAZZ 22%; AXSM negative | Good |
| Operating margin | 24% (GAAP derived); 34% (feed) | 27–33% | JAZZ 27%; AXSM −23% | Good |
| EPS growth | Volatile; 2025 +8% | FY21 base is depressed | — | Slowing |
| Revenue growth | 21% (2025); 15–20% guided (2026) | 43% → 21% | JAZZ 16%; AXSM 46% | Above JAZZ, below AXSM |
| FCF growth | OCF +58% in FY25 | Rising | — | Strong |
| Share-count growth | About flat | — | — | Neutral |
| Dividend payout | None | — | — | n/a |
| Insider ownership | 1.2% (feed) | — | JAZZ 2.9%; AXSM 15.1% | Low |
Peer set is limited to two names. Intra-industry comparisons are weak because each sleep and neurology company has different products and patent positions.
5. Competitive position
Major competitors: Jazz Pharmaceuticals (sleep franchise, including a competing narcolepsy product) and Axsome (AXSM, neuropsychiatric pipeline). Other sleep-disorder therapies are in development. General knowledge; not checked against current filings.
Where Harmony is stronger: Faster revenue growth than Jazz (21% versus 16% in the feed), lower leverage, and a simpler story with one major franchise.
Where Harmony is weaker: Smaller scale, concentration in one or two products, and less diversification than Jazz.
Valuation: Cheaper than Jazz on EV/EBITDA and on trailing P/E on the feed, despite faster growth. The discount reflects patent and concentration risk.
Expected return: See section 7. Base-case returns are high if exclusivity holds; the bear case depends on patent timing that I have not verified.
6. Valuation
Model inputs:
- Revenue FY26: $1.02B (midpoint of the $1.0–1.04B net product revenue guidance; company guidance, reported)
- SBC-adjusted FCF margin: 35% in FY25 (derived); base 32%
- Net cash: $0.51B (feed)
- Shares: 58.2M (feed)
- No dilution beyond SBC already deducted
| Assumption | Bear | Base | Bull |
|---|---|---|---|
| Revenue growth, years 1–10 | 0% in years 1–2, then 2% | 15%, then 10% fading to 3% | 18%, then 14% fading to 3.5% |
| SBC-adjusted FCF margin | 25% | 32% | 36% |
| Discount rate | 11.0% | 10.0% | 9.0% |
| Terminal growth | 2.0% | 3.0% | 3.5% |
| Exit multiple (P / SBC-adjusted FCF) | 6x | 9x | 13x |
| Scenario | Fair value / share | Implied market cap | Key assumptions |
|---|---|---|---|
| Bear | $56.53 | $3.3B | No growth for two years, then 2%; margin 25% |
| Base | $119.23 | $6.9B | Growth 15% fading to 3%; margin 32%; no loss of exclusivity |
| Bull | $196.22 | $11.4B | Growth 18% fading to 3.5%; margin 36% |
Current price vs. fair value: $37.95 is about 68% below the base value.
Margin-of-safety prices (base case): 25%: $89.42. 35%: $77.50.
Critical caveat. The bear case does not model loss of exclusivity. It models zero growth followed by 2% growth, which is a slowdown, not a patent cliff. A real loss-of-exclusivity case could cut revenue by a large share within a few years. Because I have not verified the patent expiry and settlement dates, I cannot price that downside. The base value assumes the franchise holds for 10 years, which is the single biggest assumption in this memo.
Analyst mean target: $48.25 (12 analysts, estimate). The analyst target implies much less growth or a shorter franchise life than my base case, which is a useful signal to reconcile before relying on any value here.
Sensitivity (base case): at 9% the value is $138; at 10% it is $119. Every version is above the price, but all depend on exclusivity.
Peer valuation: Jazz trades at 8.2x EV/EBITDA and 8.1x forward P/E in the feed. Harmony's 6.7x EV/EBITDA is lower, consistent with its concentration.
Historical valuation. Year-end price divided by GAAP EPS:
| Year-end | Price | GAAP EPS | P/E |
|---|---|---|---|
| FY21 | $42.60 | 0.58 | 73x |
| FY22 | $55.10 | 2.97 | 19x |
| FY23 | $32.30 | 2.13 | 15x |
| FY24 | $34.40 | 2.51 | 14x |
| FY25 | $37.40 | 2.71 | 14x |
| Now | $37.95 | 2.71 (trailing) | 14x |
The 3-year average (FY23–25) is about 14x, so the stock trades at its recent average. The FY21 figure is distorted by low earnings, so the 5-year average is not meaningful. Historical comparisons are only partly relevant: growth is slowing, the product mix is more concentrated, and the patent situation is more defined. The base case assumes no re-rating, using a 9x exit, below the current 14x trailing multiple.
7. Expected return
Annualized total return from $37.95. No dividend.
| Scenario | 3 years | 5 years | 10 years |
|---|---|---|---|
| Bear | −10.9% (price $27) | −6.0% (price $28) | −2.1% (price $31) |
| Base | +22.0% (price $69) | +15.6% (price $78) | +9.5% (price $94) |
| Bull | +48.2% (price $124) | +31.3% (price $148) | +17.6% (price $192) |
Decomposition (base, 5 years, approximate): revenue growth about 12% a year (fading from 15% to 3%); margin change near zero (32% assumed, compared with 35% SBC-adjusted in FY25); share count flat; dividend none; valuation change from 7.3x to 9x, adding about 4% a year. Derived.
Comparison with a broad index: the base case beats an index assumption of 7–8% over five years and 10 years, but the bear case is negative, and the bear case is not a true downside case because it excludes patent loss. Index assumption not verified.
8. Risk analysis
| Risk | Probability | Severity | Evidence that the thesis is breaking |
|---|---|---|---|
| Loss of exclusivity or early generic entry | Medium (timing not verified) | High | Patent litigation outcomes; generic approval dates; revenue falling below guidance |
| Competing therapies | Medium | High | Competitor approvals in narcolepsy; prescription share loss |
| Single-product concentration | High | High | WAKIX share of revenue staying above 90% |
| Reimbursement changes | Medium | Medium | Payer restrictions; higher patient out-of-pocket costs |
| Execution (pipeline and launches) | Medium | Medium | Trial failures or delayed filings |
| Management | Medium | Medium | Further senior departures after the CFO |
| Accounting quality | Low | Medium | Non-GAAP and GAAP gap widening; one-time items |
| Dilution | Low | Low | Equity issuance (none so far) |
| Valuation compression | Medium | High | Further de-rating if growth slows toward 10% |
9. Buffett and Lynch lenses
Buffett-style: partial pass with a limit. Understandability is moderate: the business is easy to describe, but its future depends on patent law and clinical data. The moat is narrow and time-limited, which Buffett would treat as a problem for a long-term holding. Returns on equity are good, cash generation is strong, and the balance sheet is net cash. Capital allocation has been conservative. The price is below the base value, but the value depends on an assumption about patent life that I cannot yet support. Does it meet Buffett's principles? Partly. It meets the business-quality and balance-sheet tests. It fails the durable-moat test until patent timing is known.
Lynch-style: Fast Grower, with binary risk. Revenue is growing 20% a year, which is fast for Lynch's categories. The story is simple: one drug, growing fast. Lynch would still ask what happens when the growth ends. PEG is not meaningful here: the 5-year EPS CAGR is inflated by the depressed FY21 base, and the growth rate is falling.
10. Capital allocation test
Why own Harmony instead of an index? Because the cash flow and the price make a strong return possible if WAKIX keeps its exclusivity for several years, which would justify the 22% three-year base case. If it does not, the stock is likely to fall well below today's price.
The evidence does not yet support a compelling case that Harmony beats an index with high confidence. The upside is large, but the downside depends on patent timing that is not verified. That makes the valuation uncertainty too high to classify with confidence.
Classification: Not classified. The valuation uncertainty is too high to say the stock is attractive at any price until exclusivity dates and litigation outcomes are verified. Once they are, re-run the model with a loss-of-exclusivity scenario, and reassess.
Comparison with Jazz: Harmony is cheaper, grows faster, and carries less debt, but has far more concentration. Jazz is the more diversified business; Harmony is the simpler bet.
11. Investment decision dashboard
| Field | Value |
|---|---|
| Current price | $37.95 (Oct. 16, 2026 close) |
| Market cap | $2.2B |
| Enterprise value | $1.7B (feed; net cash about $0.5B) |
| Business quality | Strong for the present |
| Moat | Narrow, time-limited by patents |
| Financial strength | Strong: net cash, high cash conversion |
| Growth outlook | Strong near term (15–20% guided for 2026); uncertain beyond exclusivity |
| Valuation | Low on cash flow; not usable on forward P/E until EPS basis is confirmed |
| Risk level | High (patent timing and single-product concentration) |
| Bear-case fair value | $57 (does not model loss of exclusivity) |
| Base-case fair value | $119 (assumes no loss of exclusivity over 10 years) |
| Bull-case fair value | $196 |
| 25% MoS price | $89 (base) |
| 35% MoS price | $78 (base) |
| Expected 5-year base-case CAGR | 15.6% |
| Expected 10-year base-case CAGR | 9.5% |
| Most important bull argument | Fast, cash-generative growth from one product with net cash on the balance sheet |
| Most important bear argument | Patent expiry or early generic entry could remove most of the revenue |
| Biggest unknown | WAKIX exclusivity dates and litigation outcomes |
| Thesis breaker | Verified loss of exclusivity before 2030, or revenue below guidance two quarters in a row |
Classification: Not classified. Valuation uncertainty is too high until patent timing is verified.
Final conclusion
Is this a great business? It is a strong business today: fast growth, a 24% GAAP operating margin, growing cash, and no debt that matters. It is not a durable great business until its patent protection and pipeline are known.
Is the valuation attractive? On cash flow, the stock looks cheap, at about 7x SBC-adjusted free cash flow and about 14x trailing earnings. That is only attractive if the franchise lasts. Without verified exclusivity dates, I cannot say the price is attractive.
What return appears realistically achievable? If exclusivity holds for five years, the base case is about 15% a year. If it does not, returns could be negative, and my model cannot yet quantify how negative.
What could permanently impair the investment? Early generic or competing-drug entry that removes most of WAKIX revenue faster than the pipeline can replace it.
Does it justify allocating capital instead of owning the market? Not yet. The upside is real, but the key uncertainty is one that could permanently reduce the value, and the evidence does not let me size it. Revisit once patent dates and litigation outcomes are verified. This is research, not personalized advice.
Sources
- SEC XBRL company facts for Harmony (CIK via screener cache), FY2021–FY2025 10-K values
- Yahoo Finance market and peer data, Oct. 16–17, 2026 (
memo_peers.json,screen_results_sp400_sp600.csv) - Q2 2026 preannouncement and guidance reaffirmation: Harmony preannouncement, via BioSpace
- Q2 2026 8-K exhibit: SEC EDGAR, Aug. 4, 2026
- Q2 2026 earnings summary: Quartr
- Model outputs:
memo_valuation.json
Source caveats: Patent and exclusivity dates were not verified in this pass, and that is the central unknown. The Q2 2026 figures come from the company's preannouncement and a secondary summary. The forward P/E in the feed implies an EPS basis that does not match reported figures, so it is not used. Q3 2026 results were not available at the data date.