With every week that passes, it seems a pharmaceutical company has tied a new deal with an artificial intelligence (AI) provider.
The US Food and Drug Administration (FDA) is approving more novel drugs year on year. In 2023, the agency approved 55 novel drugs 鈥� up 50% compared to 2022. With heightened drug development activity comes higher costs for companies.
AI makes bringing drugs to market cheaper, faster, and more reliable, as per its proponents. Across healthcare, the AI market is expected to be worth $908.7bn by 2030 according to . Generative AI (genAI) especially, presents a streamlined efficiency across the pharma value chain.
鈥淕enAI is revolutionising healthcare, fundamentally altering the industry鈥檚 approach to efficiency and productivity,鈥� says Saurabh Daga, associate project manager of Disruptive Tech at GlobalData.
One of its most popular applications, and what many high-value pharma deals have recently centred around, is at the pre-market stages in a medicine鈥檚 lifecycle.
But whilst the genAI buzz engulfs the pharma landscape, there remain questions over the transparency of its use in drug development and how its success is assessed as regulation struggles to keep up with technological advancements.
How well do you really know your competitors?
Access the most comprehensive Company Profiles on the market, powered by GlobalData. Save hours of research. Gain competitive edge.
Thank you!
Your download email will arrive shortly
Not ready to buy yet? Download a free sample
We are confident about the unique quality of our Company Profiles. However, we want you to make the most beneficial decision for your business, so we offer a free sample that you can download by submitting the below form
By GlobalDataBig deals and big investment
Perhaps the clearest indication of genAI鈥檚 growing role in drug development is the money exchanging hands.
One of the biggest deals in recent times involved (BMS), who tied down the services of US deep tech company VantAI for $674m in February this year. follows closely behind, having agreed to part with a potential $140m to work with Paris-based Aqemia 鈥� which uses genAI and deep physics algorithms in drug discovery. MSD and Genentech have also partnered with tech companies and to shore up their AI capabilities.
NVIDIA has quickly become an undisputed leader in providing AI tools. Having started as a video game graphics card developer, it has become one of the most profitable tech companies in the world, boasting a market capitalisation of $2.57trn. The company has made inroads into healthcare with deals with major companies like , Roche鈥檚 Genentech, and Johnson and Johnson MedTech 鈥� the first two deals being
鈥淭he pharma industry is known to be slow when adopting new technologies and innovative tools. However, in recent years, driven by the Covid-19 pandemic, the pharma industry has undergone increased levels of digital transformation,鈥� says GlobalData Healthcare analyst Wafaa Hassan.
鈥淭he availability of ultra-large datasets and technological advances has led to more interest in the use of AI and big data analytics across the pharma value chain, from drug discovery and clinical trial design.鈥�
The proliferation in deals affirms pharma’s pro-licensing outlook for genAI鈥檚. Big pharma players have invariably remained tight-lipped on in-house AI capabilities, but the size of the deals suggests that the majority of their know-how comes from bringing AI platforms in from third parties.
鈥淧harma companies are set up in a way that they usually don鈥檛 have an AI team that can do everything in-house. They are open to collaborating with outside vendors who demonstrate a tool can help develop a pipeline, says Weijie Sun, CEO of DP Technology, a software-centric startup that provides AI tools and services for companies in the life sciences space.
Do deals involve transparency?
Whilst companies tying the big dollar knots on sharing AI knowledge in exchange for upfront payments attracts much of the spotlight, questions over how the technology is used have led to the FDA and European Medicines Agency (EMA) issuing good-practice guidances.
The reflection papers, which guide drug sponsors on how to use AI and machine learning (M/L) algorithms, are intended to improve the transparency of AI鈥檚 application in drug development. The EMA has said that although AI鈥檚 application in drug discovery is low risk from a regulatory perspective, sponsors must 鈥渕itigate ethical issues, risks of bias and other sources of discrimination鈥� from a data quality point of view.鈥�
While companies are not mandated to follow these recommendations, they will become increasingly important. Data suggests the AI-discovered molecules are successful in clinical trials 鈥� and even pass through at a higher success rate than historical averages. One analysis has found a in Phase I trials for AI-derived molecules. However, this success rate drops to approximately 40% at the Phase II stage.
Despite the current lack of legislation, software companies are optimistic about how regulation will impact the space. 鈥淸New regulation] will obviously impact the way applications are designed in terms of transparency and in terms of GDPR. Highly regulated industries like pharma can give a window into what is possible with this technology 鈥� there is some constraint, but ultimately it is needed to mature applications,鈥� says Yseop CEO Emmanuel Walckenaer. Yseop, a company developing genAI tools for regulatory document generation, works with the likes of and Sanofi.
Measuring impact
On the level of a software provider and pharma company, success can even be measured by something as simple as how well the technology is actually used, according to DP Technology鈥檚 Sun.
鈥淲e have our own drug discovery team who are the first users of our software 鈥� whether our algorithms are released to the public depends on their feedback. If they open their computer up and use the software for lots of tasks, not just for one case, then this will be very successful in the long run.鈥�
The burgeoning of AI tools within drug development has also produced an unintended benefit of measuring success. Sun says, 鈥渢he good thing about drug discovery is that there are existing tools out there, so it’s quite easy to do a head-to-head comparison. If we can fit into a customer鈥檚 workflow, it鈥檚 easy to define what success looks like.鈥�
Companies with in-house tech teams that also develop their own drug pipelines, like DP Technology, have an advantage too. This new business model affords smaller biotechs agility in their chosen disease spaces, allowing AI to both guide and accelerate clinical programmes.
According to a 2023 report by the Convention on Pharmaceutical Ingredients (CPHI), biotechs with both the proprietary AI and patented pipeline are especially attractive to investors.
Steadfast investment
The CPHI report indicates that despite unsteady regulatory footing and relatively opaque applications, investment into AI-centric biotechs will continue.
Genesis Therapeutics 鈥� an AI-led drug development startup raised $200m in a Series B round last year with backing from Eli Lilly. Evozyne, which puts proteins through millions of years of simulated evolution, raised $81m in September 2023. Paris-based biotech Aqemia, which has an ongoing partnership with Sanofi, has raised $65.3m in Series A financing for its genAI-derived drug pipeline.
The McKinsey Global Institute (MGI) that genAI is estimated to produce between $18bn and $53bn in annual revenues across research, early discovery and clinical development.
Hassan believes deals are likely to increase in value, and the fruitful capital-raising landscape for emerging companies will continue. 鈥淧harma firms will continue to invest in AI as its tools play an important role in accelerating drug discovery and enhance clinical trials. In GlobalData鈥檚 Digital Transformation and Emerging Technologies report that was produced in October 2023, a survey showed that AI, followed by big data, were viewed as the most disruptive technologies over the next two years.鈥�
The dynamism of impending regulatory changes, coupled with a talent pool still in its infancy, means the road to widespread AI acceptance is still long, however.
鈥淓stablishing the safety, efficacy, and reliability of AI-based drug discovery methods to meet regulatory standards remains a significant hurdle,鈥� says Hassan.
