Reid Hoffman, Partner at Greylock and co-founder LinkedIn, speaks during the WSJ Tech Live conference hosted by the Wall Street Journal at the Montage Laguna Beach in Laguna Beach, California, on October 21, 2024.
Frederic J. Brown | Afp | Getty Images
LinkedIn co-founder and venture capitalist Reid Hoffman became a billionaire from his business social-working company, and has made lucrative bets on companies including Airbnb and Zynga while also backing nuclear fusion startup Helion Energy.
Now Hoffman is diving into the health care, which he describes as “wondrous and terrifying,” with his latest startup, Manas AI.
Hoffman and Dr. Siddhartha Mukherjee, an oncologist and Pulitzer Prize-winning author, unveiled the company on Monday. Manas will use artificial intelligence to try and accelerate the drug discovery process, starting with new treatments for aggressive cancers like prostate cancer, lymphoma and triple-negative breast cancer.
Developing new drugs is traditionally a costly and complex process. It can take more than 10 years and cost billions of dollars to develop a single medication, according to a report from Deloitte. Manas said it will use its proprietary chemical libraries and AI-powered filters to identify drug candidates more quickly, ideally reducing the decades-long discovery process to just a few years.
“Most people have had friends, family members, etc., who’ve died from cancer or had serious cancer problems,” Hoffman told CNBC in an interview this week. “If we can make a huge difference on this, and this is the kind of thing that AI can make a huge difference in, it’s the kind of reason why AI can be great for humanity.”
Manas raised $24.6 million in seed funding, led by General Catalyst and Hoffman with participation from Greylock, where he is a partner. Hoffman has been deep in AI in recent years. He was an early investor in OpenAI, when the project was still a nonprofit, and he helped start Inflection AI along with DeepMind co-founder Mustafa Suleyman. Last year, Suleyman joined Microsoft, where Hoffman is a board member, as CEO of a new unit called Microsoft AI. Several Inflection employees joined him.
Manas has also inked a partnership with Microsoft, and will leverage its Azure cloud-computing platform. Hoffman, who sold LinkedIn to Microsoft for $27 billion, said Manas is deploying several additional tools from Microsoft as well, including some that are not generally available to the public yet.
Hoffman has been working with Mukherjee to create Manas for about a year, though the process picked up steam in the last couple months. Hoffman said the team felt ready to publicly share its ambitions this week since its baseline, foundational resources are in order.
‘Totally delighted’ to see competition
The company has a long road ahead, and the drug discovery market is very competitive. Other startups along with major pharmaceutical companies like Eli Lilly, Pfizer and Merck, are also exploring how to leverage AI to accelerate drug research and development.
Hoffman said he feels confident in Manas’s approach, though he would be “totally delighted” to see multiple companies flourish.
“We also bring the thing that a startup usually brings, which is a willingness to go very hard, abandon things quickly that aren’t working,” he said. “Live like this week matters, and the result of this week matters.”
Following Manas’s launch on Monday, five different potential strategic partners have already approached the company, Hoffman said.
Hoffman said the company is in “build quickly” and “learn and deploy” mode. One of its early initiatives is called Project Cosmos, which is an effort to map out the fundamental rules of drug binding, according to the company’s website. Hoffman declined to share any additional details about the project.
Manas currently has just four employees – including Hoffman and Mukherjee – but Hoffman said it will grow. He’s been acting as the company’s “AI guy” while Mukherjee serves as the “bio guy,” he said. Ultimately, Manas is about melding the two fields.
“It isn’t just the best of science and it isn’t just the best of AI, because either of those two are insufficient,” Hoffman said. “You need to put those two together.”
As the AI guy, Hoffman was paying close attention this week to the sudden emergence of China’s DeepSeek in the U.S.
DeepSeek began generating buzz in January, when the startup released its open-source reasoning model R1, which rivals OpenAI’s o1. The model was reportedly developed at a fraction of the cost of rival models by OpenAI, Anthropic, Google and others.
Hoffman said that while DeepSeek might encourage American companies to pick up the pace and share their plans sooner, the new revelations don’t suggest that large models are a bad investment.
“The competition game is on,” he said, “But I don’t think it’s the ‘Oh my God, we’re losing!’ as American technology.”
