Roivant (ROIV) Q2 2026 earnings review
Roivant's first medicine sells $31 million, but its loss widens
Roivant started selling its first medicine and its trials met their goals, yet the reported loss widened. LISRAYA, a treatment for the muscle and skin disease dermatomyositis, brought in $31.0 million in its first weeks on the American market. Chief executive Matt Gline said the launch is "off to a strong start". Payouts to holders of Genevant stock awards then took $95.9 million out of Roivant's share of the result, which pushed the loss per share to $0.27.
| Product sales in the first five weeks | $31.0M no product sales a year ago |
|---|---|
| Loss per share, excluding the Genevant payouts | -$0.14 -$0.17 a year ago |
| Cash and marketable securities | $4.21B $3.84B at the end of June |
| Cash runway outlook | Unchanged about twelve quarters at this quarter's rate of spending |
⚖️ Verdict: 🟢 Bullish
The case got better, because Roivant now sells a medicine and has a second use for it in view. Brepocitinib cleared both of its late-stage eye trials, and Priovant plans to ask the FDA to add that disease to the label in the first half of 2027. One caution: the arthritis result came at a single dose, and Immunovant quietly dropped Sjögren's disease and halted its skin lupus study without publishing the numbers.
The question now is whether five weeks of LISRAYA sales say anything about the year ahead. One answer is that the specialists who treat this disease switched quickly and will keep prescribing; the other is that pharmacies simply filled their shelves at launch. Roivant publishes no sales forecast, so the January report, with three full months of prescriptions, settles it.
🐂 Bull Case
LISRAYA Sold $31 Million in Five Weeks
LISRAYA is Roivant's brand name for brepocitinib, a medicine the FDA approved in August for dermatomyositis, a disease that inflames muscles and skin. Priovant, the Roivant subsidiary that owns it, does the selling. No targeted treatment had ever reached this disease, so there was no earlier launch to copy.
Pharmacies and patients bought $31.0 million of it in the first five weeks. That works out near $6 million a week by our math. Chief executive Matt Gline said the medicine is "off to a strong start", and pointed to "reflecting strong early patient and prescriber demand" as the reason.
The start matters because a year ago the group sold no medicine at all, and because Gline had told investors to expect a slow, steady launch. Five weeks prove little on their own: specialty pharmacies stock up at the beginning, and this release counts dollars rather than patients.
What to watch: the quarter ending in December brings a full three months of LISRAYA sales, the first clean read on demand. Should that figure fall short of three times the five-week number, early pharmacy stocking rather than patient demand carried the launch.
Both Uveitis Trials Met Every Main Endpoint
Non-infectious uveitis is inflammation inside the eye that can take a patient's sight, and doctors hold it down with steroids that carry harms of their own. Priovant ran two identical late-stage trials, CLARITY-1 and CLARITY-2, giving brepocitinib 45 mg once a day against placebo.
Each trial cleared its primary endpoint and every ranked secondary endpoint, and Priovant said the results set "a new bar for time to failure and steroid-sparing benefit in NIU". Safety matched earlier brepocitinib trials, with no new signals, which matters for a drug class doctors watch closely.
- Time to treatment failure: more than 48 weeks on the drug, about 12 weeks on placebo
- Steroid-free quiet disease at week 48: 62% of treated patients against 20% on placebo
- Swelling at the centre of the retina cleared in 42% of treated eyes, against 4.8% on placebo
- Treatment without steroids: 40 weeks of the stretch
"Priovant intends to file a supplemental New Drug Application (sNDA) for brepocitinib for the treatment of NIU in the first half of calendar year 2027", so sales from this second use arrive later. The size of the benefit decides whether eye doctors reach for a JAK inhibitor, a class Gline expects to carry the standard warnings on its label.
What to watch: the filing itself in the first half of 2027. A slip past that window would suggest the regulator wants more from these two trials than the headline figures show.
Research Spending Fell While Launch Costs Rose
Roivant pays for trials and for the people who run them, and nothing else comes close. Research spending came to $145.7 million, $18.8 million below a year ago, as Immunovant wound down batoclimab, stopped its lupus study and carried a "lower headcount" on the payroll.
- Research and development: $145.7 million, down $18.8 million from a year ago
- Selling and administration: $154.1 million, up $11.0 million on the LISRAYA launch
- Total operating costs: $303.2 million, against $367.8 million in the June quarter by our math
- Mosliciguat: $11.6 million more, "reflecting the progression of the program"
That mix matters because the group funds everything from its own balance sheet. Cutting one franchise while paying for a launch held total costs below last year's, so Priovant's sales force did not widen the hole. The anti-FcRn franchise alone took $18.3 million less, partly offset by the Graves' disease trials.
What to watch: research spending over the next two quarters, now that the Phase 3 mosliciguat trial is enrolling. A climb back above $160 million would show the new trial costs more than the discontinued ones saved.
Mosliciguat Cut Lung Pressure by 56%
Pulmovant is testing mosliciguat, a drug patients breathe in for high blood pressure in the lungs caused by scarring lung disease. The Phase 2 PHocus trial met its main goal: the resistance the heart must push blood against fell 56.3% more than on placebo.
Roivant called that "the highest ever reported PVR reduction in any randomized controlled PH trial". Patients also walked 35.2 metres further than the placebo group at week 16, and "a lower proportion of patients experienced cough", the side effect that limits inhaled rivals. Pulmovant has started the Phase 3 PHrontier trial.
What to watch: how fast PHrontier enrolls, and whether the walking benefit survives in a trial built to measure it. Management conceded last year that the Phase 2 was not powered for that endpoint, and it is the one regulators weigh.
🐻 Bear Case
Immunovant Drops Two Diseases From Its Plan
Roivant has sold investors a wide pipeline. IMVT-1402, Immunovant's antibody that strips disease-causing antibodies out of the blood, was meant to run across ten indications. This quarter the list got shorter in two places, and the release carries no data for either one.
Immunovant now says it is "unlikely that we will advance IMVT-1402 further in SjD", the dryness disease Sjögren's, blaming "the evolving SjD landscape and recent data in the field". The skin lupus trial, "which was discontinued in September 2026", produced no published result, although Roivant had pointed to topline data in this half of the year.
Two deletions cut against the pipeline-in-a-product story. Gline said last year that the lupus readout would decide whether that programme continued, so halting it answers the question; keeping the numbers back does not. Sjögren's sat on the calendar as a 2028 readout only three months ago.
What to watch: whether Immunovant ever publishes the lupus result, and what the FDA meeting later this year does to the arthritis programme. Each deleted disease leaves fewer places for one antibody to earn back its development bill.
The Arthritis Win Covers One Dose Only
Immunovant's arthritis trial gave IMVT-1402 to 170 patients for 16 weeks, then took the drug away from those who had responded, to see who held the benefit. Only 117 of them entered that second stage, and the higher dose alone beat placebo there.
- Response kept at week 28: 92.5% on 600 mg against 75.6% on placebo, p=0.0139
- The 300 mg arm reached 75.0% and "did not show a difference from placebo"
- Two secondary measures only "numerically favored IMVT-1402 600 mg over placebo", at p=0.1096 and p=0.1299
- No serious side effects and no discontinuations in any arm
So the result rests on one dose and one measure. Gline warned a year ago that strong open-label results would make this endpoint hard to clear, and the gap of 17 percentage points by our math is thinner than the first-stage response rates implied. A registrational path still needs the regulator: "FDA discussions are planned" for later this year.
What to watch: what comes out of that meeting. A demand for a second controlled trial would push any arthritis application far behind the Graves' disease and myasthenia gravis readouts due next year.
Share Count Rises Despite the Buybacks
Roivant repurchased 2.1 million shares for about $72.5 million during the quarter. A year earlier it spent $208.7 million. The average price it paid, near $34.50 a share by our math, sits above last year's level, so the same money retires fewer shares than it used to.
The count still went up. Weighted average shares reached 722.8 million from 680.9 million a year ago, because stock awards create new shares faster than the programme cancels them. That is one reason the loss per share worsened while the consolidated loss from continuing operations narrowed.
What to watch: the share count in the January report against the pace of repurchases. Another increase would show the buyback absorbing employee pay rather than returning capital to shareholders.
👓 Other Themes
Genevant Payouts Widened Roivant's Loss
The $0.27 loss per share does not measure how the business ran. Genevant's own staff hold stock awards that shared in the Moderna money, and those "distributions on outstanding restricted stock units and options of Genevant Sciences Ltd." took $95.9 million, or $0.13 a share. Minority holders in subsidiaries also took $54.4 million of income, against a $52.5 million loss a year ago.
💲 Other KPIs
Stock awards made up 46.3% of selling and administration costs, against 49.5% a year ago. The share has sat between 44% and 50% for five quarters while the dollar amount hardly moved. These awards create most of the gap between reported costs and the adjusted figures Roivant highlights.
Earnings on the group's cash and securities came to $41.6 million, 8% under a year ago, because the portfolio is smaller than it was. Even so, this income covered 13.7% of operating costs by our math, which is why the adjusted loss narrowed.
Book value fell 13% from $5.29 billion at the fiscal year end in March, as losses and the Genevant payouts outran new share issuance. Liabilities came down too, to $380.1 million from $416.3 million, so nothing on the other side of the balance sheet caused the drop.
What Moderna owed Genevant at the fiscal year end became cash in July. Roivant collected $771.6 million, which lifted cash and marketable securities to $4.21 billion from $3.84 billion at the end of June, even after a quarter of spending and buybacks.
🔮 Guidance
Unchanged. Roivant repeats that its money lasts until it earns a profit, now from a bigger balance after Moderna paid. At the $334 million the group used this quarter, by our math the balance lasts about twelve more quarters. That sum excludes the second Moderna payment and the cost of the new Phase 3 trial.
Now dated. Priovant names a half-year for its supplemental application in non-infectious uveitis, where before it would only promise speed. Roivant will therefore earn nothing from this second use of brepocitinib during the current fiscal year, and the two trials behind it carry no further risk of failure.
Held. Immunovant keeps its own deadline and will sit with the regulator before the year ends, this time with the withdrawal data in hand. That meeting settles how many more trials the arthritis use needs. No date is attached to the answer, so it may arrive alongside the January results.
Unchanged. Immunovant says "clinical development timelines remain on track" for its two largest registrational readouts next year. These two diseases, rather than arthritis, carry most of what the antibody has left to prove. A slide into 2028 would leave the uveitis filing as the only near-term event on the calendar.
Withdrawn. Three months ago Sjögren's sat among the 2028 readouts; Immunovant now calls it "unlikely that we will advance IMVT-1402 further in SjD". The reason it offers is other companies' recent results rather than its own data. Investors lose an indication and get no numbers with which to judge the decision.
Unchanged. The trial in chronic inflammatory demyelinating polyneuropathy, a disease that damages nerves, still points to 2028. With Sjögren's gone, it is the only 2028 date the antibody programme still carries, so any delay here opens a gap after next year's two readouts.
Unchanged. Judges heard arguments on Moderna's appeal on 5 November, so an answer could come within months. Roivant books nothing for the payment, which arrives only on a resolution "favorable to Genevant and Arbutus". A loss leaves the $950 million already collected as the whole of the settlement.
❓ Key Questions
How many patients now take LISRAYA?
Roivant reported dollars, not patients or prescriptions. Without a count, nobody can separate specialty pharmacies filling their shelves from doctors writing repeat prescriptions in the first five weeks.
Why are there no skin lupus results?
Immunovant pointed to topline lupus data in this half of the year, then stopped the study in September. The numbers would show whether the antibody worked in that disease at all, or whether rivals set the bar too high.
Why did 53 arthritis patients not continue?
The trial enrolled 170 patients and analysed 117 in the withdrawal stage. The release does not say what became of the rest, and who leaves a trial shapes how convincing the 92.5% figure is.
Will buybacks keep up with new shares?
The share count has climbed in each of the last three reported quarters while repurchases slowed. Roivant has not said whether it intends to hold the count flat, which decides what the programme is for.
When will Roivant publish a sales forecast?
The group offers no revenue outlook even with a medicine on the market. An Investor Day falls on 9 December, and a first forecast there would let readers judge the launch against a stated plan.