Podcast Summary: No Priors – "How AI Will Accelerate Breakthroughs in Biotechnology" with Benchling CEO Sajith Wickramasekara

Date: November 13, 2025
Host: Sarah Guo (Conviction)
Guest: Sajith "Saji" Wickramasekara (Co-founder & CEO, Benchling)

Overview

This episode dives deep into how artificial intelligence (AI) is transforming the biotechnology sector, with a special focus on the challenges of drug discovery, the slow pace and high cost of biotech innovation, and how AI—through platforms like Benchling—could dramatically accelerate breakthroughs, reduce costs, and change the way both science and business are conducted in the life sciences. The discussion also explores macroeconomic trends, China’s increasing role in biotech, lessons from notable drug development stories, and the challenges of building companies that combine the best of both software and science.

Key Discussion Points & Insights

1. Benchling’s Mission and Scale

• Benchling was founded 13 years ago after Saji, a software engineer with lab experience, realized scientists lacked modern digital tools (00:44).
• They serve over 1,300 biotech and pharma companies and 7,000+ academic institutions worldwide, powering giants like Moderna, Sanofi, Eli Lilly, as well as cutting-edge startups (01:34).

  “When I got to the biology lab, I found that scientists had paper notebooks and spreadsheets… it was terrible.” — Saji (00:54)

2. Nature of Scientific Data in Biotech

• The complexity of drug development involves thousands of steps from target discovery to mass production and post-approval work (02:28).
• Benchling collates highly heterogeneous lab data (molecules, tests, animal studies, scale-up, etc.) to make it searchable and actionable (03:15).

3. The Biotech "Macro" Cycle & Recent Downturn

• Post-COVID, biotech funding and valuations plummeted, drawing parallels to the dot-com bust (04:10).
• Factors: rate hikes, regulatory uncertainty, overly optimistic bets on next-gen modalities (gene editing, RNA therapies), and shifts in global dynamics (04:57).

4. China’s Growing Role in Biotech

• Chinese biotechs now outpace Western companies on speed and cost, increasingly becoming sources for large pharma pipeline deals (06:40).
• Example: Johnson & Johnson partnered with China's Legend Biotech for CARVYKTI (cancer drug for multiple myeloma), now a US commercial success (07:34).

  “China is very good at things related to speed and cost.” — Saji (06:41)

5. Why Drug Discovery Is Expensive & Slow

• It takes over $2 billion and up to 10 years to bring a drug to market, with most failing late in development (10:09).
• Much of this is due to bespoke, artisanal digital workflows and lack of systematization compared to advances in the physical realm (09:29, 11:05).

  “It is probably easier at this point to send things to space… than it is to get a new medicine approved.” — Saji (10:33)

6. Critical Industry Inefficiencies

• The biotech industry struggles with buying external innovation early enough (GLP1, Keytruda stories), resulting from poor progress prediction and expensive "shots on goal" (14:31).
• Platform companies once hailed as the solution fell out of favor when investor preferences shifted, stranding startups mid-development (06:38, 13:37).

7. The Promise of AI in Biotech

• Saji is bullish on AI as the path to speed, scale, and higher quality discoveries (08:58).
• AI can lower failure rates and streamline steps beyond just clinical trials—most crucially, it can recover "lost" institutional memory (such as prior experiments or negative data trapped in old notebooks) (18:16).

  "We view this as being able to unlock memory for these organizations and help make scientific data reusable." — Saji (17:53)

8. Benchling AI—Capabilities and Applications

• Two Prongs:
  1. Seamlessly integrating models (simulation, structure prediction, etc.) into scientific workflows.
  2. Deploying AI agents (like a “deep research agent”) to automate and accelerate time-consuming research tasks (16:19).
• Example: AI surfaced prior mouse studies, saving months and major costs for a customer (17:21).

9. Role of Scientists & The AI Scientist Debate

• AI will augment more than replace scientists in the near term, similar to the “co-pilot” role in radiology (18:48).
• Full autonomy is likely further away than people wish—short-term gains will come from meaningful automation and decision support (19:08).

  “I’m a little bit more bullish on the augmentation model… make things better, one experiment at a time.” — Saji (18:52)

10. Trust, Transparency & Adoption Challenges

• Scientists remain skeptical; trust, accuracy, and workflow integration are essential for adoption, especially outside of Silicon Valley (21:13).
• The big leap will come from making AI "legible" and useful in specialist workflows (22:36).

  "The AI that wins is gonna be the one that people actually use." — Saji (22:50)

11. State of AI Model Development

• Open-source science models are rapidly growing (24:19).
• Pharma with scale are building unique, proprietary models using their data generation advantage (22:59).
• Foundational models exist for structure prediction (AlphaFold, Bolt), but fewer for manufacturing or "downstream" bottlenecks (26:02, 26:24).

12. The Future of Business Models in AI Biotech

• Pure-play AI model companies will likely need to diversify; the market for selling bespoke models to large pharma is probably too small and commoditizing fast (27:24).
• Opportunities: SaaS-like distribution for AI models, unlocking more data transactions (e.g., selling negative experimental results) if trust/fidelity/normalization challenges are solved (29:17).

13. Lessons on Building Biotech-Software Hybrids

• Saji emphasizes relentless customer empathy, product-market fit vigilance, and bridging the gap between science and software (37:15, 38:25).

  "The number one piece of advice I’d give people: go talk to customers." — Saji (37:15)

• Success with scientists as users hinges on cross-disciplinary teams, iterative deep dives with “representative customers,” and understanding both scientific and commercial incentives (40:54).

14. Culture & Cross-Pollination Lessons

• Biotech/Pharma could learn: storytelling, direct-to-public communication, building recognizable leaders, and sharing the human side of success from tech. (43:10)
• Tech could learn: rigor, safety, and the importance of systematized, validated processes from life sciences, especially when dealing with regulated, mission-critical domains (44:36).

  "Move fast and break things does work for certain domains… but once you want to deliver medicines, that stuff comes to matter." — Saji (44:18)

15. Personal Reflections on AI

• Saji has rediscovered the "whimsy" of coding through agentic AI coding tools and values seeing non-technical users (family) adopt AI through intuitive interfaces (voice, chat) (46:28).

Notable Quotes & Memorable Moments

• On Drug Discovery’s Difficulty:
  "It is probably easier at this point to send things to space or put people on the moon than it is to get a new medicine approved." — Saji (10:33)

• On AI’s Institutional Knowledge Recovery:
  "We sort of view this as being able to unlock memory for these organizations and help make scientific data reusable over time." — Saji (17:53)

• On the AI Scientist Hype:
  "Everyone is saying AI is going to cure disease… I would love that to happen and I’m maybe more optimistic long-term, but short-term, I’m bullish on the augmentation model." — Saji (08:58, 18:48)

• On Company Building in Complex, Regulated Markets:
  "All the times the company has been at its lowest is when I’ve gotten too far from customers… product market fit is a moving target." — Saji (37:15)

• On Bringing Science & Business Together:
  "In order for us to achieve our mission and to keep delivering great things to customers, we have to make money. And a lot of attention has come from the need to do that." — Saji (41:51)

• On Biotech’s Communication Challenge:
  "Most people couldn’t name five CEOs of pharma companies, but everyone knows the tech CEOs… I think biotech and pharma need to tell their stories." — Saji (43:31)

Timestamps for Key Segments

• Benchling’s Mission & Scale: 00:37 – 02:08
• Drug Discovery Complexity: 02:28 – 03:49
• Biotech Macro Trends & Downturn: 04:04 – 05:56
• China’s Role in Biotech: 06:40 – 08:13
• AI in Biotech—Promise & Use Cases: 08:58 – 18:16
• State of Model Development & Business Models: 22:59 – 29:28
• Company Building & Cultural Integration: 34:54 – 42:26
• Tech/Bio Culture Exchange: 43:10 – 46:20
• AI Tools and Personal Excitement: 46:28 – 47:49

In summary, this episode is a masterclass in understanding how and why biotech is “ripe” for technological transformation, why AI’s promise matters now, and what it takes to build an impactful company in one of the world’s most complex industries. Saji mixes industry realism with optimism, offering grounded advice for entrepreneurs, investors, and scientists aiming to catalyze scientific progress with AI.