New York-based biotech startup Proxima, formerly VantAI, has emerged from stealth with $80 million in seed funding to reshape how medicines influence cells. The round was led by DCVC, with participation from NVentures (NVIDIA’s venture capital arm), Braidwell, Roivant, AIX Ventures, Yosemite, Magnetic Ventures, Alexandria Venture Investments, Modi Ventures, and additional strategic and institutional investors
The fresh capital will fund the expansion of this data platform, deeper machine learning research, and both internal programmes and external partnerships. The company already collaborates with major players, including Johnson & Johnson, Bristol Myers Squibb and Blueprint Medicines, now part of Sanofi.
Moving beyond one-protein thinking
Most drugs are built on a simple premise: identify a faulty protein and shut it down. But biology rarely plays by such clean rules. Diseases often arise from disrupted pathways, not from single offenders.
Founded by Zachary Carpenter and Luca Naef, Proxima’s approach treats proteins as parts of a system. The company focuses on induced proximity, designing small molecules that either pull two proteins together or push them apart. These molecules act like adapters, redirecting cellular workflows rather than stopping them outright. In cases where a function needs to be restored, relocated or enhanced, this strategy offers a more nuanced lever than inhibition alone.
Rather than chasing single targets, Proxima is developing medicines that act as molecular matchmakers or separators inside cells. The idea is to change behaviour by reshaping protein networks, opening paths to diseases that have resisted traditional approaches.
This shift in thinking is why proximity-based medicines have drawn growing interest across the industry. They promise control, not just suppression.
AI as a three-dimensional design tool
At the core of Proxima’s platform is software inspired by the same foundations that power modern image and video generation. Instead of pixels, it works with molecular structures. The system explores vast design spaces to create tiny compounds shaped to fit precisely between proteins.
Drug discovery is a three-dimensional puzzle in which the right piece is needed to stabilise, disrupt, or rewire a protein interaction within a living cell. This approach enables Proxima to explore options that would be impractical to test manually, thereby dramatically expanding the search for viable drug candidates. But clever algorithms alone aren’t enough, and the biggest bottleneck remains data.
Building an advantage through biology’s blind spots
Fewer than 5% of protein–protein interactions are structurally mapped, leaving most of cellular biology poorly charted. Proxima is attacking this gap head-on with NeoLink, its structural proteomics technology designed to generate new interaction data at scale.
For investors, the appeal lies in what one partner called an unfair data advantage, a foundation that could make protein interactions programmable. For Proxima, the real proof lies in translating this data-first vision into medicines that work not just on screens, but in living systems.
“Proximity-based medicines represent one of the most powerful new ways to treat disease, but progress has been constrained by a lack of structural data and accurate design tools,” said Zachary Carpenter, Co-Founder and CEO, Proxima. “By combining proteome-scale structural data with frontier AI models, we’re building the foundation needed to make these therapies broadly accessible rather than rare exceptions.”
“Proximity-based therapeutics represent one of the most promising frontiers in modern drug discovery with the potential to treat previously intractable diseases and target ‘undruggable’ proteins,” said Jason Pontin, General Partner at DCVC, who will be joining the company’s board. “Proxima’s technology combines proteome-scale structural data with state-of-the-art generative AI foundation models, and the company’s team is uniquely well-positioned to discover and develop a new class of medicines.”
