Biotech in 2025: A retrospective 

As 2025 comes to an end, it’s a good moment to take a look at what happened this year in biotech. Unlike years defined by one big breakthrough, 2025 felt more like a phase shift, where many trends that had been building finally started to matter in practice. 

Metabolic medicines, especially GLP-1 and related incretin therapies, continued to dominate both research and development (R&D) and business headlines. At the same time, artificial intelligence (AI) moved beyond early promise into a more concrete role in discovery and development, with major partnerships and deployments across companies of all sizes. A number of collaborations are now explicitly focused on using AI platforms to accelerate molecular design and reduce development timelines, reflecting broader shifts in how science gets done. Other fields gained traction; radiopharmaceutical is a clear example. 

Funding and dealmaking remained a mixed picture, with ongoing venture activity for clinically validated platforms and significant mergers and acquisitions (M&A) as big pharma sought to refresh pipelines ahead of looming patent expiries. 

The metabolic health shockwave: GLP-1s and beyond 

If there’s one theme in biotech that defined 2025, it was metabolic health, and especially the continued spread of GLP-1-based therapies beyond their original niche.  

By early 2025, there were 11 GLP-1 drugs on the market for diabetes and weight management, with more than 40 in development. The field is now crowded and very competitive.  

Newer programs are not only chasing improvements in weight loss, but also broader cardiometabolic benefit, and insurers and health systems are having to rethink how they cover these medicines as they attract younger and healthier populations. 

“In 2025, GLP-1 medicines moved firmly into the broader cardiometabolic arena. Large clinical trials and real-world studies now show strong cardiovascular and renal protection, positioning these drugs as core tools for preventing heart and kidney disease. Early evidence also suggests potential benefits for metabolic liver disease, and observational data hint at reduced dementia risk among GLP-1 users, though definitive neurodegeneration trials are still pending,” said Momo Vuyisich, chief science officer at Viome Life Sciences. 

Strategic deals also reflected this trend. Throughout 2025, nearly every major pharmaceutical company moved to secure a position in the GLP-1 race. Pfizer struck a global licensing agreement with YaoPharma for an oral GLP-1 candidate and also outbid Novo Nordisk to acquire obesity drug developer Metsera. Roche signed a large collaboration with Zealand Pharma for a next-generation obesity therapy, and AbbVie entered the space via a multibillion-dollar licensing deal with Gubra. 

While GLP-1s were the dominant headline, 2025 also saw movement in adjacent metabolic targets. 

One notable example is AstraZeneca’s AZD0780, an oral PCSK9 inhibitor that demonstrated roughly a 50% reduction in LDL cholesterol on top of statins in phase 2. Because it is small-molecule and oral, this candidate could offer a more convenient alternative to existing injectable PCSK9 inhibitors. Biotech’s interest in the metabolic space isn’t limited to one mechanism. 

AI in biotech: from promise to practical infrastructure 

If 2024 was the year AI dominated biotech headlines, 2025 was the year it became something more ordinary, rationalized even: a standard tool in day-to-day R&D. The novelty wore off, but the impact became clearer. Across discovery teams, computational chemistry groups, and clinical operations, AI stopped being used as a differentiator and instead became part of how modern biotech works. 

A good example comes from the rise of physics-informed and structure-based models, which moved out of proof-of-concept territory and into production workflows. A growing number of companies now combine generative models with physical simulations to design proteins or small molecules in ways that avoid the pitfalls of purely data-driven approaches.  

This shift was visible at scale, too. As Stef van Grieken, chief executive officer (CEO) of Cradle, put it, “the turning point wasn’t a single breakthrough, it was the adoption of AI at scale due to the rise of software that actually works for big pharma. The barrier used to be technology, 2025 is when the accessibility barrier fell.” Pharmas that once ran limited pilot projects began integrating AI directly into their standard design workflows, making tasks like protein engineering or ligand scoring faster and more systematic.  

Indeed, in 2025, big pharma bet big on AI, from Eli Lilly to Novo Nordisk, going through AstraZeneca, to name a few, AI has been at the center of many strategic deals this year. 

According to Karine Sargsyan, director of operations at IBD Institute and Scientific Director of the OncoBiobank at Cedars-Sinai Medical Center, 2025 will be known as the year AI democratized cancer research. “AI/ML algorithms combed through petabytes of cancer biobank data for clues on tumor heterogeneity, potentially cutting preclinical timelines in half and unlocking insights into previously untreated populations.”  

Datasets are finally becoming large and curated enough for algorithms to extract insights that were previously inaccessible. 

Taken together, AI settled into the infrastructure, increasingly indispensable. The differentiation now lies not in whether companies use AI, but in how well they combine it. 

Radiopharmaceuticals become a real portfolio strategy 

Over the last few years, radiopharmaceuticals have gone from a specialized corner of oncology to one of the main strategic bets for big pharma. The M&A wave that started in 2023-2024: Novartis buying Mariana Oncology, Eli Lilly acquiring Point Biopharma, AstraZeneca snapping up Fusion Pharmaceuticals, and BMS taking over RayzeBio, set the stage for 2025. These deals were all about locking in access to platforms, manufacturing capacity and late-stage assets in a field that is capital-intensive and hard to copy. 

2025 started strong with a big win for radiopharmaceuticals as ITM-11, a targeted radiotherapy for gastroenteropancreatic neuroendocrine tumours, succeeded in phase 3. Novartis’ radioligand therapy Pluvicto also secured an expanded FDA indication in March, allowing use in PSMA-positive metastatic castration-resistant prostate cancer after androgen receptor pathway inhibitor therapy, but before chemotherapy – a move analysts said could roughly triple the eligible patient population. At ESMO later in the year, the PSMAddition trial in hormone-sensitive disease added more momentum by showing a 28% reduction in risk of progression or death when Pluvicto was added to standard of care. 

At the same time, companies like Novartis and Bayer kept pushing existing radiotherapies into earlier lines and new combinations. For example, new trials of Lutathera in first-line gastroenteropancreatic neuroendocrine tumors, and continued data generation around Xofigo in prostate cancer. Earlier-stage companies such as Actithera and Nuclidium also managed to raise significant funds. 

Market forecasts now put radiopharmaceuticals in the low-double-digit billion range by the early 2030s, with several analyses projecting high single- to double-digit annual growth. 

In 2025, radiopharma became a core part of oncology strategy, with all the associated headaches around isotope supply, specialized manufacturing, and clinical infrastructure. 

Modality bets: from neuroplastogens to precision psychiatry 

Neuroplastogens and the psychedelics’ “legitimacy moment” 

If radiopharma was one big modality bet in 2025, another came from an unexpected angle: non-hallucinogenic neuroplastogens and next-generation psychedelic-adjacent drugs for mental health. 

The headline deal came in August, when AbbVie agreed to buy rights to bretisilocin from Gilgamesh Pharmaceuticals in a transaction worth up to $1.2 billion. Bretisilocin is a short-acting 5-HT2A agonist in phase 2 for major depressive disorder, designed to deliver the therapeutic effects associated with psychedelics while limiting the duration and intensity of the psychoactive experience. 

“The AbbVie acquisition of Gilgamesh was a clear indicator because it signaled that large pharmaceutical companies are now ready to place massive bets on next-generation psychedelic treatments that deliver neuroplasticity without the patient-management burden of a psychedelic experience. It removed any lingering doubt for companies and investors about whether this class of molecules had commercial legs and shifted investor conversations from speculative interest to active pipeline building,” said Natalie Dolphin, managing director at HLTH Communications. 

On the biotech side, Delix Therapeutics spent much of 2025 turning the “neuroplastogen” concept into clinical data. Its lead candidate DLX-001 (zalsupindole), a non-hallucinogenic isotryptamine, generated positive phase 1b results in major depressive disorder, with rapid and durable antidepressant effects and biomarker evidence of target engagement.  

Precision psychiatry: biomarkers move from theory to trials 

A related, but distinct, modality bet played out in precision psychiatry. Here, the focus was less on one specific drug class and more on whether genetics and biomarkers can finally make a difference in how depression and other central nervous system (CNS) disorders are treated. 

One of the most widely discussed examples was Denovo Biopharma’s work on liafensine (DB104), a triple reuptake inhibitor for treatment-resistant depression. A study published in JAMA Psychiatry in 2025 reported that patients carrying a specific variant in the ANK3 gene responded significantly better to liafensine than to placebo, in a prospective biomarker-guided trial. This was described by the authors as the first successful use of a genetic biomarker to select responders in a prospective psychiatry study.  

According to Daniel Gehrlach at HMNC Brain Health, 2025 marked a turning point in how biotech approaches the complexity of psychiatric disease. “For years, predictive biomarkers in psychiatry were confined to academic curiosity and mostly focused on diagnosis. This year, we saw the first real evidence that genetically-guided treatment selection in depression can work at scale.” 

At the same time, he remains aware of the limits the field is still facing. “What 2025 also made painfully clear is that ‘multi-omics integration’ remains a practical challenge, not a turnkey solution. From our own multi-omic clinical trial datasets, it’s obvious why: enormous feature counts, heterogeneous phenotyping, medication histories, and batch effects rapidly drown signal unless you start with biological priors, carefully chosen candidate features, and uncompromising quality control.” 

Biotech funding in 2025: a selective market and a hollowed-out middle 

If you look at 2025 through the lens of capital, it still felt more like an extended winter than a clean rebound. Public biotech valuations stayed under pressure: sector indices remained well below their 2021 peaks, and many small and mid-cap companies continued to trade below their cash value. Initial public offerings (IPOs) did not “come back” in any broad sense. Reports on the 2025 IPO market stressed a focus on a small number of companies with late-stage data and clear paths to market, favoring quality over quantity. 

On the private side, the picture wasn’t much easier. Overall biotech venture funding, including first-time financings, fell sharply in the second quarter of 2025, reversing a better start to the year. In Europe, Citeline highlighted how fragile the situation was: total European biotech deal value in Q1 2025 dropped around 64% compared to Q1 2024, with IPOs scarce and M&A not really picking up the slack. 

There were signs of life later in the year. GlobalData reported that venture investments in biotech rose by about 70.9% quarter-on-quarter in Q3 2025, reaching roughly $3.1 billion, enough to suggest a change in mood, but not a return to the easy money of a few years ago. These rounds tended to cluster around companies with strong clinical data, clear commercial stories, or in hot areas like cardiometabolic disease and radiopharmaceuticals, rather than lifting the whole sector. 

Capital is available, but it is highly selective, and it flows to a narrow set of de-risked assets and platform stories rather than supporting a broad pipeline of early-stage innovation. 

The M&As that did happen in 2025 often reinforced the sense that biotech is being treated less as a set of companies and more as a collection of assets. A clear example came at the end of the year, when UK-based Cycle Pharmaceuticals agreed to acquire Applied Therapeutics, a once high-flying rare disease company that had traded above a $1 billion valuation earlier in its life, for just $0.088 per share in cash. After clinical and regulatory setbacks, Applied was effectively sold for parts, with Cycle mainly interested in its lead asset govorestat. 

“In the current cycle, big pharma treats most M&A as an asset trade: even when it buys entire biotechs, the economic logic is focused on a narrow set of pipeline assets, with the rest often wound down, spun off, or quietly deprioritized,” said Andrey Meshcheryakov of Recombinators. 

“The funding crunch and asset-centric M&A cycle have hollowed out the ‘middle class’ of biotech. Capital now bar-bells toward a few hyper-funded platforms and late-stage assets packaged for pharma, while many mid-cap biotechs are either trading below cash, being broken up, or sold for parts.” 

Funding pressures in 2025 played out differently across regions, but the underlying dynamic was similar with capital concentrated at the top. Europe struggled the most, while the U.S. remained the deepest market. China continued to supply innovation and partnerable assets even as geopolitical scrutiny tightened around data security and cross-border collaboration. 

2025 has seen a tension between two forces: strong innovation in specific areas and a capital model that only rewards a narrow slice of it. If 2025 had a single lesson, it might be that making progress in biotech is no longer just about having the right science or the right story; it’s about being focused enough to fit into a funding environment that has become both more demanding and more selective.