BioInvent Buys Its Own Patience: A SEK 430 Million Rights Issue Where a Partner Was Supposed to Be

BioInvent Buys Its Own Patience

For about a year, BioInvent's answer to the question "how do you fund 2027?" was partnering. On the Q3 2026 call, the answer is a rights issue: SEK 430 million gross, SEK 350 million of it locked up via subscription commitments and guarantee undertakings from Redmile, Van Herk and Forbion, plus HealthInvest. The exchange is small — market cap sits around $37 million — and the announce came the night before the call, ahead of a December 1 EGM and a subscription window running December 7 to 21. That sequencing is the story: the science got a validation upgrade, the balance sheet got a shareholder-funded extension, and the deal that was supposed to do the job instead moved further away.

The data got better. The clock did not.

The clinical update is genuinely a repeat-with-more-patients rather than a new claim. The BI-1808 plus Keytruda ovarian dataset shown at ESMO reproduced the ASCO signal in a larger set: 24% objective response for the doublet versus 8% for pembrolizumab alone and 6% for BI-1808 alone, disease control of 53%, six patients still on treatment at the September cutoff, and under 5% treatment-related discontinuation across 81 solid-tumour patients. The poster took best-in-class honours in the investigational immunotherapy section, and the compound now carries an FDA Fast Track designation in recurrent ovarian cancer on top of the CTCL one.

What is interesting is where management chose to plant the flag. Not response rate — progression-free survival. Median PFS on iRECIST was 10.3 months against a 2.1-month pembrolizumab benchmark from KEYNOTE-100, and “The PFS in the next coming readouts will be essential. I think that's where we will have, I think, a huge increase.” — Andres McAllister, Unknown · 2026-10-29 That is a deliberate repositioning against the antibody-drug conjugates. Andres McAllister's argument is structural: there are roughly twenty ADC agents in or entering Phase III chasing only two payloads, whereas in immuno-oncology there is one approved backbone and BioInvent's asset is an add-on to it. The mode of action — first-in-class TNFR2 agonism, with the company claiming to be the only one developing it in oncology — is the moat and the problem at the same time.

The next step is the triplet: randomized, three arms of 20 patients, two doses of BI-1808 plus a control, all on pembrolizumab with weekly paclitaxel plus or minus investigator-discretion bevacizumab, mirroring KEYNOTE-B96. Start Q1 2027, first data H2 2027, with targets of roughly 60% ORR and a three-month PFS add. Management also flagged a deliberate enrichment for Clear cell carcinoma — a subtype that responds poorly to chemotherapy and where a clean signal could shorten the road to approval.

The partner that keeps being six months away

Here is the tension worth naming. In February, the CEO told the same analyst crowd: “I think we might already see some activity around partnering — potential partnering, deal making, during the second half of this year.” — Martin Welschof, CEO · 2026-02-26 Asked then about funding 2027, the answer was “we're looking at partnering... we have a strong focus on partnering.” — Martin Welschof, CEO · 2026-02-26 Asked in October 2025 when the BI-1206 pembrolizumab data would land, the guidance was “during the second half of next year” — Martin Welschof, CEO · 2025-10-29 — i.e. H2 2026; by February it was still "second half of this year" and pivotal-ready programmes were the talking point. Today that readout is H1 2027, and the explanation is “the delay is basically that we want to mature the data a little bit more.” — Martin Welschof, CEO · 2026-10-29 Recruitment, he insists, was fine.

You know that what we are developing is first in class... The problem with that is the current markets, they are really mostly focused on best in class, not so much on first in class. The interest is high, so we have no problems to get meetings with corporates as well as with large institutional investors. I think the interest is actually getting better. What they will have to see in order to get comfortable or more comfortable is more data.

That is an unusually honest diagnosis, and it cuts against the global tape. Across the last 90 days, the market's clinical-stage biotech keywords are the ones advancing — novel mechanism of action, open label extensions, patients on therapy, net product sales. Investors have been paying for clinical de-risking events. BioInvent's read is that they are paying for a different kind of de-risking: same-mechanism, bigger-N, faster-to-approval. For a first-in-class asset with 81 patients, that means the ESMO repeat, however well-awarded, does not clear the bar; only the triplet does. Hence the external adviser engaged to run a business development process on both assets, and hence the equity.

What management stopped saying

The quiet parts are as informative as the loud ones. The CTCL path, previously framed as a possible standalone pivotal, is now explicitly partnered: “But obviously we will not run that pivotal study on our own. That should be done with a partner.” — Martin Welschof, CEO · 2026-10-29 In February the "dream scenario" was keeping 1808 longer and partnering 1206 first; today 1808 is the priority and 1206 is the one parked in front of a partner — the sequencing inverted by circumstance, not strategy. The organisation is being cut by up to 20 FTEs, roughly a fifth of a company this size, concentrated on the deprioritised programmes. Those are the ones that fell off the vocabulary: BI-1607 and BI-1910 now appear only as line items explaining lower cost, and BT 001, a top-five keyword two quarters ago, has vanished entirely.

The financials explain the urgency without much drama. Q3 net sales were SEK 10.8 million against 3.3 million, a SEK 7.5 million bump from antibody production for customers; the nine-month figure of SEK 37 million compares with SEK 224 million a year ago, and that gap is almost entirely the absence of the $20 million XOMA Royalty payment for mezagitamab plus a $1 million milestone — this year's equivalent was a single €1 million milestone from Bayer. Operating costs actually fell, to SEK 131 million in the quarter and SEK 404 million for nine months, with the mix shifting toward BI-1808 and away from the shelved assets. The Q3 loss was SEK 119 million, SEK 361 million for nine months, and liquid funds plus current investments ended September at SEK 225 million. CFO Stefan Ericsson's runway arithmetic is unambiguous: “With the secured part of the rights issue completed at SEK 350 million in gross proceeds, we are funded until, like Martin said, end of 2027.” — Stefan Ericsson, CFO · 2026-10-29 Without it, financing runs to the latter part of Q1 2027 — the same asset base, twelve fewer months.

The bet in one sentence

BioInvent is asking shareholders to fund a single interim look — an interim readout of 30 patients in the Part C triplet — on the theory that a first-in-class mechanism with clean safety and a three-month PFS delta sells itself to a large pharma once the pattern is visible in the clinical setting that will actually be used. Everything else has been trimmed to protect that. The risk is not scientific, it is temporal: “It's, of course, always difficult to forecast” — Martin Welschof, CEO · 2026-10-29 is what the CEO now says about partnering, and the company has just priced the possibility that it stays true for another year. The one thing that would change the story fastest is the first 10 to 20 triplet patients — or a BI-1206 lung dataset in H1 2027 that finally makes the external adviser's phone ring.

A retirement, a board nominee in Dharminder Chahal returning with Van Herk's backing, a rights issue and a fifth of the workforce gone: the corporate housekeeping of 2026 is all in service of 2027 being the year of data rather than the year of the deal.