[KongchangAI]
· 2 min read· 1,059 words

Proximal Ventures Closes $27.5M Debut Fund to Bet on AI Drug Discovery

Proximal Ventures Closes $27.5M Debut Fund to Bet on AI Drug Discovery

Former SignalFire investor founds Proximal Ventures, raising $27.5M for early-stage AI drug discovery.

Proximal Ventures, founded by former SignalFire investor Sahir Raoof, has closed a $27.5M debut fund focused on startups applying AI to drug R&D. The article analyzes the industry logic—Eroom's Law, AI's role in target ID, molecule generation, and ADMET prediction—and argues vertical funds need post-investment resources like wet labs and pharma BD to differentiate. The fundraise signals the early-stage window remains open, though the fund's capabilities await validation by its first deals.

The AI drug discovery space has a new player. According to an RSS source, Proximal Ventures, founded by former SignalFire investor Sahir Raoof, has closed its debut fund at $27.5 million, with an explicit focus on startups applying artificial intelligence to the drug R&D process.

Proximal Ventures debut fund announcement

Fund Background and Leadership

Proximal Ventures founder Sahir Raoof previously worked as an investor at SignalFire. SignalFire is known for its data-driven investment methodology, and that experience likely shaped Raoof's judgment on the intersection of AI and life sciences. The new fund's name, "Proximal," hints that its investment strategy may lean toward projects close to the core of drug discovery rather than peripheral tooling layers—though the specific portfolio has not been disclosed, so this reading remains to be verified.

At $27.5 million, the debut fund is a typical size for an early-stage fund in the biotech-AI crossover space, usually corresponding to seed to Series A checks.

Why AI Drug Discovery

Drug R&D has long been trapped by "Eroom's Law"—long cycles, high costs, and high failure rates. Traditional small-molecule drug discovery often requires screening vast chemical spaces, and preclinical research is enormously expensive with limited success rates. AI can intervene across target identification, molecule generation, ADMET property prediction, clinical trial design optimization, and more, theoretically compressing early discovery timelines and reducing trial-and-error costs.

Over the past few years, this direction has attracted a wave of specialized funds and strategic investments from large pharma companies. Proximal Ventures' entry shows that capital is still willing to bet on vertical AI applications at the early stage, rather than generalized AI platforms.

Eroom's Law is an empirical observation in the pharmaceutical industry: since the 1950s, the number of new drugs approved per billion dollars of R&D spending has roughly halved every nine years—the exact opposite of Moore's Law in semiconductors. This reflects the dual predicament of drug R&D: on one hand, easily discovered targets and molecules have already been claimed by pioneers, making the remaining opportunities progressively harder to find; on the other hand, regulatory requirements for safety and efficacy keep rising, and clinical trial scale and costs continue to inflate. Estimates suggest that a new drug takes an average of 10-15 years and over $1 billion from target discovery to market, with clinical-stage failure rates as high as around 90%. AI is seen as a potential breakthrough precisely because it can replace some wet-lab experiments with computation in areas like compound screening and toxicity prediction, thereby compressing trial-and-error costs and timelines.

ADMET is the acronym for five key dimensions of drug-likeness evaluation for candidate molecules in drug R&D: Absorption, Distribution, Metabolism, Excretion, and Toxicity. Even if a molecule shows strong activity against a target in vitro, it cannot become a drug if it cannot be effectively absorbed by the human body, distributed to the site of action in vivo, or metabolized and cleared at a reasonable rate—or if it produces serious toxic side effects. Traditionally, ADMET assessment relies on animal experiments and in vitro cell experiments, which are time-consuming, costly, and limited in throughput. AI models can learn structure-property relationships from existing compound databases and predict the ADMET performance of candidate molecules in silico, thereby eliminating large numbers of unqualified molecules before synthesis and experimentation, significantly improving early screening efficiency.

Differentiation Space for Vertical Funds

In AI drug discovery, generalist VCs and industry capital each have advantages: the former provides capital and networks, the latter provides pipelines and data. For a $27.5 million vertical fund to establish differentiation, it typically needs deep post-investment involvement—for example, helping AI teams connect with wet-lab resources, pharma BD channels, or regulatory pathway experience.

Whether Raoof's experience at SignalFire accumulated sufficient biomedical industry resources is not well documented in public information. This will be a key variable in evaluating the fund's subsequent performance.

Wet lab refers to a physical experimental space for conducting real biochemical experiments, as opposed to a dry lab (computational simulation), equipped with cell culture, molecular biology, analytical testing equipment and consumables. Although AI drug discovery companies excel in algorithms and compute, their model predictions must ultimately be validated through wet-lab experiments—synthesizing predicted molecules and testing their activity and toxicity. Building wet-lab capabilities takes a long time and significant capital investment, posing a notable barrier for early-stage AI teams. Therefore, vertical funds that can help portfolio companies access shared labs, CRO (contract research organization) resources, or pharma companies' internal experimental platforms have substantive value in post-investment support—an important dimension for evaluating the differentiation capabilities of such funds.

Signals Worth Watching

This fundraise itself offers several angles for observation: first, the early-stage financing window for AI drug discovery has not closed, and specialized new funds can still complete a first close; second, investors leaving large multi-stage funds to strike out on their own reflects that vertical deep-dive strategies are being adopted by more practitioners; third, the $27.5 million size means the portfolio will be relatively concentrated, making the quality of individual investment selections more important than coverage breadth.

For AI drug discovery founders, the emergence of Proximal Ventures means one more potential option for early-stage financing, but the fund's specific investment themes, stage preferences, and post-investment capabilities will need to be validated by its first batch of investment cases.

One-line Summary

Former SignalFire investor founds Proximal Ventures, raising $27.5M to focus on early-stage AI drug discovery investments.

Paragraph Summary

Proximal Ventures, founded by former SignalFire investor Sahir Raoof, has closed its debut fund at $27.5 million, with an explicit focus on startups applying AI to drug R&D. The article analyzes the industry logic of AI drug discovery: drug R&D is trapped by Eroom's Law—long cycles, high costs, and high failure rates—and AI can compress early discovery timelines and trial-and-error costs in areas such as target identification, molecule generation, and ADMET prediction. The $27.5 million size corresponds to the seed to Series A stage, and vertical funds need post-investment resources (wet labs, pharma BD, regulatory experience) to build differentiation. The article sees several signals in this fundraise: the early-stage financing window for AI drug discovery has not closed; investors leaving large multi-stage funds to strike out on their own reflects adoption of vertical deep-dive strategies; portfolio concentration means individual investment quality matters more than coverage breadth; and the fund's actual capabilities still need to be validated by its first investment cases.

#Proximal Ventures#AI drug discovery#Sahir Raoof#SignalFire#biotech venture fund#Eroom's Law#ADMET
Share:

Related articles