Wen Shuhao helped speed up the development of Pfizers Covid oral drug, Paxlovid. Now the cofounder and chairman of Hong Kong-listed Xtalpi wants to help make better solar panels and EV batteries.In 2016, Wen Shuhao and his team took part in a blind test held by Pfizer to predict the 3D structure of molecules that shows the efficacy of a drug. The Chinese postdoctoral physicist from MIT developed a program that uses AI and quantum physics-based calculations to find suitable structures that are fit for drug making. It turned out to work well. Very well.Wens program achieved a 100% prediction accuracy rate and his then-tiny startup, Xtalpi, was awarded a 10-year partnership with Pfizer in early-stage drug discovery. During the pandemic, the Chinese company helped speed up the development of the U.S. pharmaceutical giants Paxlovid, the U.S.s first approved Covid oral drug.During the Paxlovid development, we truly experienced how our algorithms helped solve problems, recalls Wen, the 43-year-old cofounder and chairman of Xtalpi, in a video interview from the companys headquarters in Shenzhen. It made a significant contribution to alleviating the impact of the Covid pandemic.After working with global pharma giants like Eli Lilly, Johnson & Johnson and Merck to discover new drugs using Xtalpis AI and quantum physics algorithms, Wen thought he could utilize the technology to other industries. He soon found two new industries for his 10-year-old company to disrupt: solar panels and electric vehicle batteries.Wen expects revenue from Xtalpis new materials discovery business, which includes solar panels and EV batteries, to be at least on par with that of the companys drug discovery business within the next three years. Together with potential royalties from Xtalpis 39 ongoing drug discovery programs, he aims to reach Ebitda breakeven as soon as this year.Especially with the materials business, which has a shorter monetization cycle, we will be able to break even faster, says Wen. Once we cross that point, our profit margins will continue to grow because we will earn royalties from the projects we took part in without having to get involved in sales and productions.From left to right: Lai Lipeng, Xtalpi's cofounder and chief innovation officer; Wen Shuhao; Ma Jian, cofounder and CEO.XtalpiIn the first half of 2024, the most recent financial period publicly available, Xtalpi earned almost all of its 102.6 million yuan ($14 million) revenue from AI drug discovery, with the rest from new materials discovery services. The Hong Kong-listed company attributed the 28% growth in sales partly to its first clinical milestone payment, for an undisclosed amount, from Chinese biotech startup Signet Therapeutics; Xtalpi is an early backer of Signet. This came after a drug candidate Xtalpi co-developed for diffuse gastric cancer, a form of stomach cancer, received approval from the U.S. Food and Drug Administration (FDA) to begin human trials.Xtalpis net loss, however, widened 46% to 1.3 billion yuan during the same period. The company blamed it on the significant fair value loss of convertible redeemable preferred shares. Meanwhile, it has continued to invest heavily in research and development, spending 210.4 million yuan in the first half of 2024, though thats roughly 10% less from the previous year.Wen was inspired to dive into drug discovery while helping a cousin in China find cheaper liver cancer drugs. He saw a significant difference in the level of advancement in drug development between the U.S. and his home country, where most cancer patients relied on imported drugs. Wen wanted to use his expertise in using sophisticated math equations to study molecular behaviors and narrow the drug development gap.The stability of a drug molecule and its interactions within the human body are governed by the principles of quantum mechanics, explains Wen. We were very idealistic, believing that, together with AI and robotics, we could enhance the efficiency of drug development and reduce its cost, so that pharma companies could still make a profit while making medication affordable for the general public.In 2015, Wen went back to China and started Xtalpi with MIT postdoctoral fellows Ma Jian and Lai Lipeng. Xtalpi wasand still isone of the few AI drug discovery companies in Asia and quickly garnered investor interest. According to Xtalpis IPO prospectus, filed in 2023, the company said it is the worlds most well-funded startup in the space, raising $732 million in total funding from the likes of Tencent, Chinese venture capital firms Hongshan and 5Y Capital, Google, Susquehanna International Group, Chinas Sino Biopharmaceutical and SoftBanks Vision Fund 2. It went public in a $127 million Hong Kong IPO last June.Xtalpi's robotic lab in Shenzhen.XtalpiXtalpi seeks to distinguish itself with its quantum physics algorithms, which Wen says could increase the accuracy of predicting molecular interactionscrucial to the efficacy and safety of drugs. The quantum physics algorithms calculate the molecular structures, and the resulting datasets are used for machine learning to train its AI models. The AI model then screens through billions of molecules and generates ones that could potentially be made into new drugs. The potential molecules are later tested in Xtalpis robotic labs in Shenzhen and Shanghai, as well as in Somerville, Massachusetts.Unlike peers such as Qiming-backed Insilico Medicine in Hong Kong and Nvidia-backed Recursion Pharmaceuticals in the U.S., Xtalpi doesnt develop its own drugs from start to finish. Instead, it provides proprietary algorithms designed for early-stage drug discovery, which Wen says could speed up the process by about 50%. These algorithms help pharma companies identify molecules that can interact with their drug targetsspecific molecules involved in the disease processto cause therapeutic effects, and optimize the promising molecules so that they are ready for human clinical trials.Xtalpis technology combines quantum physics, AI and robotics so it is one of a kind, says Ted Jing, managing director of 5Y Capital. Xtalpis business model allows it to spend the resources on improving its technology, rather than risking them in human trials. While its reaping a smaller return than the drug makers on each drug, its leading technical infrastructure will continue to attract clients.The stability of a drug molecule and its interactions within the human body are governed by the principles of quantum mechanics,To be sure, there are some doubts over Xtalpis technological capabilities. Derek Lowe, a U.S. pharma veteran who has spent nearly four decades working at several leading drug makers, says calculating how molecules interact to cause therapeutic effects is computationally intensive and difficult to get to work because its a very tricky computation problem.So far no quantum mechanical simulation has been reliable enough to generate data that you would train a machine learning model on, Lowe says. If you were to take a huge pile of 100,000 quantum mechanical estimates of compound binding and feed it into a machine learning model, right now, my belief is you would get a lot of garbage.Over the past four decades or so, scientists have been trying to adopt computational methods to speed up the time-consuming drug development process and improve its success rate. The release of the popular AI chatbot ChatGPT in 2022 sparked renewed hope for machine learning on the medical front. On Monday, Manas AI, a U.S. AI drug discovery startup cofounded by billionaire Reid Hoffman, raised $24.6 million in a seed round. It marked the latest major investment in the space, which saw Advanced Micro Devices inject $20 million into Canadas Absci in early January and Nvidia pour $50 million into Recursion in 2023. The World Health Organization has identified at least 73 AI-derived potential drugs. Some have failed in human trials, others are still testing on patients, but none has hit the market.Lowe adds that even if Xtalpis technology works, the company and its peers are only speeding up the shortest and cheapest part of the entire drug discovery process. Human trials are the hardest part, with an 85% failure rate, because they are based on biology and computational techniques right now cannot help.What Xtalpi and a lot of other companies are doing is really worth something, but its not the incredible revelation of, Oh my God, drug discovery has been solved, says Lowe. If you could compute your way to finding those molecules, that would be terrificthis lets you get up to the shredder, being that 85% failure rate, faster. Now instead of taking six months or two years to get to the shredder, you can get there in just a month or two.Wen thinks that the fact that the 2024 Nobel Prize in chemistry was jointly awarded to the team that developed Alphafold showcases AIs potential on the medical front. AlphaFold is an AI tool co-developed by Google DeepMind scientists Demis Hassabis and John Jumper that can predict the structures of proteins in minutes, which helps scientists better understand how the building blocks of life work and thus open up new avenues for drug development.Wen adds that the pharma industrys AI adoption is a testimony to the technology. All of the worlds top 10 biggest drug makers by market cap have invested in AI through in-house development and startup partnerships. Among them is Eli Lilly from the U.S., which inked a deal worth up to $250 million with Xtapi in 2023 to discover potential medicines for an undisclosed disease. All pharma giants that are working on the most challenging drugs are embracing AI, so why would anyone still have doubts about this? says Wen. AI will become a fundamental tool in the pharma and materials industries. Everyone has to use it; if you dont, you will be less efficient than the others.Xtalpi debuted on the Hong Kong stock exchange in June 2024.Shanshan Kao/Forbes AsiaBefore Xtalpi could help deliver the next life-saving drug, it has ventured into unearthing new materials. Wen first came up with the idea about three years ago when he was looking at how he could expand the application of Xtalpis technology. Quantum physics underpins much of modern materials science, so combining AI screening of massive datasets and robotic testing, Xtalpi could help make better solar panels, batteries, fertilizers and other applications by adding new properties to existing materials.Drugs are the most challenging type of material to make because they involve the complexity of human biology. Making other materials are much simpler, says Wen. Unlike drugs, materials dont require years of clinical trials. Once they achieve the desired effect, they can quickly realize good commercial returns.Wen sees potential in helping commercialize a more energy-efficient solar panel material called perovskite, which scientists have been struggling for two decades to mass produce. By combining perovskite with conventional silicon solar cells, a solar panel can convert up to 43% of sunlight it receives into electricity, versus the current limit of less than 30%, scientists have found. While the process of making perovskite is cheaper and faster, the material is hard to mass-produce because of its low durability and stability.Another area Wen is looking at is helping commercialize the next-generation EV battery. Many major car and battery makers have been racing to mass produce solid-state batteries but no one has succeeded. An upgrade to the current lithium-ion batteries, solid-state batteries have been dubbed the holy grail of batteries because of its promise of more energy storage, faster charging and higher safety for EVs. One of the difficulties of producing solid-state batteries is improving the solid electrolyte component, which is the solid version of the lithium-ion solution used in the current generation of EV batteries.Some companies are convinced that Xtalpis technology could help make the commercialization of these cutting-edge materials possible. In August, GCL Holdings, a Chinese energy company controlled by billionaire Zhu Gongshan, agreed to pay Xtalpi more than $135 million to upgrade materials including perovskite and those for EV batteries. GCLs subsidiary, GCL Technology, is one of Chinas largest solar panel materials makers.If you told me that I have to start a quantum physics-based discovery company, I would immediately jump at materials, says Lowe, who has a PhD in organic chemistry at Duke University. Xtalpi has a better chance of success there because the test cycle is so much faster and cheaper. That drug discovery business is a distraction from materials science.Wen is confident that Xtalpi can excel at both its drug and materials discovery services. Last month, Xtalpi entered into a strategic partnership with Microsoft China, in which they will co-develop AI models for drug and materials discoveries. Every industrial revolution faces resistance, says Wen. Our technology is rare, and we are able to create high value-added super medicine and super materialswe are now in the early-stage of commercial explosion.MORE FROM FORBES