ARTV — Q2 2026 Financial Report Analysis

Overview

Artiva Biotherapeutics reported a second-quarter 2026 net loss of $25.0 million, up from $21.3 million a year earlier. The extra $3.8 million came almost entirely from research spending on AlloNK, its lead drug, as clinical trials in autoimmune disease grew. The quarter matters much more for what happened off the income statement. In May the company released its first efficacy data in refractory rheumatoid arthritis (RA). It said it had agreed with the FDA on a single Phase 3 trial that could support approval. It then raised $282.7 million net in a public offering. Cash, cash equivalents and investments ended June at $349.4 million, up from $108.0 million at the end of 2025, and management says that funds the plan "into 2029".

Some background: AlloNK is an allogeneic NK cell therapy, an off-the-shelf treatment made from natural killer immune cells taken from donor cord blood rather than from the patient. It is frozen in advance and shipped to the clinic. Artiva pairs it with rituximab, an approved antibody that targets B cells. The aim is to wipe out the B cells that drive autoimmune disease and let the immune system "reset", much as personalised CAR-T cell therapies have shown in early studies. Artiva has no product revenue, and in this quarter it recognised no collaboration revenue either.

At a glance

• $349.4M in cash and investments, runway "into 2029": the May offering ($300.0M gross) more than tripled the cash pile. Management expects primary Phase 3 data in the second half of 2028 and a possible approval filing in 2029, so the money is meant to last to roughly the data readout, not through an approval.
• R&D up 22.8% to $21.9M: external spending on AlloNK rose 36.6% to $11.9M, which the 10-Q attributes to "ongoing clinical trials exploring AlloNK in autoimmune diseases". This is where the money is going, and it should rise further once the Phase 3 trial starts.
• Loss per share fell to -$0.63 from -$0.87 even though the loss grew: the weighted share count rose 61.7% to 39.4 million after the offering. The smaller per-share loss reflects dilution, not better economics.

Results table

For the first half, the net loss was $48.5M versus $41.6M, and R&D was $41.2M versus $34.9M. With no revenue, operating margin is not a meaningful figure, so it is left out.

Where the money went

The 10-Q explains the $4.1M rise in R&D as "a $3.2 million increase in external research and development expense, and a $0.9 million increase in internal research and development expense". The external increase it attributes to "an increase in AlloNK costs related to product candidate development and ongoing clinical trials exploring AlloNK in autoimmune diseases". The internal increase was mostly "a $1.0 million increase in personnel-related expenses due to increased headcount". Almost all of Artiva's spending now goes to AlloNK: other programs got $18 thousand of external spend this quarter. Last year the company completed its lymphoma trial early so it could focus on autoimmune disease.

General and administrative costs were flat at $5.0M. Interest income rose to $1.9M from $1.6M, which the filing puts down to "a higher money market and investment balance in the second quarter of 2026 from the May 2026 Offering". The full quarter of interest on the new cash will show up in Q3.

The clinical news behind the raise

The financing followed directly on the May 8 data update (data cutoff April 3, 2026):

• Efficacy in refractory RA: 21 patients had at least 12 weeks of follow-up, and 13 of them had six months. All 21 had high disease activity, and 81% had failed two or more classes of advanced RA drugs (biologic or targeted synthetic DMARDs). More than 50% of patients followed for six months reached ACR50, the standard RA response measure meaning at least a 50% improvement in symptoms and signs. 19 of 21 showed clinically meaningful drops in two disease-activity scores, CDAI and DAS28-ESR. For comparison, Artiva cites published data suggesting this group has only an 11%–19% chance of reaching ACR50 on existing drugs.
• Safety: there was no cytokine release syndrome or neurotoxicity, the two serious immune reactions that have held back CAR-T therapies. No serious adverse events were related to AlloNK and no patient stopped treatment because of side effects. The grade 3 or higher infection rate was 2% (one patient). Every patient was treated as an outpatient, mostly in community rheumatology clinics.
• Mechanism: all 51 evaluable patients showed B-cell depletion in the blood by Day 13. Where B cells returned, they were mostly "naïve/transitional" cells, which fits the "immune reset" idea.
• Regulatory: Artiva has agreed with the FDA on a single registrational trial. About 150 patients will be randomised 2:1 to AlloNK plus rituximab or rituximab alone, with ACR50 at six months as the primary goal. In June the FDA granted RMAT designation, an expedited-review status for regenerative medicines. AlloNK already had Fast Track status in refractory RA.

Two caveats. These are early, open-label results from a small number of patients, combining a company trial and an investigator-run trial, with no control group. The Phase 3 trial adds rituximab alone as a control arm, which Artiva expects to reach about 20%–25% ACR50. Also, the regimen includes cyclophosphamide/fludarabine "conditioning" chemotherapy before dosing. The filing says the most common side effects were those linked to rituximab or that conditioning, so conditioning is part of the risk profile a regulator and a rheumatologist will weigh.

Takeaway: In Q2, Artiva turned early RA data into an FDA-agreed path to a single Phase 3 trial and $282.7M of new cash. The company is now funded "into 2029", past its expected second-half 2028 primary readout. The price was roughly doubling the share count. The investment case now rests almost entirely on whether AlloNK plus rituximab can beat rituximab alone on ACR50 in a controlled trial.

What the headline numbers hide

• Cash burn is a little lower than the reported loss: first-half operating cash outflow was $41.3M against a $48.5M net loss. The gap comes from $3.5M of non-cash stock-based compensation, $1.3M of depreciation and a $2.7M working-capital tailwind, mainly a $1.9M fall in prepaid expenses. At roughly $20.7M a quarter, the H1 pace would stretch $349.4M well beyond 2029. Management's more cautious "into 2029" guidance reflects its expectation that spending will "increase substantially" once a 150-patient trial across more than 80 sites is under way.
• The smaller per-share loss is a dilution effect: the net loss grew 17.6%, but loss per share improved because the weighted share count jumped 61.7%. The full effect is still to come. Shares outstanding went from 24.6M at year-end to 48.8M at June 30, plus 2.17M pre-funded warrants that already count in basic EPS, so about 51.0M share-equivalents in total, a roughly 107% increase. Q3's per-share loss will reflect a full quarter of the larger count. A further 4.8M options and RSUs were excluded from diluted EPS as anti-dilutive.
• No going-concern doubt: the notes say cash is sufficient "for at least one year from the issuance of these financial statements". The company still says it "will need to secure additional funding in the future" to complete development and commercialise, so another raise before any approval is likely. An $11.95M at-the-market program with Leerink is suspended but still in place.
• No collaboration revenue to watch: Artiva recognised no revenue and received no payments from GC Cell, its Korean licensor, manufacturing partner and shareholder, in either period. Artiva's rights cover the world outside Asia, Australia and New Zealand, and it owes milestones and royalties to GC Cell on future sales. There is no big-pharma partner funding the program, so all trial costs fall on Artiva's own balance sheet.
• No one-offs: apart from the offering, the quarter has no unusual items. Expenses are R&D and G&A only, and there is no debt.

Outlook

Management's stated timeline is to start the Phase 3 trial in refractory RA in the second half of 2026, report primary efficacy data in the second half of 2028, and potentially file for approval (a BLA) in 2029. It plans a safety database of more than 250 patients treated with AlloNK plus rituximab across its trials. Artiva estimates 150,000–200,000 US patients have failed two or more advanced RA drug classes.

Our read: the financial picture is now simple. Losses will rise as the Phase 3 trial enrolls, and the cash is sized to get through the readout. The next few quarters' 10-Qs should show external AlloNK costs rising well above the current $11.9M a quarter. Things to watch: confirmation that the first Phase 3 patient is dosed in the second half of 2026; longer, larger updates on how long responses last in the existing RA groups, since only 13 patients had six months of data at cutoff; any safety signal tied to the conditioning chemotherapy as more patients are treated in community clinics; and whether the quarterly burn rate fits management's "into 2029" runway. The single-trial design means one readout will largely decide the outcome.