Oxford Biomedica: The AAV Overtake and a Guidance Cut That Wasn't Demand

A timing-driven 2026 trim, a resolved Durham delay, and a first-ever pipeline crossover as AAV opportunities top Lenti

OXB.L · Earnings Call · 2026-09-22

The timing-versus-demand distinction

Oxford Biomedica walked into its H1 2026 interim print with a guidance cut already public — the August trading update trimmed full-year revenue to GBP 180–200M. The whole job of this call was to prove the miss was clock-and-paperwork, not demand. Management spent the hour doing exactly that, and the strongest single adverb was "unchanged." “our full year 2026 guidance remained unchanged at GBP 180 million-GBP 200 million” — Frank Mathias, Chief Executive Officer · 2026-09-22, Frank Mathias repeated, with roughly GBP 168M already covered by contracted client orders. The cut itself, CFO Lucy Crabtree explained, was a stack of small, boring things: client programs deferred on strategy or clinical data, clients ordering in smaller work packages, one larger client changing approval pathway, and — the one that made headlines — a delayed Durham ramp. On that last point the tone was resolved, not defensive: “our Durham operations saw a delay of around six months on GMP operational readiness. However, we have now completed the first GMP run and execution is back on track” — Lucy Crabtree, Chief Financial Officer · 2026-09-22. The market's structural growth numbers were trotted out to separate cycle from stock: CGT assets rose from 2,155 (Q2 2025) to 2,217 (Q2 2026), but late stage assets — Phase II, III and pre-registration — grew 18% year-on-year. That is the nuance management wants investors to underwrite: the gene therapy market is maturing even where headline asset counts look flat. Under the hood, H1 revenue rose 10% constant-currency to GBP 80.2M, manufacturing up 20% on more clinical and commercial launch batches, while licenses and royalties slipped to GBP 1.2M as Kymriah matures. Operating EBITDA loss narrowed to GBP 7.8M from GBP 8.3M, and adjusted EBITDA loss improved to GBP 2.5M from GBP 3.9M. A GBP 7.6M impairment of French PP&E was taken on lowered near-term France expectations — the one genuinely negative footnote. The balance sheet closed with GBP 75.3M cash and GBP 21.4M net cash.

The AAV overtake — a genuine first

The freshest company-unique signal here isn't the guidance line; it's a pipeline crossover. For the first time, AAV opportunity value exceeds Lenti. “it is the very first time in the company's history that the volume of opportunities on the AAV side is above the volume of opportunity on the Lenti side” — Sébastien Ribault, Chief Business Officer · 2026-09-22, Chief Business Officer Sébastien Ribault said, quoting $289M of AAV opportunities against $265M of Lenti opportunities. Active client programs swelled to 59 from 44 a year earlier (+34%), with 29 onboarded and 14 finished, and the geographic mix shifted hard toward the U.S. (24% to 40% of pipeline) and France (13% to 24%). That mix change ties directly to the global footprint strategy — the U.K. Oxford site serves Lenti, not AAV, so the growth is loading into Durham and Lyon. The pipeline grew to $730M at the half (+30% YoY, $659M by end-August), contracted orders rose from GBP 97M to GBP 127M, and revenue backlog jumped from GBP 193M to GBP 299M. Nine commercial programs and late-stage assets now sit in the portfolio (up from seven), with several BLA submission dates expected over the next 12–18 months. The awkward part is ordering behavior. Rather than a one-off, Ribault framed it as a structural funding phenomenon.

Not a one-off and has nothing to do with the clinical data. It is purely a funding question or a financial risk mitigation plan that they have, where they go really step by step and save as much money as they can at each stage. It delays some of the decisions. The delay is minimal. It is couple of weeks between two stages. But if you sign the contract in five different tranches... at the end of the program, it is 8 to 10 weeks delay.
Sébastien Ribault, Chief Business Officer · 2026-09-22

He pointed to a sector snapshot showing $9.4B of worldwide H1 CGT investment, of which Europe was just $0.7B — thin funding begets tranche-by-tranche European clients. The offsetting resilience argument is client diversification: the largest single pipeline opportunity is now just 6% of pipeline value, with 71% spread across ~200 opportunities. Some of that is real conversion — 17 new clients signed in H1, already 21–22 since. But rising attrition rate is baked in as normal, with Ribault noting that 80–95% of projects fail for clinical or funding reasons, and he sees no meaningful difference versus 2025.

The CAR T wildcard and the market's odd silence

One under-discussed item deserves a flag: a safety scare in CAR T trials for autoimmune disease. Mathias addressed it head-on rather than ignoring it, drawing a boundary: “we have not seen any related issues in the few programs we are involved in, and as a result, our outlook remains unaffected” — Frank Mathias, Chief Executive Officer · 2026-09-22. The CAR T autoimmune exposure runs through just two named client programs — Cabaletta and Kyverna — with both having disclosed strong clinical data. This is a genuine two-sided risk: OXB's late-stage funnel leans on autoimmune CAR T, so a broadening safety concern in the field would hit the commercial-programs leg of the GBP 500M-by-2030 ambition. That ambition itself (roughly 30% EBITDA margin by 2030–31, ~18% market CAGR, ~6% current share) is a long bridge from a loss-making H1, and the call's most revealing structural data point is the process development tranche behavior — clients buying feasibility now and GMP later, which slows recognition even as backlog builds. Meanwhile the broader market tape is voting elsewhere entirely: 360/90/30-day advancers and decliners are dominated by AI data-center, memory and specialty-semiconductor names, with CGT and biotech largely absent from the top movers. OXB is riding a wave the tape isn't yet pricing — which cuts both ways. The August cut did its damage; this print argues the demand is intact, the AAV inflection is real, and the miss was sequencing. Whether the market forgives the sequencing is the open question, and no price record was available to adjudicate it.