ABOS — Acumen Pharmaceuticals

Company context · prepared 2026-08-07 · company sweep 2026-08-07 · USD · framework v5.5.0

How to read this. Written for a reader who has not studied pharmacology or finance. General terms are defined once in framework/04-glossary.md. Tags: [VERIFIED — source] / [UNVERIFIED] / [WEB ESTIMATE — source, date].

Program analyses for this company are the sub-folders listed in C.2.

C.0 Company-tier coverage — CLEARED

C.1 What the company is

Acumen Pharmaceuticals is a clinical-stage biotechnology company in Newton, Massachusetts, and it is a single-asset company in every practical sense [VERIFIED — BPIQ fetch_company_info 2026-08-07]. It has no product revenue of any kind. Every one of its four pipeline programs is in Alzheimer’s disease, and three of the four are variations on one molecule or one delivery technology.

The lead asset is sabirnetug, also called ACU193. It is a laboratory-made antibody — a protein engineered to stick to one specific target in the body — designed to bind soluble clumps of a protein called amyloid-beta while largely ignoring the hardened deposits of the same protein [VERIFIED — J Prev Alzheimers Dis 2025, DOI 10.1016/j.tjpad.2024.100005]. It is in a Phase 2 trial called ALTITUDE-AD whose results are due in Late 2026. That single readout is what the equity is [VERIFIED — BPIQ fetch_company_drugs 2026-08-07].

Two things about the financial position changed materially in 2026 and both are in C.3. The company now carries bank debt, and its own accounts state that there is substantial doubt about its ability to continue as a going concern without more money [VERIFIED — Acumen Q1 2026 Form 10-Q, filed 2026-05-12]. It is not funded through to a Phase 3 trial, and on its own guidance it is funded only into early 2027 — a few months past the readout it is waiting for. The next scheduled disclosure is the Q2 2026 report on 2026-08-12 [VERIFIED — press feed 2026-08-05].

C.2 Pipeline — every program

All four rows returned by fetch_company_drugs on 2026-08-07. The catalyst column reports catalyst_date_text exactly as disclosed; the “to a day?” column is yes only where that text names a day (01-rules.md, rule 23).

No row is marked Failed or Discontinued. That is a finding rather than a comfort: Acumen has never shut a programme down, so there is no record of how management behaves when something does not work, and there is also no second molecule that has ever been given to a patient.

All four rows carry the indication text “Alzheimer’s disease”, so the indication does not separate them. The integer bpiq_drug_id does.

C.3 Financial position

Dilution history. Every raise in the last 24 months, plus the resale registration.

Is a raise likely before the next catalyst? The honest answer is that it is likely either just before or immediately after, and the company would prefer after. It has already declared substantial doubt about going concern, it has an undrawn $20.0M debt tranche it has chosen not to draw, and its own stated runway ends within roughly one quarter of the readout it is waiting for. The March 2026 placement was explicitly earmarked for the preclinical brain-delivery portfolio, which tells you that the Phase 2 trial itself was already paid for and that the balance sheet has no room for anything beyond it. A holder should assume that a positive Late 2026 readout is followed within weeks by an equity raise at whatever the new price is, and that a negative one is followed by a financing done from a position of no strength. The 2026-08-12 Q2 report is the next place a shelf registration or at-the-market facility could surface.

C.4 Valuation frame

Two facts about this range matter more than the level. First, the whole 52-week range sits between $1.19 and $3.60 on a stock that floated at $16.00 in July 2021, so the shares are down about 86% from listing and the entire recent history is a penny-stock range [VERIFIED — price cache; the 2021 flotation price appears in the insider file as a $16.00 disposal line on 2021-07-06]. Second, the intraday low of $1.19 printed in late January 2026 and the high within the following two months; on closing prices the stock ran from $1.72 to $3.37 (+96%) between January and early April 2026 and then gave half of it back, with no clinical data released in either direction [VERIFIED — data/prices/ABOS.json, closes between 2026-01-01 and 2026-04-15]. That run is the run-up baseline for every program under this ticker, and it is recomputable from the committed cache for any date.

C.5 Ownership and flow

Insider flow. There has never been an open-market purchase by any Acumen officer or director. That is a statement about the entire 233-row file, which runs from the 2021-07-06 flotation to 2026-06-17, with no new row since the 2026-08-04 sweep [VERIFIED — BPIQ fetch_company_insider_transactions 2026-08-07]. Every acquisition in that file at a non-zero price is one of three things: an option exercised at its strike and paired with a matching option disposal on the same date; a private placement; or the conversion of preferred stock at flotation. The one line that looks like conviction buying, RA Capital’s 6,060,606 shares at $3.30 on 2026-03-16, is the March private placement.

Selling is continuous and it is mechanical. In the first three weeks of January 2026 the chief executive, chief financial officer, chief operating officer, chief medical officer, chief legal officer and chief regulatory officer all sold, at prices between $1.71 and $1.99, in the days immediately after the annual equity grant on 2026-01-16 — the classic sell-to-cover-tax pattern. In June 2026 director Derrell Porter exercised options at a $1.07 strike and sold 32,300 shares at $2.325 to $2.345, and director Nathan Fountain sold 6,400 shares at $2.2301 on 2026-06-09 [VERIFIED — same source]. None of this is a signal about the trial. What it establishes is the absence of the opposite signal: with a company-transforming readout roughly a quarter out and the stock at 1.9× its 52-week low, not one insider has bought a share in the market.

Concentration. Three funds held 24.20M shares at 2026Q2, about 33.5% of the shares outstanding: RA Capital Management 20,992,669 shares valued at $49.542M, ADAR1 Capital Management 2,156,334 at $5.088M, and Ikarian Capital 1,053,797 at $2.486M [VERIFIED — BPIQ fetch_company_hedge_fund_holdings 2026-08-07, unchanged since 2026-08-04]. All three implied prices are $2.36 per share, so the quarter is internally consistent.

RA Capital is the fact that matters. It holds 29.1% of the company, it has a board seat, and it has bought in three separate rounds at descending prices: 9,770,773 shares before 2023, 5,161,290 at $7.75 on 2023-07-21, and 6,060,606 at $3.30 on 2026-03-16. Every one of those tranches is under water at $2.28, and the most recent is 31% under water after five months. ADAR1 and Ikarian both roughly doubled or quadrupled their positions in 2026 at the same placement price. The register is therefore concentrated, anchored well above the current price, and in no position to be a seller into weakness — but also unlikely to fund a fourth round on the same terms.

Deep Track Capital held 2,513,044 shares at 2023Q4 and was gone by 2024Q1; Perceptive Advisors and BVF exited earlier [VERIFIED — same source]. The company was added to the Russell 2000, Russell 2500 and Russell 2000 Value indices on 2026-06-29, which brings passive holders but no view [VERIFIED — press feed 2026-06-29]. George Golumbeski was appointed chairman on 2025-11-10, taking the board to eight members [VERIFIED — press feed].

C.6 Options chain and its readability

Plain takeaway. The chain still cannot price this event, and it cannot even bracket it. There is no listed strike below $2.50 on a $2.28 stock, so nothing trades at the money. Total open interest across the entire December 2026 expiry is 153 contracts and total volume there on the day was 64 contracts, all in one $5.00 put. Any implied-move figure quoted from this chain would be a construction, not a price. The expected move for this readout has to be built from the scenario anchors in the program document instead, and that is what is done there. The one new fact is a modest accumulation of January 2027 calls — 320 contracts traded on the $5.00 call today against 186 of open interest — which is the first visible options positioning across the readout window.

[VERIFIED — BPIQ fetch_company_options_data 2026-08-07]

C.7 Reaction to past catalysts

Seven distinct event dates from fetch_company_historical_catalysts, six of which carry price data. The two 2026-07-15 rows are the same press release filed under two drug names and are shown once. Unchanged since the 2026-08-04 sweep.

The read that matters. Six measured reactions span −12.60% to +19.38%, and not one of them is an efficacy readout. Every positive row is a conference presentation, an enrolment milestone or a dosing announcement; the only negative row is a preclinical presentation that the market disregarded. Nothing in this feed tells you how ABOS trades on the ALTITUDE-AD result, because the company has never reported a controlled efficacy result in its life. Any expected-move estimate for Late 2026 that leans on this table is leaning on the wrong precedent, and the single most useful thing the table does say is that the market’s tolerance for non-ALTITUDE news has turned negative.

C.8 Company data-quality flags

• BPIQ max_52_week_position reproduces the documented previous-close bug exactly. The field reads 0.43568, which is what the formula returns on the previous close of $2.24. Since v5.1.0 this row is derived from the price cache instead; this sweep the two coincide because the cache’s last bar is the same $2.24 print.
• BPIQ market capitalisation is computed on the previous close ($161.8M = 72,227,586 implied shares × $2.24). The share count itself has been corrected since the 2026-08-04 sweep, when it was 657,000 shares stale against the 10-Q; it now matches the filing to within six shares. The previous-close bug family persists; the stale-count flag is resolved.
• BPIQ enterprise value is stale rather than wrong. The $58.8M figure deducts the 2026-03-31 cash balance from a 2026-08-07 market capitalisation and ignores the term loan. Recomputed: roughly $103M.
• BPIQ short_float units are not resolved for this ticker. The field reads 0.07616 and no second source was found. Treated comparatively, not as a precise level.
• BPIQ insider_own and tute_own carry no unit label, read here as fractions (6.80% and 64.74%). The institutional figure is corroborated by the 13F detail; the insider figure is not corroborated by anything.
• The 2023Q4 Boxer Capital 13F value reproduces the documented 1000× bug. It reads $9,158,000,000 on 2,206,983 shares, which is $4,149 per share. The share count is usable; the value is not.
• The insider-transaction feed is not limited to four filing quarters for this ticker, contrary to the tool’s own documentation. It returned 233 rows back to 2021-07-06, so the “no open-market buying ever” finding in C.5 rests on the full history rather than on one year.
• The historical-catalyst feed double-counts one press release. The 2026-07-15 AAIC announcement appears as two rows, ids 8189 and 8195, under ACU401 and Sabirnetug (ACU193), carrying identical price fields. Counting rows rather than dates would overstate the event history by one.
• Options implied volatility is unusable in a new way. On 2026-08-04 the December and January $2.50 calls were byte-identical at 1.99532; today they differ (1.65386 / 1.25387) but the September and October $2.50 calls print 0.21 and 0.16 against zero open interest, and three puts still sit exactly on the documented 0.01488 floor. Neither set is usable.
• BPIQ open_price_gap_percent and intra_day_price_change_percent disagree in sign on the 2026-07-15 row (+0.40% against −12.60%). Per 02-connectors.md the intraday figure is preferred and both are reported.
• Cash burn exceeds the reported net loss and the reported operating expense, at $8.08M a month against $6.90M and $7.07M. The difference is term-loan interest and working capital. The cash figure is used because it is the one that determines when the money runs out.