BioNTech proved mRNA worked. Now its founders are leaving, and the pipeline is still unproven.
Ugur Sahin and Özlem Türeci built BioNTech into the poster child of the mRNA revolution: a German biotech that, alongside Pfizer, turned a pandemic into proof that an entirely new drug platform could work at commercial scale. In 2022, Comirnaty generated €17.3 billion in revenue, according to BioSpace. The company used that windfall to buy a manufacturing facility in Singapore, promising a regional hub that would create more than 100 jobs and be operational by late 2023, per BioNTech's acquisition announcement.
On March 31, 2026, BioNTech announced it would close that facility by the end of February 2027. The site never fully came online. It has 85 employees today, according to the Straits Times.
The Singapore exit is a facilities story and a statement about where BioNTech thinks its future is. The company is pivoting hard to oncology, betting that seven late-stage clinical readouts due in 2026 will justify the pivot from the pandemic-era revenues that funded it. The readouts cover a PD-L1xVEGF-A bispecific antibody called pumitamig developed with Bristol Myers Squibb, an ADC dubbed BNT323, and an anti-CTLA-4 candidate called gotistobart. BioNTech expects to have 15 Phase 3 trials running by year-end, per its January 2026 strategic update.
Those are real trials. Phase 3 is not a suggestion. But late-stage readouts are not approved products, and oncology drug development has a well-documented habit of breaking hearts in the final act. BioNTech's 2025 revenues were €2.9 billion, up roughly 4% from 2024's €2.751 billion, according to its Q4 2025 financial results, and they are not going to sustain the investment without a pipeline win.
The financial pressure is real. BioNTech guided to €2.0–2.3 billion in 2026 revenue, roughly a 25% decline from 2025, per the same financial results filing. Capital expenditures were €307.1 million in 2025, according to BioSpace, and the Singapore facility was a contributor to that spend. Closing it removes a line item and a platform the company had explicitly framed as its gateway to Asia-Pacific.
Sahin and Türeci announced on March 10, 2026 that they would depart by year-end to start a new mRNA venture. BioNTech retains a minority stake and will license technology to the new company. The stock fell roughly 17% that day. Sahin and Türeci are the company's scientific identity and its narrative spine. A company that spent five years saying "mRNA is a platform" is watching its platform architects leave to build something new, while the original company pins everything on readouts that are still, in clinical terms, a hypothesis.
Meanwhile, Chinese mRNA companies are not retreating. Stemirna Therapeutics operates manufacturing lines capable of producing 100 million mRNA vaccine doses annually, funded by nearly $200 million in equity rounds, according to Mordor Intelligence. These are not science projects. They are industrial-scale operations targeting the same Asia-Pacific markets BioNTech is exiting.
[Correction: The Stemirna capacity figure originates from a BioNTech investor presentation and has not been independently verified.]
Singapore's biotech sector accounts for 1.9% of the city-state's GDP and employs more than 26,000 people, according to the Economic Development Board. BioNTech's departure is a footnote in that ecosystem, not a structural blow. But it is a signal. A company that proved mRNA could work at commercial scale looked at Asia-Pacific, built something, and decided it was not worth staying for.
The oncology bet is coherent. mRNA's versatility is genuinely attractive as a cancer therapy approach, and BioNTech's checkpoint and ADC programs are not trivial. But "poised to be a catalyst-rich year," as the company's January 2026 strategic update put it, is not the same as "here is an approved drug." The gap between those two sentences is measured in failed Phase 3 trials and rejected applications. Whether BioNTech can clear it without the scientists who built the original platform is the question the 2026 readouts will eventually answer.