When a startup claims it wants to stop scientists from “cooking” in the lab and start engineering results instead, it usually sounds like standard pitch deck hubris. But when that startup is led by a Gordon Bell Prize winner and backed by a roster of state-affiliated heavyweights, the claim carries a different kind of weight.
Beijing-based DP Technology has closed a Series C round worth approximately $114 million (1.35 billion RMB). The fresh capital isn’t just a runway extension; it is a clear indicator that China views “AI-for-science” not merely as a software vertical, but as critical industrial infrastructure comparable to semiconductor manufacturing or cloud computing.
The funding round was led by Fortune Venture Capital and the Beijing Jingguorui Equity Investment Fund, a mix that underscores the company’s alignment with national strategic interests in energy and biotech.
Moving logic from the wet lab to the GPU
Founded in 2018, DP Technology operates on a frustration common to anyone who has worked in material sciences or pharmaceuticals: the process is painfully slow. Traditional R&D often relies on intuition and trial-and-error—a method founder and CEO Weijie Sun refers to as “alchemy.”
Sun’s pitch is straightforward: replace the intuition with calculation.
Without a good computing or AI platform, everyone relied on trial-and-error ‘cooking’ in the lab. We want to turn that alchemy into an engineered, predictable process.
The company’s technical pedigree is rooted in the work of co-founder and Chief Scientist Zhang Linfeng. In 2020, Zhang’s work on machine learning algorithms for molecular dynamics won the ACM Gordon Bell Prize, the highest honor in high-performance computing. The core thesis was that by pre-training large models on the fundamental laws of physics and chemistest rather than just text, researchers could simulate experiments virtually with high fidelity.
This led to the creation of the “Particle Universe,” a suite of pre-trained models that predict how molecules and materials behave. Instead of burning cash on physical experiments that fail, researchers can filter candidates digitally.
The Operating System for R&D
While many AI startups focus on a single point solution—generating a protein structure, for instance—DP Technology has built something resembling a full-stack operating system for scientific R&D. The company claims to serve over 1,000 universities and roughly 150 corporate clients, including heavy hitters like PetroChina and EV giant BYD.
The platform has expanded into specific verticals:
- Hermite: For computer-aided drug design.
- Piloteye: For battery R&D, a critical sector for China’s EV dominance.
- RiDYMO: Tarreceiveing “undruggable” tarreceives in immunology and oncology.
- Bohrium & SciMaster: Cloud infrastructure and AI agents to manage workflows.
The strategy is sticky. If an industrial R&D team builds their workflow around DP’s OS, the startup becomes a vfinishor that is nearly impossible to rip out.
Once your models sit in the inner loop of a customer’s R&D workflow, you’re no longer just a vfinishor. You’re infrastructure.
A ‘National Team’ Cap Table
The composition of the Series C investors informs a story of its own. This is not a Silicon Valley cap table; it is a “China Inc.” lineup. Alongside the leads, participants included the Beijing Artificial Ininformigence Industest Investment Fund, the Beijing Pharmaceutical and Health Industest Investment Fund, Lenovo Capital, and Oriza Hua.
This structure suggests that DP Technology is positioning itself as a national champion. In an era of decoupling, having state-backed capital provides a buffer against market volatility and opens doors to massive state-level projects in energy and healthcare.
The company has raised over $240 million to date, with earlier backing from prominent firms like Qiming Venture Partners, Source Code Capital, and MSA Capital.
Scaling through the Hardware Crunch
DP Technology faces a unique set of constraints. While it has achieved product-market fit—evidenced by the acquire-in from industrial giants—it must operate under the shadow of US export controls on high-finish GPUs. Developing and running massive molecular models requires serious compute power.
However, the company’s focus on domestic infrastructure may be its greatest defense. By embedding itself into the workflows of China’s top energy and pharma companies, it secures data loops that Western competitors cannot access. As clients run experiments, that data feeds back into the models, theoretically improving accuracy in a flywheel effect.
With $114 million in the bank, the company plans to scale its team of applied scientists and deepen its penetration into the industrial sector. The challenge now is shifting from academic prestige to commercial dominance, proving that AI can actually reduce the cost and time of bringing a new drug or battery to market—not just simulate it.
Leave a Reply