AI Drug Discovery: From Fee-for-Service to Pipeline Equity Goldmine

The whole business model for AI-driven drug discovery is changing. It’s moving fast from a simple fee-for-service setup to a model built on pipeline equity and milestone payments. For biotech VCs and life science crossover investors, you have to get this shift if you want to accurately value these companies and find the next high-growth plays. The industry is maturing, and the limits of just selling software are obvious. AI platforms now need to be co-creators of therapeutic assets, not just computational vendors.

The Inherent Limitations of the Pure SaaS Model in Drug Discovery

For a long time, many AI drug discovery platforms just ran on a software-as-a-service (SaaS) or fee-for-service basis. A pharma company would hire an AI firm for a specific job: maybe target ID, hit generation, lead optimization, or even designing a molecule from scratch (de novo). This gave the AI companies a predictable, if small, revenue stream, but it completely capped their upside. They were spending a ton on R&D to build their algorithms and compute power, but their compensation was disconnected from the actual billion-dollar potential of a drug. The incentives were totally misaligned. Success for the AI firm was measured in contract renewals, not in whether a molecule they helped find actually worked in the clinic and made money. What’s worse, the pure SaaS model made them look like simple tool providers, easily commoditized or replaced once Big Pharma built its own in-house AI teams. They weren’t even fully using their own critical “data moat” to build proprietary assets. Because of this setup, valuations got stuck on software multiples. They missed out on the much higher biopharma multiples that come with valuable intellectual property and clinical wins.

From Service Provider to Pipeline Partner: Recursion and Insilico’s Strategic Pivot

The smartest companies in AI drug discovery see this problem and are changing their models to get a bigger piece of the pie. They’re becoming real pipeline partners, sharing the risks and the rewards that come with getting a drug through clinical translation. This shift means a bigger focus on milestone payments, royalties, and direct equity in the drugs themselves. Recursion Pharmaceuticals is a perfect example. They were first known for their massive biological dataset and AI-powered phenotypic screening, but they’ve aggressively signed partnerships that are heavy on milestones. Their deal with Roche and Genentech is a monster, potentially worth up to $12 billion with its mix of upfront cash, research funding, and huge clinical and commercial milestones across neuroscience and oncology programs. This deal structure aligns Recursion’s money with the actual success of the drugs coming off its platform. Of course, scaling these kinds of partnerships is expensive, which is why Nvidia just dropped a $50 million investment into the company to accelerate its computing infrastructure. Nvidia’s check shows everyone that AI drug discovery is capital-intensive and requires serious computing power to build these pipelines. Insilico Medicine is another great case study. As an AI-native company, they’ve shown they’re all-in on co-development by pushing their own AI-designed molecules into the clinic. Their lead program, INS018_055 (now called Rentosertib), is an antifibrotic molecule for idiopathic pulmonary fibrosis (IPF) that was designed by their AI and has already reached Phase 3 trials. Going from purely in silico work to running their own clinical trials is what really sets them apart. It makes Insilico a biopharma company with its own proprietary pipeline. By shouldering the enormous risks and costs of clinical development (think IND-enabling studies, manufacturing, and the trials themselves, all while working through tough FDA IND approval requirements), Insilico stands to see a much, much bigger payday if Rentosertib or another one of its assets gets approved and to market.

Evaluating Clinical Milestone Assets Versus Software Revenue

If you’re a venture capitalist or crossover investor, this move to pipeline equity means you need a new way to value these companies. Sure, software metrics like Annual Recurring Revenue (ARR) and customer acquisition costs still matter for their software business, but they don’t tell the whole story anymore. Investors have to build a solid model for valuing multi-year pipelines and milestones. What should you be looking at?

• Probability of Technical Success (PTS): You’ve got to estimate the real odds a drug candidate makes it through each clinical phase (Phase 1, 2, 3) and actually gets regulatory approval at the end. This takes deep domain expertise in drug development.
• Peak Sales Potential: Forecast the market size and what peak annual sales could look like for each drug. You have to consider the competitive environment, any unmet medical needs, and pricing.
• Discounted Cash Flow (DCF) Analysis: You’ll need to run a DCF on all those future milestone payments, royalties, and potential equity stakes, applying the right discount rates to account for the time value of money and the insane risk inherent in drug development.
• Partnership Economics: Dig into the partnership deals. What are the upfront payments, the research funding, the milestone triggers (are they for development, regulatory, or commercial events?), the royalty rates, and who owns the IP? The deal terms can make or break the value that flows back to the AI company.
• Capital Efficiency: Figure out if the company can actually afford to fund its pipeline and partnerships without diluting shareholders into oblivion. That means looking hard at burn rate, cash runway, and their financing strategy. This shift from SaaS to co-owning the pipeline completely re-rates these AI drug discovery stocks. A pure SaaS model has lower risk and predictable revenue, but the upside is always capped. Co-owning the pipeline, on the other hand, brings the high risk of clinical trials but also the massive potential returns you see in traditional biotech. As an investor, you have to balance today’s software revenue against the massive, uncertain future value of clinical assets. A company like Insilico, with its AI-designed drug in Phase 3, gives you a tangible asset to value with biotech metrics, right alongside its platform technology.
  

  Methodology and Source Note

  This analysis is based on public info, corporate partnership filings, SEC disclosures, and clinical trial press releases from these companies. The numbers on Recursion’s Roche deal, for example, come right from their SEC filings on the total potential value of the partnership. We got the clinical trial status for Insilico Medicine’s INS018_055 from their official press releases and confirmed it in public clinical trial registries. Going forward, we’ll keep using these authoritative sources, plus things like FDA IND database entries and other regulatory filings, to give our readers, biotech venture capitalists and life science crossover investors, a solid, evidence-based view. With this method, you can do a real multi-year valuation of the pipeline and understand the partnership economics, which is a lot more useful than just looking at top-line revenue.