According to Forbes, developing a new drug costs more than $2 billion and takes over 10 years, with 90% of drug candidates failing in clinical trials. Compugen, founded in 1993 when AI was called “computational discovery,” has transformed from a computational service provider to a clinical stage innovator with four clinical stage programs and partnerships with Bayer, Bristol Myers Squibb, Gilead, and AstraZeneca. Executive Chairman Anat Cohen-Dayag emphasizes that success requires combining AI with deep drug development expertise, noting “We have failed many times, but we’ve learned how to feed back the learning from failures” into platform refinements. The broader industry sees massive investment, with Isomorphic Labs raising $600 million last March and AI funding in drug R&D reaching $3.8 billion in 2024, while CB Insights projects AI could generate over $350 billion in annual value for pharmaceuticals. This convergence of technology and biology signals a fundamental shift in how medicines will be discovered and developed.
Sponsored content — provided for informational and promotional purposes.
The Expertise Gap in AI-Driven Discovery
The critical insight from Compugen’s three-decade journey is that successful AI implementation in drug development requires more than just algorithms and data—it demands deep biological intuition. Many current AI startups operate under the assumption that sufficient computational power and large datasets automatically translate to drug discovery breakthroughs. However, the drug development lifecycle involves complex biological systems where context, mechanism of action, and clinical translatability matter more than pattern recognition alone. Compugen’s pivot from service provider to clinical innovator reflects this understanding—true value comes from integrating computational capabilities with therapeutic expertise to make strategic decisions about which targets to pursue and how to develop them effectively.
Separating AI Substance from Hype
The current funding environment—$3.8 billion in AI drug R&D investment last year—creates both opportunity and distortion. While the potential $350 billion annual value projection from CB Insights’ analysis attracts legitimate innovation, it also fuels an AI bubble where companies may overstate their capabilities. The key differentiator will be which organizations can demonstrate actual clinical progress rather than just computational prowess. Nvidia’s support for Eli Lilly’s “AI factory” and similar initiatives represent serious infrastructure investments, but the real test comes when these systems produce candidates that successfully navigate clinical trials. Companies that treat AI as a magic bullet rather than a tool integrated within established biological understanding risk repeating the genomics disappointment that Compugen’s leadership witnessed earlier in their careers.
The 2025 Tipping Point and Beyond
If 2025 indeed marks the tipping point for AI in life sciences, we should expect to see several key developments over the next 24 months. First, the market will begin separating platforms that generate biologically relevant insights from those that simply identify correlations. Second, successful companies will increasingly look like Compugen’s integrated model—combining computational discovery with internal development capabilities rather than operating as pure service providers. Third, partnerships between AI specialists and pharmaceutical giants will evolve toward more equitable risk-sharing as computational capabilities prove their value. The companies positioned for long-term success will be those that view AI not as a replacement for drug development expertise but as a powerful amplifier of human intelligence in navigating biological complexity.
The Hard Road to Clinical Validation
What most AI drug discovery discussions overlook is the implementation gap between computational prediction and clinical success. Compugen’s experience of “failing many times” and learning from those failures mirrors the broader challenge facing the industry. The computational infrastructure investments by companies like Nvidia and Lilly provide necessary foundation, but biological systems introduce complexities that algorithms alone cannot navigate. The most significant barrier isn’t technical capability but understanding which computational outputs translate to clinically viable therapies. Over the next several years, we’ll see increased focus on validation frameworks and standards for evaluating AI-generated drug candidates—moving beyond in silico predictions to demonstrated clinical benefit.
