[Recursion Pharmaceuticals]
In its early days, the silicon chip industry was laboratory-based, empirical-based. Since then, the industry has shifted to computer-aided design and in silico simulation. A fundamental belief in the power of accelerated computing and graphics-based processing has guided the company ever since. And now, the company has transformed from video game graphics pioneer to AI juggernaut.
“What lessons can we learn in the biology field from that transition [from empiricism to computer simulation-driven design]?” asked Chris Gibson, Ph.D., CEO of Recursion, in a fireside chat at Download Day, a gathering of scientists, investors, and entrepreneurs focused on accelerating drug discovery. The event, which took place on June 24, 2024, provided a platform for Recursion to share updated pipeline guidance, including seven expected clinical trial readouts, as well as updates on their partnerships and platform advancements.
From pixels to proteins
“There are absolutely parallels,” NVIDIA CEO Jensen Huang said, referring to the transition from empiricism to computer simulation-driven design in the biology field. “My career started 41 years ago, right at the time when computer-aided design started to make progress,” Huang added. The combination of algorithms, compute, and domain expertise led to a new approach. “Know-how was introduced into modern chip design at that time,” Huang said.
Huang sees a similar metamorphosis happening in biology. Three core factors mirror the chip design breakthrough: increasingly powerful deep learning algorithms, access to growing supercomputing capabilities, and the integration of robotics and automation to generate and analyze biological data. But what Huang also underscored wasn’t just the technology itself, but the importance of fostering a culture of bold, yet calculated, risk-taking. As Huang put it, companies need to “believe that in order to discover knowledge… principles and simulations will simply not get us there. You have to be willing to take that leap.”
In 2023, Recursion announced an NVIDIA collaboration that involved a $50 million to support the development of novel foundation models in AI-enabled drug discovery.
Supercomputing meets biology
In modern chip design, the recursive nature extends beyond the iterative process of designing and testing. “You get these ‘Inception’ levels,” he said, referencing the 2010 sci-fi film in which characters could enter multiple layers of dreams, creating dreams within dreams. There are “chips that are creating next-generation chips so that we can design algorithms to design next-generation chips,” Huang explained. “It’s almost like recursion,” he quipped, drawing a parallel to the company hosting the fireside chat.
Recursion’s SEC filing highlights its own recursive approach: “[Recursion] integrates physical and digital components as iterative loops of atoms and bits, scaling wet lab biology and chemistry data organized into virtuous cycles with computational tools to rapidly translate in silico hypotheses into validated insights and novel chemistry.”
This commitment to merging the digital and physical realms of research is exemplified by Recursion’s recent unveiling of BioHive-2, a supercomputer purpose-built to propel drug discovery. As reportedly the fastest supercomputer wholly owned and operated by a pharmaceutical company, BioHive-2 is evidence of Recursion’s belief in the power of AI and sizable computational resources applied to biological complexities.
Recursion’s decision to invest in such cutting-edge technology, typically associated with tech giants and government labs, represents a paradigm shift in the pharmaceutical industry’s approach to drug discovery. “Is it surprising to you that the fastest supercomputer in biopharma is being built and operated by little Recursion and not one of the massive biopharmaceutical companies?” Gibson asked in the interview.
“Tesla is the first car company to build supercomputers. Not a bad company to be compared to,” Huang replied. “If it was obvious to everybody, then it would have already been done. Somebody’s got to go do it.”
Magic lies “somewhere between confidence and insecurity”
While building its computational prowess, Recursion is also forging strategic partnerships. The company has inked deals with firms like Helix, creating a feedback loop where de-identified clinical data and genomic insights continuously refine its AI engine. The ultimate aim: more precise drug target identification.
This ambition, however, comes at a price. Recursion, like many companies aiming to drive new drug discovery approaches, faces a steep climb towards profitability, and its reliance on yet-to-be-proven technology introduces an element of risk that has made some investors wary.
It’s a delicate balance, one that Huang identifies with. “Somewhere between confidence and insecurity… is where we live,” he said. But beyond the financial pressures, Huang emphasizes the importance of a deeper motivation. “If you choose well,” he said, referring to the challenges a company tackles, “not only will you enjoy the journey, you can make a real contribution, and you might be able to live a life of purpose.”
Balancing risk and purpose
For its part, NVIDIA, founded in 1993, has seen dramatic stock increases recently, rising more than 200% from $39.23 on October 31, 2023, to a high of $132.84 on June 14, 2024. But Huang, despite now leading a tech giant, hasn’t forgotten the precariousness of the early days and the importance of long-term thinking.
Huang recounted a company meeting where an employee, noticing a financial detail, asked, “Jensen, are our cash runs red in June?” Huang confirmed, adding, “If we, if we don’t make money before, then we’ll be out of business.” When the employee pressed him on what that meant, Huang stated plainly, “It just means that if we don’t make more money in 30 days, we’ll be out of business.”
Since then, the existential mindset has become part of NVIDIA’s mindset. “If we don’t keep doing amazing things, we don’t get to do this anymore,” Huang said.
Being willing to take the leap
To be successful, companies need to be willing to counter uncertainty with bold action, Huang said. He emphasized that sometimes, established methods aren’t enough: “principles and simulations will simply not get us there.” Instead, companies must have the courage to venture into the unknown. As Huang put it, “You have to be willing to take that leap.”
This willingness to leap, to invest in cutting-edge technology even when the path forward isn’t entirely clear, is what Huang sees as the hallmark of true innovators. “This is a once in a lifetime opportunity for all of you,” he told the Recursion team. “This is a once in a lifetime company and a once in a lifetime circumstance.”
Huang echoed this commitment to a purpose beyond profit, noting that while many investors prioritize short-term gains, “our personality [at NVIDIA] is to go do something nobody’s ever done before. Something that if we didn’t do it, nobody’s going to do it. And if you choose well, not only will you enjoy the journey, you can make a real contribution.”
But with great opportunity comes great responsibility, especially in the field of drug discovery. Gibson emphasized this, highlighting the human element driving Recursion’s work. “In our industry, we get to know some of the patients we hope to one day be able to treat,” Gibson said. “And those patients are waiting. Every day matters.”
Filed Under: Data science, Drug Discovery, machine learning and AI
