DNTH — Q2 2026 Financial Report Analysis

Dianthus Therapeutics lost $50.2 million in the second quarter of 2026, up from $31.6 million a year earlier, because spending on research and development nearly doubled as its lead drug, claseprubart, moved into three late-stage trials at once. That loss matters less than the balance sheet: after a $719 million share sale in March, the company held $1.19 billion in cash and investments at June 30 and says that is enough to fund operations into 2030. Dianthus has no product on the market; its only revenue is a small licensing stream from a partner in Greater China.

The numbers

Operating margin is not meaningful here: with under $1 million of revenue against $62 million of expenses, the ratio is a large negative number that tells a reader nothing.

Where the money went

The $22.4 million rise in R&D splits into two parts, per the filing's MD&A:

• $15.1 million more in outside costs. Of that, $13.0 million was claseprubart — $7.8 million more for running clinical trials and $6.0 million more for CMC (chemistry, manufacturing and controls, i.e. making the drug supply for trials), partly offset by $0.7 million lower preclinical study costs. Another $2.1 million went to discovery work on the two earlier drugs, DNTH212 and DNTH312.
• $7.3 million more in internal costs: $3.3 million in personnel, $3.2 million in stock-based compensation and $0.8 million other, from building out the in-house team for its Phase 2 and Phase 3 trials.

General and administrative costs rose $4.7 million, of which $3.1 million was stock-based compensation and $1.3 million personnel, from higher headcount.

Revenue of $0.8 million comes from Tenacia, which licensed claseprubart rights in Greater China and reimburses part of the trial costs; it rose because those reimbursable costs went up. For the six months it actually fell, to $1.2 million from $1.4 million. It is a cost-sharing line, not a business.

What the headline numbers hide

• Interest income is cushioning the loss. Interest and investment income of $11.5 million offset almost a fifth of the $61.5 million operating loss. The operating loss widened 76%; the net loss only 59%. That cushion depends on both the cash pile and interest rates, and it shrinks as the cash is spent.
• Cash burn is lower than the reported loss — for now. For the first half, cash used in operations was $77.8 million against a net loss of $91.0 million. The gap is mostly $22.5 million of stock-based compensation (a real cost to shareholders, paid in shares rather than cash), partly offset by $6.9 million of interest that accrues on discounted bonds but is not yet received in cash. At the first-half pace, the $1.19 billion would last well beyond 2030; the company's more cautious "into 2030" guidance implies it expects spending to keep climbing as three late-stage trials enroll, which the MD&A says outright ("research and development expenses will increase substantially").
• The runway was bought with dilution. In March 2026 Dianthus sold 8.47 million shares plus 0.40 million pre-funded warrants (shares bought in advance for a $0.001 exercise price) at $81.00 each, raising about $719 million gross ($676.1 million net), and sold another 1.12 million shares through its at-the-market program for $58.7 million net. Shares outstanding rose from 43.2 million at year-end to 54.8 million at June 30, about 27%, before counting remaining pre-funded warrants. The price matters: the September 2025 offering was at $33.00, so the March raise brought in far more cash per share given up. The at-the-market program now has no remaining capacity; a $600 million shelf registration filed in January 2026 remains available.
• Stock-based compensation doubled. It was $22.5 million in the first half vs $11.0 million a year earlier, and accounts for $6.3 million of the quarter's combined $27.1 million rise in R&D and G&A.
• No one-offs in the quarter. Apart from a $0.1 million gain on a small holding of Zenas stock, there were no unusual items in either year. Dianthus pays no income tax and records no tax benefit on its losses.
• The DNTH212 licence will cost more. Dianthus has paid Nanjing Leads Biolabs $30.0 million so far and owes an $8.0 million milestone when a Dianthus-led Phase 1 study begins, plus up to $962 million more in development, regulatory and sales milestones if the drug succeeds.

Takeaway: Q2's widening loss is the expected cost of running three late-stage trials at once, and the March raise at $81 a share means Dianthus can see all of them through without needing more money soon; what decides the company's value over the next 15 months is the data — MMN results in December 2026, CIDP Part B guidance by year-end and gMG Phase 3 results in the second half of 2028.

Pipeline and what to watch

The filing lists these milestones:

• CIDP (chronic inflammatory demyelinating polyneuropathy, a nerve disease): in the Phase 3 CAPTIVATE trial, 75% of the first 40 patients responded in the open-label Part A, against a 50% target, with no related serious infections or drug-related serious side effects. Because the target was reached early, the trial was resized to enroll up to 256 patients (from up to 480) and randomize 128 into Part B (from 192). Management expects to give Part B top-line guidance by the end of 2026 and believes this single trial can support an approval filing.
• MMN (multifocal motor neuropathy): the Phase 2 MoMeNtum trial finished enrolling in August 2026 with 46 patients, above its target of 36. Top-line results are expected in December 2026 — the next hard data point.
• gMG (generalized myasthenia gravis, a muscle-weakness disease): after positive Phase 2 data in September 2025 and an end-of-Phase 2 meeting with the FDA, the roughly 195-patient Phase 3 EMERGE trial started in June 2026, testing dosing every two or four weeks. Results are expected in the second half of 2028. Claseprubart received Orphan Drug Designation for myasthenia gravis on May 1, 2026.
• FDA protocol changes (March 2026): the agency agreed to drop routine anti-nuclear antibody screening and testing, a common reason patients failed screening across all three programs, which should help enrollment.
• DNTH212: Phase 1 healthy-volunteer data in China expected by year-end 2026. DNTH312 aims to be ready for Phase 1 by year-end 2027.

Our read: expect quarterly losses to keep rising through 2027 as EMERGE and CAPTIVATE enroll and manufacturing scales up; the cash pile makes that a non-issue for funding. The near-term risk is concentrated in December: the MMN readout is a smaller Phase 2 trial, and a weak result there would raise questions about how far claseprubart extends beyond CIDP and gMG. The CIDP Part A response rate is encouraging, but Part B — whether patients who stop the drug relapse more than those who stay on it — is the test regulators will weigh.

This is the first Dianthus report on this site, so there is no earlier outlook to check against.