Collegium Pharmaceutical: $936 million of net debt against a $740 million market value

Collegium's shares are worth about $740 million. Its net debt, after a $650 million cash purchase of AZSTARYS in May, is about $936 million (my estimate for October). A screen shows the stock at 3.1 times adjusted earnings. That is the multiple of a company whose best-known painkillers are about to meet generics and whose newest asset was bought with borrowed money.

At the October 16, 2026 regular session close of $23.19 the shares trade 14.8% below my triangulated Base value of $27.22, upside of 17.4%. The Bear value of $1.81 is 92.2% below the price and the Bull value of $51.88 is 123.7% above it. Confidence is grade C (13/25), the required margin of safety (MOS) is 45%, and the maximum buy price is $14.97. The latest company report is the second quarter release of August 6, 2026. Third quarter results and the final settlement of the August share repurchase are not in these numbers.

Valuation route: commercial specialty drug company with finite product lives. Primary method: hard EBITDA (adjusted EBITDA after stock pay, the JORNAY financing royalty and recurring legal costs) at a 5 times exit multiple in December 2028, with two forward years of cash going to debt, discounted at 10%. Secondary method: a franchise by franchise cash runoff to 2035. Classification: standard company, 4 of 12 on the healthcare scorecard. Values are present fair value estimates, not 12 month price targets.

What Collegium is

Collegium sells branded specialty drugs in two groups. The pain group is Belbuca, Xtampza ER and the Nucynta franchise. The ADHD group is JORNAY PM, which came with the 2024 purchase of Ironshore, and AZSTARYS, bought from Corium Therapeutics on May 12, 2026 for about $650 million in cash, with up to $135 million more if milestones are met. In the second quarter pain contributed $140.9 million, or 70.5% of $199.9 million of sales, and ADHD $59.0 million, or 29.5%. Belbuca alone was $57.7 million, 28.9% of the company. Three wholesalers accounted for 97% of fiscal 2025 revenue, according to the annual report, which is a channel concentration and not a statement about how many patients buy the drugs.

The company spends nothing on research: R&D expense was zero in fiscal 2025, because it has been a buyer of commercial assets since 2022. That makes cash generation easy and replacement hard. Physical capital spending is under 1% of sales, so the real reinvestment is the next drug purchase, and my model assumes none.

The dates when generics can arrive matter more here than the last patent date. As the company's own January 2026 presentation sets them out:

• Belbuca: generic entry is permitted from January 2027 under a settlement with Teva. Other patent litigation can run to 2032. In Base I keep $150 million of residual brand and authorized generic sales in 2027 and $120 million in 2028, against $230 million this year (my allocation).
• Nucynta: authorized generics launched in the first quarter of 2026. Lower net prices on them are the reason the company gave for cutting guidance in August.
• Xtampza ER: Teva may enter in September 2033.
• JORNAY PM: Orange Book patents run to March 2032 and the presentation lists no generic filers yet.
• AZSTARYS: six Orange Book patents, most to December 2037. This is the longest runway in the portfolio.
• Symproic: patents to November 2031. A small asset.

A permitted entry date is not a launch date, since a generic still needs approval and supply. That uncertainty is why my three scenarios differ so much.

The latest quarter

JORNAY prescriptions grew 13.1% and prescribers 17.6%, so most of the 41% revenue growth is real demand, though net price and mix always need checking. AZSTARYS prescriptions grew only 1.9%. The GAAP loss comes from $63.0 million of amortization and $24.1 million of acquisition costs in the quarter, which is why management and I look at EBITDA and cash instead. Over the half year, product revenue was $393.4 million and adjusted EBITDA $217.7 million.

The midpoints came down by $40 million of revenue (4.5%) and $30 million of EBITDA (6.2%), and the company blamed the Nucynta authorized generics. The year is two thirds reported, so guidance is not a wide target, but the cut tells you that the pain group can still surprise on the downside.

Which cash to value

Adjusted EBITDA adds back a lot. For the first half of 2026 the add-backs were $118.4 million of amortization, $25.4 million of stock-based compensation (SBC), $30.3 million of acquisition costs, $5.4 million of inventory step-up and $1.4 million of executive transition costs. Some of those are not cash, and the stock pay is a real cost to shareholders. I therefore value what I call hard cash: cash from operations, less capital spending, less SBC.

The three-year average of hard cash for 2023 to 2025 is $234.6 million and the five-year average is $176.2 million. I accept neither as permanent: 2024 carried payments of liabilities assumed with Ironshore, 2025 had a full year of that business, and the spread between them is the warning against treating the best year as the base. In the first half of 2026, hard cash was 47.0% of adjusted EBITDA.

Hard EBITDA is 82.8% of the headline figure and hard owner cash 50.9% of it. This is a run rate built from guidance, not the company's own free cash flow guidance, which it has not given. The 23% cash tax rate is a proxy that gives no value to deferred tax assets or amortization shields, and I test 15% to 28%. The JORNAY royalty sits in the EBITDA line because the company pays it out of operating cash through March 2032, so I do not also deduct the $121.4 million royalty liability as debt.

Debt, shares and other claims

In December 2025 Collegium closed a $980 million credit facility: a $580 million term loan, a $300 million delayed draw term loan used for AZSTARYS and a $100 million revolver, priced at SOFR plus 2.75% to 3.75%. At June 30 the term debt was $865.5 million in principal and the convertible notes $241.5 million, $1,107.0 million together, against cash of $129.5 million. Both the cash and the share count have moved since then, so I roll them forward to October 19 and label the result as an estimate.

The $655.4 million net outflow for AZSTARYS and the $299.5 million of loan proceeds are already in the June balances, so I neither subtract the price nor add the loan again. Net debt of $936 million is about 2.5 times my hard EBITDA run rate and 2.0 times guided adjusted EBITDA (my arithmetic).

Two other claims matter. Purchase consideration payable is $29.7 million: $17.6 million to former Ironshore holders, matched by restricted escrow, which I offset, and $12.2 million for AZSTARYS, which I deduct. Cash held in escrow beyond that gets no credit.

The August repurchase was $50 million paid to Jefferies, with 1,556,420 shares delivered up front, about 80% of the expected total at the $25.70 reference price of August 12. The rest is due by the fourth quarter, and at $25.70 the remaining $10 million would buy about 0.39 million shares. At the 2026 diluted count that goes into adjusted EPS, 39.9 million shares, the notes are included as if converted; the GAAP loss quarter used only 32.5 million. Neither is right for a spot valuation, so I keep the notes as debt today and convert them only at the 2028 exit, and only if the equity is then worth more than $36.56 a share, the conversion price. Conversion removes $241.5 million of debt and adds the shares in the same step. The 2.5 million unvested awards and the 0.5 million options are tested with a share count up to 34.0 million.

Share count since 2021 has fallen about 2.2% a year on a debt-retained basis (my arithmetic, over four years), yet SBC rose from $24.3 million to $41.9 million. Buybacks do not make that cost go away. In the first half of 2026, SBC was 19.9% of reported free cash flow.

Sales and cost build behind the Base

The 2026 total and the ADHD midpoints come from guidance. The split of the pain group is my allocation. JORNAY grows 17.7% a year from $195 million to $270 million in Base. AZSTARYS rises from $70 million to $145 million, but 2026 holds it for only about seven months, so that is not a growth rate. Its 2025 revenue under previous ownership was about $113 million in the acquisition pro forma, and my 2027 figure of $115 million is a flat year.

Base 2028: $790 million x 85% - $290 million - $26.2 million - $10 million = $345.3 million. Platform costs rise while total sales are flat because the ADHD drugs need promotion while the pain sales fall. The company's target of over $50 million of AZSTARYS synergies is already inside those costs and is not added again.

Primary valuation: hard EBITDA with a two year transition

Normal earnings are more than two years away and differ from today's, so multiplying them by a multiple and calling the result today's value would be a trick. The model runs forward to December 31, 2028, 2.2 years from the valuation date. Cash generated before then pays down debt or builds liquidity. Nothing is paid to shareholders, since there is no dividend, and cash is counted once: as lower debt at exit. Owner cash each year is (hard EBITDA - interest) x (1 - 23%) - capital spending - working capital.

The stub is the 73 days to year end. Interest falls from $66 million to $60 million and $50 million as debt is repaid. The $90 million added to liquidity is built out of cash above the scheduled repayments of $11 million in the stub and $66 million in each of 2027 and 2028, so the model does not pretend a reserve can be funded by skipping them. By my formula 2027 owner cash is $222.0 million, $2.0 million more than the table. I kept the lower figure, and using the higher one would add about $0.05 a share.

Equity before conversion is $38.22 a share, above the $36.56 conversion price, so the notes convert in Base: $30.76 today. That makes the Base sensitive to a conversion threshold, and below roughly a 4.9 times multiple the notes would stay as debt.

Why 5 times? It is my estimate for a portfolio with generic risk and a finite life, not a measured peer median. Year-end enterprise value to EBITDA from a data aggregator ran from 4.4 to 8.0 times in 2021 to 2025, with a median of 5.0, but those use a softer EBITDA than mine, so I give them no weight. Bear uses 4 times for a shorter remaining life and Bull 6 times for a better one. The milestone payments are $20 million, $45 million and the full $135 million.

Sensitivity: primary value per share by Belbuca sales and exit multiple

Each row changes the 2028 Belbuca sales and the debt paid down with them. These are isolated sensitivities, not three extra scenarios.

Secondary valuation: franchise cash runoff

The second method drops the multiple and values each drug's cash to 2035. For every product it takes sales times a contribution margin after manufacturing, non-financing royalties and direct selling costs (JORNAY 65%, AZSTARYS 60%, Belbuca and Xtampza 72%, Nucynta 55%, Symproic 45% in Base). It subtracts central platform and stock pay costs of $99 million rising to $103 million and then falling to $75 million, $10 million a year of legal costs and the JORNAY royalty, applies 23% tax, and deducts $3 million of capital spending and $5 million of working capital each year.

Present value at December 2028 of the 2029 to 2035 cash is $1,212.5 million. A residual of 2 times the 2035 cash adds $157.0 million, 11.5% of the exit enterprise value. It stands for what AZSTARYS and the generic tails may still earn after 2035 and is not a growth formula. Enterprise value at exit is $1,369.5 million, $357 million below the 5 times multiple, or 20.7% lower. Less $507.3 million of net debt and other claims leaves $862.2 million of equity, $27.03 a share, below the conversion price, so the notes stay as debt here. Discounted, that is $21.91 today.

• Shared by both methods: the drug portfolio, 2027 and 2028 sales, the patent and generic dates, the starting cash and debt and the conversion rule.
• Different in each: the first capitalizes consolidated 2028 EBITDA at a multiple I choose, the second builds product margins, central costs and a full runoff. The second has no perpetuity.
• Treated the same: all capital spending is deducted, rent is inside costs, and neither credits any minority or joint venture cash.
• What agreement would prove: little. More than half of the value in each method depends on the same drug franchises, so the methods cannot confirm that the sales forecasts are right. The two do not agree, and that lowers my confidence.

History, peers and growth diagnostics

Normalized hard cash has to be compared with where the market already is. At $23.19 the diluted equity is $739.8 million and enterprise value $1,675.8 million. That is 3.66 times the guided adjusted EBITDA and 4.43 times my hard EBITDA. The only close peer I could price on the same date is ANI Pharmaceuticals (ANIP) at $72.28, about 6.2 times its guided 2026 EBITDA, with revenue growing 22% to 29% (the ANIP figures are from its own release and a balance sheet at June 30, so this is a partial check). A growing peer deserves a higher multiple than a company with eroding legacy sales, and I do not claim that 6.2 times applies to a harder metric. I left out Supernus because its announced merger distorts its price. I have no full peer median, and I did not make one up.

The growth-adjusted check does not work either. Third-party consensus adjusted EPS is roughly $7.4 to $7.5 for 2026 and a few percent lower for 2027, so the 2026 adjusted P/E is about 3.1 times and the growth rate is negative. A PEG ratio on negative growth is not meaningful, and a vendor PEG of 0.66 uses a growth window I cannot reconcile, so I discard it. Consensus does not enter my Base.

For free cash flow growth I take the 2023 to 2025 average of hard cash, $234.6 million, as the normalized starting point and compare it with my 2028 cash three years later.

The 38.6% historical growth comes from $77.4 million in 2021 to $285.7 million in 2025, a span in which two acquisitions changed what the company owns, so I do not read it as organic. The forward number is a decline despite AZSTARYS, because pain losses outweigh ADHD gains, and that difference is the thesis in one number. Under full conversion the 2028 cash per share would be about $5.84, not $6.88, because conversion adds shares as well as removing interest. The reliability of these estimates is low. I found no dispersion of cash flow estimates and no revision history.

These are cash diagnostics and do not replace the scenario values. Growth quality is acquisition-affected and uncertain.

Valuation triangulation

The two Base values are 33.6% apart, measured against their midpoint, above my 30% line for a serious divergence. I lowered Confidence one grade for it, and added no separate margin of safety. The first method gets 60% because the continuing platform and an observable exit multiple matter, the second 40% because finite drug lives are a binding limit. Neither gets 100%. Equal weights would give $26.34. The weights are judgments about methods, not probabilities of the three scenarios, and I do not publish a probability-weighted price.

What is the market already pricing in?

These are separate answers, not one forecast. The price does not need Collegium to fail. A Belbuca franchise 36% smaller than my Base in 2028, or JORNAY sales 16.5% lower, or a lower multiple is enough to explain $23.19, and ADHD value need not go to zero. Buying here and receiving the Base outcome would earn about 18.3% a year to the end of 2028, a modeled return for information and not a forecast. The same outcome bought at $24.68 would earn 15% and at $22.47 20%.

Bear, base and bull scenarios

Bear: generics permanently shrink the pain cash, ADHD growth is not enough, and the fixed commercial cost cannot fall as fast as sales. The 4 times multiple is for a shorter remaining life, not a cyclical trough. Common equity is worth only $1.81 even without a legal payment, because the debt takes most of the enterprise value. Base: JORNAY and AZSTARYS partly replace the pain losses, sales hold at $790 million, and rising promotion and royalties stop revenue from turning into easy cash. Bull: better generic resilience and stronger ADHD uptake give $995 million of sales and $503.6 million of EBITDA, with the whole $135 million milestone paid and the notes converted.

The range is wide enough to matter: the Bear is 92% below the price and the Bull 124% above it. No financing or legal event is stacked into the ordinary Bear, and the floor for a share price is zero.

How the buy price is set

A Confidence score of 13 is grade C, whose base margin of safety runs from 35% to 50%. I use 35%, the low end, because the operating erosion is already inside the scenarios, and add 10 points for a legal tail I cannot model: an arbitration claim of more than $500 million (see the risks). The required MOS is 45% and the maximum buy price is $27.22 times 0.55, or $14.97. That is 35.4% below the price. The leading one-input tests move it between $11.84 and $17.75. The price is above that whole band, and a price below it would call for a check for fresh impairments before any purchase.

Changing the Belbuca sales alone moves the value from below the price to well above it, so this is a verdict that depends on one assumption, and the maximum buy price should be read as a range with a conditional answer. A joint optimistic case can change the action, and I show it in the stress test below.

I keep four different numbers apart. Fair value today is $27.22. The modeled return at $23.19 is 18.3% a year. An entry price for a 15% return, using the same Base exit, is $24.68, and for 20% it is $22.47; those rates are illustrations, not a hurdle I recommend. The margin of safety against model error gives $14.97. Where each risk is priced:

• Belbuca and Nucynta erosion: in the product scenarios and the post-2028 runoff. No extra concentration charge.
• ADHD execution and payer access: in the sales ranges and platform costs. The discount rate stays at 10%.
• Finite patents: in the explicit runoff, the 4 to 6 times multiple range and the 0 to 3 times residual test.
• Leverage and the 2028 to 2029 financing: in the debt deducted, the repayments, conversion and the liquidity reserve. No leverage add-on.
• Method divergence and the October estimates: in the Confidence grade and the sensitivities.
• The North Sound claim: in separate stress tests of $100 million and $500 million, and the 10 point margin add-on. I assign no probability.

Dividend and buybacks

Collegium has never paid a cash dividend and says it has no intention to, so this is not an income stock and the payout ratio is not applicable. Cash goes to debt and to buying back shares. In the 2024 to 2025 program it bought 2.70 million shares for $85.0 million, an average of $31.43, per the filings, 35% above today's price. The new $50 million repurchase paid for shares at about $25.70. Buybacks help the shares that stay, and they are not income for holders who sell.

Reasons to own Collegium

• The cash is reported, not adjusted. Cash from operations was $329.3 million in 2025 and $128.4 million in the first half of 2026. The question is how much of it survives the generics.
• JORNAY has real demand. Prescriptions rose 13.1% in the second quarter, and its patents run to March 2032.
• Deleveraging is built into the Base. Cash can take net debt from $936 million to $450 million by the end of 2028 without a dividend competing for it.
• AZSTARYS has the longest life. Most of its patents run to December 2037, which offsets the near-term pain erosion, though the $650 million price and up to $135 million of milestones need real returns.
• The Bull case has room. It pays the whole milestone cap and the convert dilution and still reaches $51.88.

These justify watching the stock. At $23.19 they do not remove the risk that the residual equity gets small if the pain cash falls quickly.

What could go wrong?

• Thesis breaker: pain cash falls faster than ADHD cash grows. Belbuca can face generics from January 2027 and Nucynta authorized generics are already selling at lower net prices, while promotion and manufacturing cannot fall in step. My ordinary Bear leaves $1.81 a share after debt.
• Legal exposure. A claim by North Sound Partners over alleged pre-acquisition conduct seeks damages above $500 million, according to the 10-Q as I have it, and the company says it cannot estimate a loss. A $100 million payment at exit would take the Base to $24.82 and $500 million to $14.65, which is below today's price. These are tests, not estimates, and an earlier payment would be worse. Patents, payer access, controlled-substance supply rules and opioid litigation are separate risks.
• Refinancing. The credit facility matures in December 2030, but I understand it can spring forward to November 18, 2028 if more than $50 million of the convertible notes remains and liquidity is under $350 million. The notes themselves mature on February 15, 2029. My Base holds $250 million of cash and assumes the $100 million revolver is available, subject to the actual definitions. A rate 2 points higher on about $700 million of debt for two years takes the Base to $26.73, and that does not test closed credit markets.
• Capital allocation. Acquired growth has to earn a return on $650 million. The first half of 2026 added back $30.3 million of acquisition costs and $25.4 million of SBC, and the annual report says adjusted EBITDA affects incentive pay, so repeated exclusions deserve scrutiny.
• Receivable and rebate accruals. Accrued rebates, returns and discounts were $406.9 million at June 30, up from $318.3 million at year end. I assume no release of that working capital.

The 2027 interest and repayment schedule is my reading of the credit terms. Scheduled term repayments of $11 million in the stub and $66 million in each of 2027 and 2028 are covered even in the Bear, where cash is about $117 million and $97 million.

Management execution

The record is good in selected years and weaker lately. Fiscal 2023 revenue of $566.8 million landed within the late-year guidance of $565 to $570 million. In 2024 revenue of $631.4 million beat the initial $580 to $595 million, but about $37.2 million of that was acquired JORNAY revenue, so legacy sales of roughly $594.2 million were at the top of the range, not above it. In 2025 revenue of $780.6 million and EBITDA of $460.5 million beat initial ranges of $735 to $750 million and $435 to $450 million with the same perimeter. Then in 2026 the May guidance after AZSTARYS ($865 to $895 million) was cut in August to $825 to $855 million. The refinancing in December lowered borrowing costs and pushed out maturities, and the company then borrowed another $300 million while buying back stock, which makes visible debt reduction more important. The Ironshore earn-out hurdle was not met, a reminder that revenue growth is not the same as every deal milestone. Assessment: 11/20, good delivery in selected years with a real 2026 warning; capital allocation average.

Stock Analyza scorecard

A score of 4 sits at the top of the standard class, and I say so because 5 would start the quality growth range. Four things keep Collegium out of it: growth bought with debt, no proven return on the acquisitions, negative character change in the legacy cash flows and no demonstrated replacement pipeline. The change from a pain portfolio to an ADHD portfolio is the character change, and it lowers the verdict by a grade from a mechanical Buy on the discount.

Business quality is 48/100. The platform turns established drug sales into cash with almost no physical capital, which is its best feature. Low physical capital is not low reinvestment, though, because mature drugs have to be replaced. I give no premium for return on capital, since an acquisition-adjusted return is not available. A low research ratio is not evidence of innovation here, because R&D is zero.

The valuation score is 48/100. A discount of 17.4% to Base earns 24 of 40 points, and protection against the Bear case earns none because the Bear value is 92% below the price.

Confidence score

The key operating inputs are my estimates, which rules out an A, and the gap between the methods takes the provisional B down to C. Agreement between the methods would not have restored it, because they share the same drug forecasts.

What I would watch from here

Today's readings: guidance sits at the green line, JORNAY prescription growth of 13.1% is green, and AZSTARYS prescription growth of 1.9% is a warning even for a partial period. Liquidity is my estimate and cannot be graded until the company reports it. The next review point is the third quarter report and the final settlement of the repurchase. A screening calendar suggests November 5, but I found no date confirmed by the company. Later checkpoints are Belbuca's permitted entry in January 2027 and an April 2027 trial setting against Alvogen, as I have them from the company's filings. I would re-underwrite after any guidance change, material acquisition, large issuance or buyback, financing event, major legal or regulatory decision, or leadership change.

The three most important thesis breakers

1. Pain cash erodes faster than ADHD owner cash replaces it, pushing hard cash toward the ordinary Bear without a replacement.
2. Financing is not available before the 2028 and 2029 obligations.
3. A material adverse arbitration, patent or regulatory event changes the cash or the asset lives.

Stress test and independent check

I rebuilt the model in a separate script from the raw inputs: the run rate bridge, hard EBITDA for each case, the cash allocation, the conversion rule, the runoff, both values and the triangulation, the three scenarios and their returns, the reverse solutions, the 3 by 3 grid and the one-input sensitivities. Every value reproduced to the cent, apart from the $2.0 million in 2027 owner cash that I noted. I checked the price, the guidance, second quarter results, cash, the share repurchase, the AZSTARYS purchase terms, the credit facility size and the generic entry dates against the company's releases and filings. The price is the October 16 close, the later after-hours quote excluded.

The sales test is my own: I set 2027 sales to about $845 million, near third-party consensus (the version of that consensus I could read stopped at 2026, so treat the level as a guide), held 2028 flat and scaled the later runoff by the same ratio. The generous joint case needs $180 million of 2028 Belbuca, $320 million of JORNAY, a 6 times multiple and a 15% cash tax rate all at once. At those inputs the current price would meet my margin of safety. They are the top of my sensitivity ranges together, and none of them is shown by results yet.

The strongest case against my Hold is that the cash is real, the generic entry may slip for approval or supply reasons, and the Bull value is $51.88. I take it seriously, and the stronger sales case is worth $32.92. Its entry price is still $18.10, below the quote. The strongest case against my Base is that 5 times may overcapitalize a portfolio whose main growth drug loses patents in 2032, and the runoff check gives only $21.91. I keep both views in the range and do not average them away.

Main data limits: October cash, the share count and the final settlement of the repurchase are my estimates from the June balance sheet and the announced transactions. Hard EBITDA, the platform costs, the 5 times multiple, the 23% tax proxy, the 2027 to 2035 product paths, the contribution margins, the milestone amounts and the interest schedule are my estimates. I could not read the 10-Q's legal and debt notes line by line, so the size of the North Sound claim, the springing maturity terms and the $865.5 million of term principal rest on the company's filings as I have them. Consensus for 2027 sits behind a paywall in the version I read. The ANIP and aggregator multiples are partial checks. The call transcript was not used.

Final verdict: hold at $23.19

Collegium makes a lot of cash and the price already allows for a lot of erosion, but the discount does not pay for the uncertainty about what remains after the generics and how long the remaining drugs last. The nominal Undervalued label is cut by one grade for the negative character change in the legacy cash, the wide gap between the methods and a price that fails the margin of safety. For an existing holder this is not an automatic sell signal, but the cash and refinancing thesis should be checked before adding, and a low forward P/E alone is no reason to average down.

Verdict: HOLD at $23.19, watchlist for new money. Triangulated Base value $27.22, upside 17.4%, Confidence C (13/25), required margin of safety 45%, maximum buy price $14.97, entry range about $11.84 to $17.75.

Net debt is $936 million today and, in my Base, $450 million at the end of 2028. That is the whole thesis: the cash has to arrive before the generics do.

Sources

Primary sources include Collegium's second quarter 2026 results release of August 6, 2026, the 10-Q for the period ended June 30, 2026, the fiscal 2025 10-K and fiscal 2023 10-K, the August 13 announcement of the share repurchase, the May 12 release on the AZSTARYS acquisition, the December 30, 2025 credit facility announcement, the January 8, 2026 corporate presentation for generic entry and patent dates, the initial 2025 guidance, the initial 2024 guidance and the AZSTARYS acquisition financial information. For the peer check I used ANI Pharmaceuticals' second quarter release and Supernus's second quarter release. Price history, consensus EPS, year-end multiples and the calendar estimate come from a third-party market data aggregator (COLL quote, forecast, ratios, statistics, ANIP quote). Broker price targets and ratings are not used.

The reference market price is $23.19, the regular session close on October 16, 2026. Values per share use 31.9 million economic diluted shares, with 38.5 million if the convertible notes convert at exit. Net debt of $936 million is $1,096 million of principal less $160 million of cash, both estimates for October 19. The earnings path, exit multiples, scenario values and robustness tests are analytical estimates, not company guidance.