Propelled by ‘science for humanity,’ this Chinese AI startup sets sight on US – TechCrunch

Image Credits: DP Technology

Amid rising geopolitical tensions, many Chinese tech companies find themselves recalibrating their overseas pursuits, often sidestepping any reference to their origin. One bold startup called DP Technology stands out from the crowd. Working to apply artificial intelligence to molecular simulations, DP, short for “Deep Potential”, believes that the unifying power of “scientific research for humanity” will pave the way for its global expansion.

Founded in 2018 with renowned mathematician Weinan E as its advisor, DP provides a set of tools to conduct scientific computing, a process in which “computer simulations of mathematical models play an indispensable role in the development of technology and in scientific research,” according to a definition by the University of Waterloo. Areas that can benefit from scientific computing range from biopharmaceutical research and car design to semiconductor development.

While the world is currently fixated on using AI to generate text, images and videos, DP finds itself in a less-tapped field: combining machine learning, which allows computers to automatically learn from given data, with molecular simulations, which analyze real-world products and systems through virtual models. When applied together, machine learning can improve the speed and accuracy of simulations to solve problems in the physical world.

“In the past, without a good computing or AI platform, everyone relied on experience-based trial and error. That process was often referred to as ‘cooking’ or ‘alchemy’,” DP’s CEO and founder Sun Weijie told TechCrunch in an interview.

“This approach was relatively effective in the early stages of industrial development because user expectations for iteration weren’t that high, but now there is a growing demand for [technological] advancements,” he continued. “For example, consumers expect an increase in battery capacity each year and anticipate better performance from each new generation of vehicles. The traditional R&D model is no longer able to sustain these rapid market changes.”

“A breakthrough in the research and development approach is necessary to keep up with these expectations of rapid iterations,” he added.

To that end, DP has devised a suite of software for industry players to discover and develop new products more efficiently. For one, it runs a scientific computing platform that enables simulations of physical properties such as magnetism, optics and electricity; the results of running these models in turn allow materials like semiconductors and batteries to be designed in a faster and cheaper way. It also operates a SaaS platform specifically for preclinical studies on drug discovery.

Aside from supplying software to industrial researchers and designers, DP goes a step further by selling services tailored to their needs and carrying out R&D processes for its customers who might not otherwise fully leverage the potential of its tools.

This mix of SaaS and service business models has proven some initial success in China. In 2023, DP is expected to rack up nearly 100 million yuan ($14 million) worth of contracts, up from “tens of millions of yuan” last year. Now it’s gearing up to take that strategy to Western markets where the field is dominated by deep-pocketed giants like DeepMind.

“There’s an old saying in China: the children from the poor become mature early. With much less funding at hand, we are the poor kids compared to the likes of DeepMind and OpenAI,” Sun said.

To date, DP has raised around $140 million from a lineup of top Chinese VC firms such as Qiming Venture Partners and Hillhouse Ventures. For some comparison, 13-year-old DeepMind was bought by Google for over $500 million back in 2014. The London-based AI powerhouse reported £44 million ($60 million) profit in 2020, up from a whopping £477 million ($650 million) loss in 2019.

Sun asserted that DP, despite its physical headquarters in Beijing, was conceived with a global mindset owing to an open-source scientific computing community it founded, DeepModeling. Its early anchor in China was also more accidental than deliberate. “The COVID-19 pandemic put a stop to international exchange, so we decided to just stay put and work on monetization [in China] the first two years,” said Sun.

DP’s international expansion is starting with the U.S., where it will open an office and work with a partner to distribute its products and services. Looking to establish a presence in the new market, the startup looks to ramp up its reputation by leveraging its open-source community and attending trade shows in what Sun described as a relatively “close-knit circle” of basic research.

In the meantime, DP’s international ambitions might encounter roadblocks from the ongoing decoupling that’s dividing the U.S. and China across many areas, including scientific research. Back in August, for instance, the Biden administration narrowly extended a science partnership that had underpinned U.S.-China relations since 1979.

Sun, however, exuded confidence in science’s resilience in the face of geopolitical complications. “Both the fields of basic science and biopharmaceuticals are shared by all of humanity, and they are relatively open and inclusive. Comparatively speaking, I believe that these areas will be fine,” he said.