CGTX — Q2 2026 Financial Report Analysis

Overview

Cognition Therapeutics, a clinical-stage drug developer working on a pill for Alzheimer's disease and dementia with Lewy bodies (DLB), cut its second-quarter net loss to $3.9 million from $6.7 million a year earlier. The improvement came almost entirely from spending less on clinical trials: research and development (R&D) expense fell to $5.1 million from $11.5 million because its large Phase 2 START study finished enrolling patients in December 2025 and the costly recruitment phase is over. The company has no product sales. Its main source of non-equity funding, government grant reimbursement, fell in step with that spending, to $3.5 million from $7.1 million. It ended June with $34.8 million of cash (including $0.75 million restricted) and says that lasts into the fourth quarter of 2027.

At a glance

• R&D down 56% to $5.1 million. Spending on outside clinical research firms fell $5.3 million as START moved from enrolling patients to simply following them.
• Net loss down to $3.9 million, or $0.04 a share. Lower trial costs did the work; grant income, which reimburses those same costs, roughly halved alongside them.
• $34.8 million of cash, runway into Q4 2027. That covers the expected START readout, but not a Phase 3 program, which management says will need new money.

The quarter in numbers

Source: Form 10-Q for the quarter ended June 30, 2026. "Grant income" here is money from the National Institute on Aging (NIA), part of the US National Institutes of Health, which pays back a share of eligible trial costs as they are incurred. The company reports it below the operating line, so its "loss from operations" looks larger than its true net cost.

Why spending fell

The 10-Q breaks R&D into its parts. In the quarter:

• Clinical programs: $2.9 million vs $8.2 million. The company attributes the $5.3 million drop to "decreased Phase 2 trial activities with contract research organizations" (the outside firms that run trial sites). START, a 545-patient study in people with mild cognitive impairment or early Alzheimer's, finished enrolling in December 2025; what remains is 18 months of treatment and follow-up per patient.
• Personnel: $1.9 million vs $2.5 million, from "reduced professional fees and headcount."
• Manufacturing: $0.2 million vs $0.5 million, from less replenishment of clinical-trial drug supply.

General and administrative cost (overheads such as finance, legal and insurance) was roughly flat at $2.6 million, with higher professional fees offsetting lower stock-based and employee pay.

For the six months, R&D was $11.2 million against $22.3 million and the net loss $8.5 million against $15.2 million. The one line that rose was preclinical work (lab and animal studies), up $0.5 million to $0.7 million, which the company says supports "registrational programs", meaning the late-stage trials that could lead to approval.

What the headline numbers hide

• The grant income drop is not a funding cut, but it is a mirror. NIA grants reimburse eligible costs as they are spent, so when trial spending halves, grant income does too. The net cost to shareholders of running the business (operating loss minus grant income) was about $4.2 million in the quarter versus $6.9 million a year earlier. The real saving is that $2.7 million gap, not the $6.4 million drop in R&D alone. As of June 30 the company had about $21.6 million of NIA money still committed for future expenses, with grant project periods running through May 31, 2027.
• Per-share loss improved faster than the loss itself because of dilution. The net loss fell 42%, but the average share count rose 42% to 90.6 million after an August 2025 share sale (14.7 million shares at $2.05, about $27.9 million net) and at-the-market sales this year (5.5 million shares for $6.9 million gross in the first half). Had the share count stayed at last year's 63.7 million, the quarter's loss would have been about -$0.06 a share; so roughly 30% of the move from -$0.11 to -$0.04 comes from spreading the loss over more shares rather than from a smaller loss.
• Cash burn was helped and hurt by timing, roughly evenly. Six-month operating cash outflow was $8.6 million, close to the $8.5 million net loss. Underneath, the company collected $7.0 million of overdue grant receivables but also paid down $8.6 million of accounts payable and accrued expenses (accrued expenses fell from $12.0 million at year-end to $3.8 million, consistent with settling the bills from START's enrollment push). Neither tailwind nor headwind is likely to repeat at that size.
• The cash balance only edged down because of share sales. Cash, equivalents and restricted cash fell $2.2 million in the half: $8.6 million used in operations, offset by $6.4 million of net financing, almost all from the at-the-market program (selling shares into the market gradually). $68.1 million of capacity remains under that program.

Pipeline: what the money is buying

The company has one drug, zervimesine (CT1812), a once-daily pill designed to stop toxic protein clumps (amyloid-beta and alpha-synuclein "oligomers") from binding to the connections between nerve cells.

• Alzheimer's, Phase 2 START (545 patients): enrollment complete; topline results come after every participant finishes 18 months of treatment. With enrollment ending in December 2025, that points to mid-2027 at the earliest; the filing gives no specific date. START is funded by an $81 million NIA grant.
• Alzheimer's, Phase 3 plan: after an end-of-Phase-2 meeting in July 2025, the FDA agreed with a design randomizing patients to 100 mg or placebo for at least six months, enriched for patients with lower blood levels of p-tau217 (a marker of brain amyloid and tau). That subgroup is where the earlier SHINE study showed its strongest signal. European regulators have said they would prefer a longer trial.
• DLB psychosis, pivotal plan: in May 2026 the company and FDA "aligned on key aspects" of a pivotal study in DLB patients with hallucinations and delusions, nine months of 100 mg or placebo. The primary endpoint's statistical details are still to be settled with the FDA, and more meetings are planned.

Takeaway: Cognition's lower loss is the natural lull after a big trial stops enrolling, not a change in the business. $34.8 million of cash covers operations into late 2027 and should see the company through START's readout, but neither the Alzheimer's Phase 3 nor the DLB pivotal study is funded. Whatever START shows will decide the price of the next raise.

Outlook

Management expects expenses and losses to "increase substantially" as programs move into later-stage trials, and says it will need "substantial additional funding." Its runway estimate (into Q4 2027) assumes no further at-the-market sales, so the guide is not flattered by planned dilution. That said, the company sold shares under the program in the first half anyway, and the share count was 95.1 million at June 30 against a 90.6 million quarterly average, so further dilution is likely before any trial readout.

Our read: the next few quarters should look like this one, with R&D in the $5-6 million range and partly offset by grant income, until the company commits to starting either pivotal trial. The things to watch are (1) a confirmed START readout date, (2) a finalized FDA endpoint for the DLB psychosis study, and (3) whether the company finds a partner or new non-dilutive (non-share) funding for Phase 3 before it starts, rather than leaning on the at-the-market program. With the remaining NIA project periods ending May 2027, grant income should keep shrinking unless new awards arrive.