CABA — Q2 2026 Financial Report Analysis

Cabaletta Bio, a Philadelphia biotech developing a one-time CAR T cell therapy (rese-cel) for autoimmune diseases such as myositis and lupus, has no product sales yet, so its quarterly results are really a report on spending and cash. In the second quarter of 2026 it lost $51.0 million, 13% more than a year earlier, as research spending rose with its first registrational trial (the trial meant to support an FDA approval filing). Yet the loss per share halved to $0.34, because the company had issued so many new shares that the average share count was 2.4 times larger than a year ago. Cash rose to $225.1 million after a May share sale, but management still says that only lasts "into mid-2027", which is the same time the key trial is due to report.

At a glance

• Net loss of $51.0M (+13% year on year): the dollar loss grew; the better-looking per-share figure is entirely a share-count effect.
• Cash of $225.1M at June 30 (vs $133.6M at December 31, 2025): raised mainly from a $140.2M net share sale in May at $2.90 a share, not from operations.
• Runway "into mid-2027": the registrational myositis data are expected in mid-2027 and the approval filing (a Biologics License Application, or BLA) in the second half of 2027, so the money as guided runs out around the moment it matters most. After the quarter, warrant holders paid in another ~$115.0M (see below).

The numbers

Cabaletta reports no revenue, so the table focuses on what drives its cash needs. "Operating cash outflow" is the cash actually spent running the business, which differs from the accounting loss.

Quarterly operating cash outflow is the six-month figure ($84.1M in 2026, $61.2M in 2025) minus the first quarter ($42.6M and $30.8M). Year to date, the net loss was $94.5M versus $81.1M (+16.6%) and the loss per share $0.72 versus $1.44.

Where the money went

R&D rose $6.8 million. The 10-Q breaks this down:

• +$3.0M of licence milestones: $2.5M to Nanjing IASO (which licensed Cabaletta the part of the CAR that recognises B cells) and $0.5M to Oxford Biomedica (the viral-vector supplier), both triggered when the first patient in a rese-cel pivotal trial was dosed. These are one-time payments tied to an event.
• +$2.4M in clinical-trial costs from "higher enrollment and clinical trial sites across multiple clinical studies."
• +$2.2M in manufacturing, "primarily due to expanded cell processing capabilities, including those for commercial readiness."
• +$0.8M in personnel from more R&D staff.
• -$2.0M in development services, from lower spending on lab consumables and outside research.

Without the $3.0M of milestones, R&D would have been about $41.4M, up roughly 10% rather than 18%. G&A fell $0.7M, mainly from $1.1M lower personnel costs due to a smaller headcount, partly offset by early commercial-readiness spending.

What the headline numbers hide

• The per-share improvement is dilution, not progress. The loss per share fell from $0.73 to $0.34 while the dollar loss rose from $45.1M to $51.0M. The cause is the share count: 100.5M shares at the end of 2025, 167.7M at June 30, 2026, after the 51.7M-share May offering, 10.1M shares sold through an at-the-market program (selling stock gradually on the open market, $28.7M net) and warrant exercises. Each existing share now owns a much smaller slice of the company.
• And it has diluted further since. The June 2025 offering came with 53.1M warrants (rights to buy shares at $2.50) that expired on September 12, 2026. A September 14, 2026 8-K says 46.0M were exercised between July 1 and that date, bringing in about $115.0M gross. Together with 1.8M more at-the-market shares sold after June 30 ($5.5M net), the share count is now roughly 215M by our arithmetic from these filings, before 6.0M pre-funded warrants (shares already paid for that can be claimed for almost nothing).
• The going-concern warning stays despite the cash. Management "concluded that substantial doubt exists about the Company's ability to continue as a going concern", in plain words a formal warning that it may not be able to pay its way for the next year without raising more money. This was already in the first-quarter filing and the cash pile has grown, but the company still says it "will need to raise additional capital to fully implement its business plan" and has "no credit facility or committed sources of capital."
• Cash burn is larger than operating cash flow shows. Year to date, $84.1M left through operations, plus $7.1M in finance-lease principal payments, which sit in the financing section of the cash flow statement but are essentially rent for dedicated manufacturing suites (at Minaris and Lonza). The total is about $91M in six months. Operating cash outflow was smaller than the $94.5M net loss mainly because of $10.8M of stock-based pay and $7.0M of non-cash lease expense, partly offset by $9.9M of working-capital outflow (the company paid down $4.7M of accrued liabilities and $1.2M of payables).
• Interest income is helping only slightly. It was $1.6M in the quarter, up from $1.4M because of higher cash balances. Year to date it fell slightly ($2.7M vs $2.9M) on lower government-bond yields.

Trial progress the spending is buying

• Myositis (lead program): the 17-patient registrational cohort in dermatomyositis and antisynthetase syndrome began enrolling in December 2025. Its primary endpoint is a moderate or major Total Improvement Score response at 16 weeks with patients off all other immune-suppressing drugs. In Phase 1/2 data presented at EULAR in June 2026, 8 of 10 such patients would have met that bar, and no myositis patient had neurological toxicity (ICANS). Topline data are guided for mid-2027 and a BLA for the second half of 2027. A juvenile cohort could support a Priority Review Voucher, a transferable FDA credit that companies have historically sold for cash.
• Systemic sclerosis with lung disease: a single-arm registrational study of about 25 patients is expected to start in the fourth quarter of 2026. In Phase 1/2, patients with lung involvement showed a median 7.5% improvement in predicted lung capacity (FVC) at 36 weeks.
• Lupus and myasthenia gravis: the company announced two registrational lupus cohorts (about 25 patients each) in January, but now says it will decide next steps after dose-ranging a version given without "preconditioning", the chemotherapy normally used to clear space for CAR T cells. In the first two lupus patients the lowest dose without preconditioning looked like a threshold dose: one patient had deep B-cell depletion and the other about 90%. The myasthenia gravis program is adding the same no-preconditioning dose exploration.
• Manufacturing: Cabaletta let its Minaris contract expire in August 2026 and now relies on Lonza, ElevateBio (added as a second manufacturer) and a 10-year commercial supply deal with Cellares for automated production.

Takeaway: On paper, Cabaletta's per-share loss halved, but that came from more than doubling its share count. The underlying burn is rising (about $91M in six months including manufacturing-lease payments) and guided runway ends right at the mid-2027 myositis readout. The ~$115M of warrant cash received in September is the most important number for the next filing, because it is not in the "into mid-2027" guidance.

Outlook

Management's guidance is unchanged: cash on hand "into mid-2027", myositis registrational data in mid-2027, a BLA in the second half of 2027, and the systemic sclerosis registrational study starting in Q4 2026. Its R&D commentary says spending should "increase substantially for the foreseeable future", and starting a second registrational study would add to that.

Our read: if spending holds at roughly $45M a quarter (operating outflow plus lease payments), the June 30 cash alone covers about five quarters, consistent with the guidance only if costs keep rising, as management says they will. The September warrant proceeds could plausibly push the runway past the myositis readout and BLA filing, but the company has not yet restated its runway in a filing, so treat that as unconfirmed until the third-quarter 10-Q (due around mid-November 2026). Points to watch in that filing: whether the runway guidance moves beyond mid-2027, whether the going-concern language is removed, whether registrational myositis enrollment is still on track for mid-2027 data, and whether R&D growth speeds up as the systemic sclerosis study starts. The outcome still depends on the trials: with no revenue, Cabaletta's value rests on the 17-patient myositis cohort.

This is our first published analysis of Cabaletta Bio, so there is no earlier outlook to check against.