Pharmaceutical companies are struggling with artificial intelligence – but do not blame artificial intelligence

more than He immersed $ 18 billion in more than 200 biotechnology companies Amnesty International is promoted to accelerate development, with 75 drugs or clinical experiments entering, according to L. Boston Consulting Group. Now, the investor’s confidence – financing – has begun to hesitate.

In 2021, investment capital investment in artificial intelligence pharmaceutical companies reached a top with more than 40 deals worth about $ 1.8 billion. This year, there were less than 20 deals worth about half of this peak amount, the Financial times I mentionedQuoted from the data from Pitchbook.

Some current companies have struggled to face challenges. In May, the biotechnology company returns Three of potential drugs provided In an effort to reduce costs after merging with Exscientia, a similar biotechnology company, last year. luck I previously mentioned that any of the obstacles to the artificial intelligence has reached the market as approved drugs. After a The main restructuring In December 2024, Biotech Company Bevolentai was deleted from the Euronext Amsterdam Stock Exchange in March before merging with Osaka Holdings.

A spokesman for Odeh said luck The decision to attach the drugs was “paid to the data” and as a planned result to integrate them with Exscientia.

“The rate of 90 % failure in our industry is unacceptable when patients wait, and we believe that the methods of us that merge the tools and advanced technologies will be the best in the long term,” the spokesman said in a statement.

Beevolentai did not respond to the comment.

Industry conflicts coincide with a broader conversation about the failure of artificial intelligence to connect more quickly to its noble promises to productivity and efficiency. Massachusetts Institute of Technology a report Last month, 95 % of the artificial intelligence pilots in companies found that they failed to accelerate revenues. American census office reconnaissance This month, the adoption of artificial intelligence in large American companies has decreased from its peak by 14 % earlier this year to 12 % as of August.

But the artificial intelligence technique used to help develop medicines is completely different from those resulting from the large language models used in most of the workplace initiatives, and therefore should not be held according to the criteria themselves, according to Scott Shawainhaus, the administrative director and stock research analyst at Keybanc Capital Markets Inc. The industry faces a set of challenges.

“Regardless of the amount of data you have, human biology is still a mystery,” said Shwainhaus luck.

The total and political factors that dry up the financing of drug development artificial intelligence

Shwinhaus said that the essence of slow financing and development results are slower and they may not be restrictions on the technology itself, but have reached a large number of wider factors.

He said: “Everyone admits that the financing environment has dried up.” “The biotechnology market is strongly affected by low interest rates. Low interest rates are equal to more financing in biotechnology, which is why we see vital technology financing at its lowest levels over the past few years, because interest rates have been high.”

It was not always that way. Take advantage of artificial intelligence in the development of drugs is not only an increase in access to semiconductor chips, but also how technology allowed fast and cheap ways now to draw the entire human genome maps. In 2001, it cost more than $ 100 million to draw the map of the human genome. After two decades, this pledge costs about $ 1,000.

Besides the epidemic to thank them for the future interest rates in 2021, Covid also impeded partnerships between AI’s drug development companies and major pharmaceutical companies. In early 2022, ABCELLERA and the start of vital operation and Eli Lily I got an emergency FDA approval For an antibiotic used in early Covid vaccines, a tangible example of how technology is used to help with drug discoveries.

But since then, there have been obstacles in other industry, as Shawin Hasus, including Big Pharma, said Reducing research and development costs Amidst the slowdown, as well as the uncertainty surrounding whether President Donald Trump will impose a A tariff for medications The American and European Union also wanders in a commercial deal. Trump A note signed This week is threatened to ban direct ads for the consumer of prescribed medicines, as the theory leads drug revenues.

Restrictions of artificial intelligence

This does not mean that there were no technological hiccups in this industry.

“There is an audit about the technology itself,” said Schwinhaus. “Everyone is waiting for these readings to prove this.”

Shwinhaus said that the next 12 months of data from emerging companies to develop medicines from artificial intelligence will be decisive in determining the success of these companies. Some results have been mixed so far. For example, return Data released From a clinical trial in the middle of the stage of a drug to treat neurological vascular condition in September last year, it was found that the drug was safe but there was little evidence of its effectiveness. The shares of the company decreased double numbers after the announcement.

These companies are also limited to how they are able to take advantage of artificial intelligence. The process of developing medicines is 10 years old and is intentionally penetrated to ensure the safety and effectiveness of relevant medicines, according to David Sidrovsky, head of pharmacy and neuroscience at North Texas University, who previously worked with pharmaceutical development companies in the private sector. Biotechnology companies that use artificial intelligence to make these operations are usually more efficiently treating a small part of this bottle neck, such as the ability to examine a drug -like molecule faster than before.

“There are many stages that must be jumped before you can actually announce the (European Pharmaceutical Agency), the FDA (FDA), or Canadian Health, whatever, this will appoint this as a safe and approved drug that is marketed to patients in the world,” luck. “The early bottleneck of test compounds is not everything from satisfactory shareholders through advertising,” we have approval of this compound as a drug. “

The smaller companies in this sector also made concerted efforts to partnership with large pharmaceutical companies, and instead exposed the construction of their pipelines, even if he no longer has access to franchise resources in the industry giants.

“They want to be able to follow their technology and show verification of their platform in a closer time, not later,” said Shwainhaus. “They will not wait for a partner molecule.

Schoenhaus believes that this strategy is a means of companies that look forward to proving themselves by mastering the use of artificial intelligence to understand the best slippery, mysterious boundaries and are largely known to human biology.

He said: “It is just a more complicated application than artificial intelligence, and for this reason I think we still see these companies focus on their internal pipelines so that they can focus directly on their resources on trying to improve their technology.”