When a concept forms a consensus in a short period of time and everyone thinks it's a big deal, it may be a dangerous signal for entrepreneurs. To some extent, AI pharmaceuticals are just like that.
In the past few years, pharmaceutical companies of all sizes around the world have been embracing AI. Their purpose is very clear: to optimize the discovery time of innovative drugs and increase the success rate of drug research and development through the empowerment of AI technology.
At present, it seems that AI-empowered drug development is not impossible. However, as a seller of water, the rise of AI pharmaceutical companies may still need time.
Especially in the cold winter, the AI pharmaceutical boom is gradually fading, and challenges keep coming. In the past year, the first batch of molecules created by artificial intelligence can be described as a large number of casualties, some directly suspended research and development, some lowered the priority of clinical trials.
AI pharmaceutical companies have to start looking for a way out. On August 8th, the largest merger in the AI pharmaceutical industry was born: Recursion and Exscientia announced that they have reached a final merger agreement.
Behind the merger, the most fundamental reason is that they have not yet produced impressive clinical data, and this merger is more like seeking warmth in unity.
Is the merger a good way out? At least in the eyes of the co-founder of Recursion, this is a very complementary choice, &#34;it feels like we are just getting started.&#34;
But as a star company in the field of AI pharmaceuticals, Exscientia will no longer exist after the merger. A dream of 12 years, Exscientia has become a footnote to the outbreak and burst of the AI pharmaceutical field.
To some extent, Fu Sheng's reflection on the big model bubble also applies to AI pharmaceuticals. &#34;I was wrong to treat the means as the end, thinking that I was the pride of heaven, and created a very awesome company, that was the fantasy I gave myself.&#34;
Of course, in the biotechnology industry, such phenomena may not only occur in the field of AI pharmaceuticals.
<h2>01 Merger = Seeking Warmth in Unity</h2>
In times of industry difficulty and transformation, mergers are always highly anticipated.
The merger of Recursion and Exscientia is the largest recent merger in the field of AI pharmaceuticals. Although the two companies have not officially disclosed the transaction amount, Reuters reported that the transaction amount is $688 million.
According to the agreement, the new company will retain the name Recursion and be headquartered in Salt Lake City, Utah, USA. Recursion's existing shareholders will ultimately hold about 74% of the shares, with the remaining 26% held by Exscientia shareholders.
In other words, Exscientia, founded in 2012, is one of the earliest AI + drug development companies in the world and will no longer exist.
However, this is also understandable. As of the first quarter, Exscientia had only $178 million in cash, and the company lost $146 million last year. Since last year, it has experienced quite a bit of turmoil, pipeline failures, founder dismissal for misconduct, layoffs
Of course, the situation with Recursion can only be said to be slightly better.
As a AI pharmaceutical company strongly supported by NVIDIA, Recursion is simultaneously building a large-scale AI model for generating biological molecules and advancing a huge pipeline combination of clinical trials across multiple therapeutic areas. Due to the need for platform expansion, Recursion has started acquiring small AI pharmaceutical companies. On May 8th last year, Recursion acquired Cyclica and Valence for a total price of $87.5 million.
Operating on multiple fronts has led to a high loss of $328 million for Recursion last year. As of the second quarter, Recursion only had $474 million in cash, facing considerable cash flow pressure.
More importantly, both Exscientia and Recursion have not yet produced impressive clinical data. In order to survive, they have chosen to merge to stay warm. It is expected that after the merger, they will save approximately $100 million annually, and the cash flow will extend until 2027.
On one hand, they hold the two largest collaborations in the AI pharmaceutical field. At the end of 2021, Recursion reached a potential collaboration worth $12 billion with Genentech; at the beginning of 2022, Exscientia reached a collaboration worth approximately $5.2 billion with Sanofi. In the next two years, there may be approximately $200 million in milestone payments.
On the other hand, their drug pipelines do not overlap, planning to have 10 clinical data readouts in the next 18 months. Facing the scrutiny of clinical trials is a necessary step for all AI drug molecules to address doubts and hype. If successful, it will further boost market confidence.
So, is mergers and acquisitions a good way out?
At least in the eyes of the parties involved, this is a very complementary choice.
Chris Gibson, co-founder of Recursion, said: &#34;Exscientia's precision chemistry tools and capabilities, including its newly deployed automated small molecule synthesis platform, will enhance our technological support for biological and chemical exploration, hit discovery, and conversion capabilities. I am excited to continue building the best example of the next generation of biotechnology companies. I feel like we are just getting started.&#34;
<h2>02 Revenue generation capabilities below expectations</h2>
There was a time when the concept of AI pharmaceuticals was popular, and the first pioneers hoped to use artificial intelligence technology to create higher-quality drugs in a shorter time and at lower costs.
However, now, the first batch of AI-designed drugs entering clinical trials have encountered setbacks one after another, and all AI pharmaceutical companies are facing market skepticism: when will the profit model work out?
So, how can AI pharmaceuticals generate profits? Currently, there are three main business models:
One is the AI-CRO model that provides outsourcing services to drug development companies, jointly advancing pipeline development;
Another is the SaaS vendor model that primarily provides software platform services;
The third is similar to Biotech, conducting new drug research and development on their own
Exscientia is a typical representative of the first model. It mainly advances the development of pipelines in collaboration with a large number of external companies, using extensive cooperation to accumulate more data to support the optimization and iteration of its algorithm models.
In 2021, Exscientia advanced the world's first AI-designed new drug, DSP-1181, into clinical trials, but the following year, its partner Sumitomo chose to stop development. In early October 2023, Exscientia halted the Phase I/II study of EXS-21546 because the drug did not achieve the expected efficacy.
Despite collaborating with many pharmaceutical companies and having AI capabilities, the research and development of innovative drugs remains challenging.
Recursion's business model can be summarized in three points: allocating some funds for self-developed product pipelines; collaborating with pharmaceutical companies to discover drugs using AI; and expanding research assets through licensing agreements.
In the first half of 2023, the company's founder, Chris Gibson, stated that the future focus would be more on advancing the development of its proprietary pipeline. However, by the end of the year, he believed that platform companies would have more opportunities in the future because the funding required to support both the platform and pipeline simultaneously was too large.
At the same time, he also mentioned exploring new models, hoping to license their software to pharmaceutical companies for their own projects and data, similar to the approach taken by the established company Schrodinger.
In other words, Recursion is currently more inclined towards a business model of selling software and services. Interestingly, Schrodinger is accelerating the layout of its self-developed pipeline.
On the surface, whether selling software or services, both are considered good models.
Because AI pharmaceutical companies essentially play the role of CRO/platform, or &#34;water sellers.&#34; However, the ceiling for AI pharmaceutical companies in this model remains unknown.
Taking selling software as an example, although the top 20 pharmaceutical companies globally are clients of Schrodinger, and 1,250 research personnel from academic institutions also use its drug discovery software, the revenue generated by selling software products for Schrodinger in 2021-2023 was $113 million, $136 million, and $159 million respectively.
Clearly, if the path to commercialization is limited to selling software, Schrodinger will find it difficult to make significant profits. This is also why it is expanding into innovative drug research and development.
In the end, the market has long seen through that AI pharmaceuticals only bring a possibility of increasing success rates. Whether this is truly feasible still requires subsequent research such as animal experiments, clinical trials, and a series of studies to verify.
Currently, pioneers like Exscientia have not yet presented impressive clinical data and have not truly helped pharmaceutical companies reduce costs and increase efficiency.
As for the third model, self-research and development of innovative drugs, ultimately relying on commercialization for monetization or earning milestone payments through external licensing. In this way, AI pharmaceutical companies are no different from biotech companies
However, if one is not careful, the shortcomings of AI companies and innovative pharmaceutical companies will be concentrated. For example, BenevolentAI, whose AI-developed local pan-Trk inhibitor BEN-2293 for treating atopic dermatitis failed to treat patients in the Phase IIa clinical trial, which is the company's only clinical pipeline and its reliance for going public.
After the clinical failure, BenevolentAI had to undergo large-scale layoffs and restructuring.
<h2>Escape from AI Pharmaceuticals</h2>
The prospects of AI pharmaceuticals have long been fully hyped.
For example, Morgan Stanley pointed out in a report that the global market size of AI pharmaceuticals has reached $50 billion in the short term and may continue to rise.
However, this can only explain that the AI pharmaceutical market has entered a fully activated state, not necessarily that the market trusts that its business model will definitely yield results. As mentioned earlier, as the model has not yet been proven, almost all AI pharmaceutical companies are experiencing unprecedented pressure and difficulties.
The bubble created by AI pharmaceuticals has been burst. Not only have the stock prices of various companies plummeted, but more importantly, insiders have also begun to &#34;escape&#34;.
The protagonist of this acquisition, Recursion, positions itself as a TechBio company, using computational tools and emerging machine learning/artificial intelligence tools to understand data. The acquisitions of Cyclica and Valence Discovery last year were also aimed at enhancing its capabilities in digital chemistry, machine learning, and generative AI.
In Recursion's words, the company is leading the transformation of biotechnology (BioTech) to tech biology (TechBio), with a team of over 500 employees (Recursionauts) consisting of 35% with a life science background and 40% with a computational and technical background. This is also an important factor in Recursion defining itself as a TechBio company.
As early as last year, facing the broad prospects of AI pharmaceuticals, Schrödinger was trying to draw a clear line with AI. When an analyst referred to Schrödinger as an AI pharmaceutical company, the company's CFO Jeffrey Porges immediately interrupted and stated that Schrödinger is not an artificial intelligence company but a pharmaceutical company with proprietary software.
Schrodinger's CEO stated, &#34;To me, being described as an AI company is like describing oneself as a company that uses Office software.&#34;
On its official website, the latest introduction has been changed to: Schrödinger's computational platform is supported by physics and is changing the way therapies and materials are discovered. The application areas of this material science research and development include petroleum and natural gas, aviation, as well as automotive, consumer goods, and other fields.
If trapped in AI, Schrödinger's ceiling is very limited, but if the label is torn off and it jumps into the pharmaceutical or even manufacturing industry, things may be different
From AI pharmaceuticals to AI manufacturing, it is definitely a broader and more imaginative market. However, after experiencing the burst of this round of bubbles, everyone should be clear that in this field full of hope and change, what should be done now is to give up fantasies, dispel myths, and make choices.
As for the business model, should one choose to make wedding dresses for others, or choose a grander business ideal?
Perhaps, only through round after round of elimination can the initial answer be obtained.
Author: Zhang Xi, Source: Amino Observation, Original Title: &#34;When AI Pharmaceutical Companies Start to Huddle for Warmth&#34;

In the past year, AI pharmaceutical companies have faced challenges with limited clinical trial results. In search of a way out, AI pharmaceutical companies Recursion and Exscientia announced a merger. This is the largest merger in the AI pharmaceutical field, with an estimated merger amount of $688 million. This move is seen as seeking warmth in unity, but the founder of Recursion believes it is a complementary choice. However, the merged Exscientia will no longer exist, becoming a footnote in the AI pharmaceutical field. This indicates that the AI pharmaceutical field still needs time to develop

- AI pharmaceutical companies are facing challenges as the initial hype fades, with Recursion and Exscientia merging to survive the tough industry conditions.
- The merger, valued at $6.88 billion, aims to save costs and extend cash flow, highlighting the struggle for AI pharmaceutical firms to deliver significant clinical data.
- While AI promises innovation, the industry's future profitability remains uncertain, prompting companies to explore diverse business models amidst the burst of the AI pharmaceutical bubble.

When AI pharmaceutical companies start to huddle together