Photo credit: www.cnbc.com
Demis Hassabis, the Co-Founder and CEO of Google DeepMind, and recipient of the 2024 Nobel Prize in Chemistry, participated in the Artificial Intelligence (AI) Action Summit held at the Grand Palais in Paris, France, on February 10, 2025.
Benoit Tessier | Reuters
On a sunny day in London, Colin Murdoch received an unexpected bank deposit from his employer—not his salary, but rather an investment from Google in his startup.
This new venture, Isomorphic Labs, emerged as a life sciences startup spun out of Google DeepMind. This development followed a significant biological breakthrough that was considered one of the most important in the past half-century. The investment occurred in November 2021, and Murdoch was a member of the founding team.
In his role as chief business officer at DeepMind, Murdoch aimed to identify practical applications for the AI innovations cultivated by the organization. With Isomorphic, he sought to focus specifically on the potential of AlphaFold2, an AI technology capable of predicting protein structures.
This groundbreaking program, created by DeepMind in 2020, tackled a long-standing challenge that had perplexed biologists for decades. If utilized effectively, it could lead to the discovery of new treatments for serious conditions such as Alzheimer’s, Parkinson’s, and cancer. Its significance was underscored when it received a Nobel Prize for chemistry.
Moreover, the insights gained from AlphaFold2 helped Murdoch and Isomorphic’s creator, Demis Hassabis, to establish a bold vision: leveraging AI for comprehensive drug discovery aimed at “solving all disease.”
“When AlphaFold emerged, Demis and I realized its potential. It was a clear indicator that we could apply AI in drug discovery,” Murdoch, now president of Isomorphic, stated during a video interview with CNBC. “This moment indicated that we could indeed start making strides in pharmaceutical development.”
Although Hassabis continues to lead Isomorphic as CEO, he also manages DeepMind, which plays a crucial role in Google’s strategy for competing in the fast-evolving AI landscape. Even though Isomorphic operates as an independent entity within Alphabet, Hassabis’s dual positions illustrate the close ties between the two organizations.
Isomorphic has not disclosed the precise amount of Google’s initial investment from 2021, but recently, the company successfully secured its first round of external funding at $600 million, led by Thrive Capital, along with contributions from GV (formerly Google Ventures) and additional support from Alphabet. This influx of cash will enable Isomorphic to venture into biologics and enhance its capabilities with improved computational resources for AI applications.
The drug discovery industry is anticipated to reach a value of $71 billion by 2025, as AI continues to revolutionize the sector.
Typically, it can take more than a decade—sometimes reaching costs of over $2 billion—for a drug to progress from ideation to FDA approval. Additionally, the costs associated with clinical trials for new drugs can reach hundreds of millions, often resulting in lost investments for unsuccessful trials. However, applying AI to drug discovery could streamline this process, ultimately reducing expenses and increasing the likelihood of successful products reaching the market.
Since its inception, Isomorphic has expanded to over 200 employees and formed partnerships with two of the largest pharmaceutical companies, Eli Lilly and Novartis. These companies engage Isomorphic to ensure their current investments yield successful outcomes in the long run. The partnerships, which include upfront and milestone payments, could potentially be valued at nearly $3 billion for Isomorphic, excluding royalties from future drug sales.
Recently, Google DeepMind and Isomorphic made AlphaFold3 publicly accessible for non-commercial academic use. This updated system can “predict the structures of and interactions between all life’s molecules,” as stated by Isomorphic. The enhanced capabilities extend beyond protein predictions to include DNA, RNA, ligands, and more, significantly contributing to advancements in drug discovery and our overall biological understanding.
Harnessing AI for Health Solutions
Murdoch, now 46, once aspired to be a doctor but shifted his focus to technology and engineering due to his discomfort with medical settings. However, his interest in medicine was reignited through DeepMind’s protein structure research.
The introduction of AlphaFold2 in 2020 marked a pivotal advancement in life sciences, with significant implications for understanding proteins, the fundamental components of cellular structure that conduct numerous life-sustaining chemical reactions. The shape a protein adopts determines its function, so the ability to accurately predict protein structures is valuable in addressing numerous diseases linked to protein misfolding.
Following the launch of AlphaFold2, Murdoch and Hassabis assembled a team from DeepMind to identify potential applications for the AI system in drug development. After months of exploration, they pinpointed areas where AI could enhance drug discovery and presented their business model for Isomorphic to Alphabet.
Alphabet responded positively, recognizing the idea’s potential.
Subsequently, Alphabet facilitated the spin-off of Isomorphic from DeepMind, assigning it the mission of harnessing the AlphaFold technology to drive biomedical breakthroughs and advance drug design initiatives.
At the start of 2022, Isomorphic began with just over a dozen individuals working in a temporary office in King’s Cross, a tech-centric location in London.
Since then, the company has operated largely under the radar, choosing to remain discreet about its advancements and progress. Over the past few years, representatives have indicated that Isomorphic was not yet prepared to disclose updates and that leadership preferred to avoid media engagements while they refined their technology.
Not all employees from DeepMind who participated in Isomorphic’s founding transitioned to the independent company. Therefore, Hassabis and Murdoch had to recruit a new team, starting with key executive roles and subsequently hiring specialists in therapeutic areas and machine learning.
“We’ve invested tremendous energy into fostering an interdisciplinary culture that blends these diverse forms of expertise,” Murdoch emphasized. “This combination of AI proficiency and drug discovery knowledge is crucial for our success.”
Demis Hassabis: A Pioneering Innovator
At 48, Hassabis commands significant influence within Google. He co-founded DeepMind in 2010 with a vision of developing “general-purpose” AI, later selling the startup to Google in 2014 for an estimated $400 million to $650 million. DeepMind has been pivotal in Google’s AI research efforts.
The landscape became more competitive in late 2022 with the emergence of OpenAI’s ChatGPT, propelling Google to accelerate its AI initiatives.
Since that shift, the relevance of Hassabis and DeepMind has grown. In April 2023, Google integrated DeepMind with Google Brain, another of its AI research frameworks, forming Google DeepMind, and Hassabis was appointed to lead the merged unit. His influence grew further in October when responsibility for the Gemini app, encompassing Google’s direct-to-consumer AI tools, was placed under the Google DeepMind jurisdiction.
Through these profound changes, Hassabis has successfully maintained his role at Isomorphic.
“The AI talent rooted in this organization, under Demis’s leadership, possesses high-caliber research skills,” noted Thrive Capital’s Vince Hankes. “Importantly, many of these individuals have the option to work in any AI lab; their choice to contribute to a bio-focused endeavor is noteworthy.”
People who have known Hassabis for years regard him with a mix of admiration, respect, and a sense of astonishment at his achievements. His profile has been enhanced as a chess prodigy, a Nobel laureate, and, as of 2023, a knight—Sir Demis Hassabis.
The Nobel laureate in Chemistry received his award from Sweden’s King Carl Gustaf during the ceremony held on December 10, 2024.
Pontus Lundahl | TT | Via Reuters
Krishna Yeshwant, a managing partner at GV, recounted that meeting Hassabis at Oxford University felt like encountering a celebrity. Even then, Hassabis’s remarkable accomplishments were well-known, yet he remained approachable, communicating about his AI work without technical jargon.
“Even before the acquisition of DeepMind, he was already contemplating breakthrough possibilities,” Yeshwant reflected.
GV’s investment in Isomorphic was driven by the unique capabilities of the startup in both AI and therapeutics. Yeshwant pointed out, “In biotech, you often conduct experiments that yield various results, and each success can advance human knowledge significantly. In tech, measurements often revolve around customer acquisition and feature development.” He added, “It’s rare to find an individual who adopts this experimental mindset within technology, but Demis has been doing so for quite some time.”
Hassabis expressed to CNBC his belief in the critical application of AI for improving human health, stating, “We possess an extraordinary tool that could finally reveal answers to some of the most perplexing health challenges we face and empower us to overcome them. The prospect of a healthier future for individuals is what truly motivates me.”
Future Directions for Isomorphic Labs
As Isomorphic begins to reveal more about its operations, the startup plans to utilize its inaugural round of external funding to expand into the biologics sector.
Currently, Isomorphic is adept at leveraging its general-purpose technology to rapidly design small-molecule drugs—think oral medications. However, the exploration into biologics will enable the startup to target a broader spectrum of diseases since biologics generally require injection rather than oral administration, as noted by Murdoch.
“Our primary areas of focus are cancer and immunology, through our internal programs. This will enable us to accelerate our initiatives and introduce more programs,” he explained.
Additionally, Isomorphic aims to explore new applications for AlphaFold3.
The newly acquired funding will also be allocated towards computational resources, a significant challenge in the AI realm, where access to graphics processing units (GPUs) is paramount. Isomorphic aims to build models that function at larger scales with enhanced data integration, both in terms of AI capabilities and therapeutic developments.
“Establishing a comprehensive life sciences company requires significant investment—particularly in computing, talent acquisition, data collection, and asset development,” Hankes from Thrive remarked.
Source
www.cnbc.com