Startups face funding squeeze, but AI helps

Biotech startups are getting squeezed from both sides. Capital is harder to raise, and big drug companies are getting pickier about what they want to license. But AI is changing one important part of the equation — the cost of getting to a credible early result. That doesn’t fix the whole biotech funding problem. It does change which founders can get in the game, and which startup stories still sound fundable in 2026. ### What’s the squeeze? The simple version is that there are more young biotech companies chasing fewer willing checks. HSBC’s mid-2025 data showed first financings for biotech startups fell from $2.6 billion in Q1 2025 to $900 million in Q2, while total venture funding dropped from $7 billion to $4.8 billion. PitchBook saw the same pattern later in the year — biopharma VC in Q2 2025 hit multiyear lows, with seed and early-stage dealmaking stalling as investors turned risk-off on preclinical programs. (pitchbook.com) ### Why are investors acting like this? Because the old biotech financing ladder is wobbling. IPOs slowed sharply, crossover investors pulled back, and late private rounds stopped looking like reliable bridges to the public market. When exits get murky, venture firms stop wanting long, expensive science projects with unclear timelines. That pushes money toward companies that can show sharper milestones, faster learning cycles, or assets that already look partly derisked. (biopharmadive.com) ### Where does pharma fit in? Big pharma still needs startups. McKinsey notes that since 2018, more than 70% of new molecular entity revenue has come from externally sourced products. But pharma is also derisking its shopping list — doing fewer deals, paying up for stronger ones, and leaning toward later-stage or more validated opportunities. In plain English, startups are not just competing for VC dollars. They are also competing for a narrower definition of what counts as licensable. (biopharmadive.com) ### So why does AI help? AI can compress the most expensive part of company creation — the wandering around before you know whether a biological idea is worth real lab spend. Models for protein design, cell-state prediction, and virtual-cell-style simulation let teams test more hypotheses in silico before committing to full wet-lab programs. PitchBook’s 2025 note says investors increasingly favored AI technologies that promise to reduce development timelines and cost. (mckinsey.com) That’s the key shift — not that biology became cheap, but that the first proof point got cheaper. ### Does that mean a tiny team can now start a biotech? Sometimes, yes — but only for certain kinds of biotech. A founder with strong models, public datasets, and a tight biological question can now get much farther before building a huge lab operation. That favors AI-native drug discovery, design tools, and software-plus-biology workflows. It is much less helpful for startups whose real bottleneck is manufacturing, scale-up, or long experimental validation — the classic pain points in synthetic biology. (pitchbook.com) ### Why does this hurt broad synbio platforms? Because the old synthetic-biology pitch often needed a lot of capital up front — automation, foundry infrastructure, strain engineering, process development, and long iteration loops. In a loose market, investors would fund the platform story. In a tight market, that story can look too broad, too slow, and too expensive. Meanwhile, narrower AI-first companies can show progress earlier and raise on the promise of efficiency rather than scale. (pitchbook.com) That is a real competitive disadvantage for capital-heavy synbio plays. ### Are there signs this is already changing who gets funded? Yes. PitchBook’s 2025 AI drug-development note says AI-native drug developers captured most of the capital going into that category, and some became unicorns within their first two institutional rounds. Crunchbase also showed biotech’s share of total U.S. startup funding fell to just over 8% in 2025 — the lowest in more than 20 years — even as a handful of jumbo rounds still went to companies with unusually strong narratives or technical leverage. (pitchbook.com) ### What’s the bottom line? The funding squeeze is real. The pharma bar is higher. But AI gives founders a new wedge — start with computation, buy time with cheaper early evidence, and only then spend heavily on biology. That does not eliminate biotech risk. Basically, it changes where the risk sits, and who can afford to take the first shot. (pitchbook.com 1) (pitchbook.com 2)