Sahir Raoof is not new to the world of biotech investing. Before founding Proximal Ventures in 2026, he was the first commercial hire and VP of Business Development at Iterative Health, built biopharma research partnerships at Flatiron Health — which was acquired by Roche — and covered biotech and outsourced pharma services at J.P. Morgan. That trajectory, spanning venture investment, operational roles and investment banking in healthcare, is precisely what he is now deploying in Proximal's first fund: a $27.5 million fund, above its initial $25 million target, dedicated to startups at the intersection of artificial intelligence and the life sciences.
Announced on November 10, the Proximal Ventures Fund I focuses on early-stage — pre-seed and seed — investments, with typical checks between $500,000 and $1.5 million per company. Unlike many generalist funds that position themselves as "AI-first" but end up spreading capital across diverse sectors, Proximal is intentionally hyper-specialized. The fund covers the entire pharmaceutical value chain: from computational molecule discovery through clinical development, manufacturing and market access. The thesis is straightforward but ambitious: AI can accelerate every stage of the drug development process, and the sector needs directed capital to build the infrastructure that will support that transformation.
Since closing the fund, Proximal has already deployed capital into six companies. The most visible is ChronicleBio, co-founded by Fidji Simo — a former OpenAI executive who left the general-purpose AI company to pursue a more concrete path at the intersection of data and biology. The portfolio, however, is only beginning to take shape. Raoof told Scrip the intention is to invest in about 20 startups over the life of the fund, focusing on AI-enabled platforms that can shorten development timelines across all phases.
The fund's governance structure is another indicator of what Proximal aims to offer. Startup founders receive not just capital — they gain access to an advisory board comprising figures like Amy Abernethy, former Principal Deputy Commissioner at the FDA, and Kim Branson, Chief AI Officer at GSK. The logic is not merely to connect startups to investors, but to anticipate regulatory hurdles and validate business models with pharma executives before companies are even incorporated.
But the macro context is as important as the fund's thesis. The AI drug discovery market grew from approximately $1.54 billion in disclosed funding in 2024 to $1.84 billion in 2025, and already exceeded $2.4 billion in the first half of 2026. The problem is that this growth is misleading: Isomorphic Labs, DeepMind/Alphabet's AI drug discovery arm, accounted for roughly 87% of total early-stage funding in 2026. This means that while aggregate numbers are impressive, most smaller funds and very early-stage startups continue to struggle to raise capital.
This extreme concentration of capital explains much of Proximal's strategy. Raoof explicitly acknowledges that "some of the greatest opportunities lie in areas that have historically attracted less venture capital investment" — meaning not molecule discovery itself, but everything that follows: clinical translation, regulatory navigation, manufacturing optimization and commercialization. It is a market thesis that many early-stage investors have overlooked, believing innovation would remain concentrated in the pure AI models.
Pharma industry adoption data supports this conviction. AI partnerships for drug discovery jumped from 84 in 2024 to 114 in 2025, a 36% increase, and the potential deal value leapt from $11.8 billion to $43.4 billion over the same period. Lilly and Nvidia announced a joint commitment of up to $1 billion over five years for an AI drug discovery laboratory combining researchers, compute, robotics and automated experimentation. Schrödinger, one of the most established computational discovery tool platforms, has seen steady software demand.
What Proximal is trying to solve, in this context, is the so-called "valley of death" in drug development. AI can increasingly predict molecular properties, optimize potential candidates and reduce time to first clinical trials — but the distance between a good candidate in silico and an clinically approved drug remains enormous. The Phase I success rate for AI-discovered molecules is estimated at 80% to 90%, compared to a typical 3% success rate for all drugs reaching Phase III. The bottleneck is no longer initial discovery, but translation: biological target validation, clinical trial design, preclinical data interpretation and, crucially, regulatory navigation.
There is also a risk Proximal cannot avoid: the restructuring that affected BenevolentAI, which halted investment in its BEN-2293 program after the drug met safety and tolerability objectives in a Phase IIa atopic dermatitis study but failed to produce sufficient efficacy. The case illustrates a recurring pattern in the field — AI can model physicochemical and safety properties with increasing accuracy, but the underlying biological complexity, especially in novel mechanisms of action, still challenges even the best models.
What Proximal is doing, in practice, is betting that the next wave of innovation in AI and healthcare will not be led by molecule discovery platforms, but by companies that build the infrastructure tools connecting discovery to the patient. It is a bet that venture capital will finally look beyond what is sexy and toward what is necessary.
Whether a $27.5 million fund will be enough to prove this thesis right — or whether the market will need greater scale to be convinced — remains to be seen.
Sources: Business Wire, Citeline / Scrip, New Market Pitch
✓ Independent sources cross-checked and verified before publishing
