May 13, 2025

Let’s get clinical

Hi there, 

The president of the US famously gets a detailed briefing in the morning, customized by a large, well-resourced research team.

You can get the same kind of tailored insights — delivered to your inbox — with CB Insights’ Personal Briefings. 

It’s more than a summary — it’s what to do next. And it’s just for you, based on your role, company, and interests.  

Sign up for a free trial to get your first Personal Briefing.

And now, let’s get into this week’s deep dive.

Clinical collab

It’s rumored that OpenAI is in talks with the FDA to develop an AI tool called “cderGPT” — a reference to the FDA’s Center for Drug Evaluation and Research (CDER) — that FDA scientists could use to accelerate the drug evaluation process.

While the partnership hasn’t been confirmed, the FDA recently announced it completed its first-ever AI-assisted scientific review and plans to deploy AI across all of its centers by June 30.

AI could accelerate multiple aspects of the drug evaluation process, from checking applications for completeness to actually assessing the likelihood of a candidate’s success. 

But the FDA evaluation process, which typically takes <10 months, is not the biggest time or cost-sink in the drug development pipeline. The real opportunity for AI in pharma lies in streamlining the discovery and trial phases, which can take 10+ years and cost billions of dollars.

AI is poised to transform 3 specific phases, which we'll break down below: 

1. Drug discovery
2. Pre-clinical development
3. Clinical trials

1. Drug discovery

AI capabilities are broadening from their foundation in small molecules to include larger and more computationally complex therapeutics, like biologics.

CB Insights data shows AI drug discovery funding surged in 2024, with biologics developers securing $1.6B — more than double 2023 funding. 

Eight of the ten largest acquisitions in this space have occurred since 2023, with companies like Recursion leading the charge, making three AI acquisitions in just over a year.

All 10 of the largest big pharma players globally have partnered with AI drug discovery startups since 2023. Simultaneously, all 10 are developing in-house AI capabilities.

Get our analysis of over 500 companies applying AI to the discovery phase in this report.

2. Pre-clinical development

For pharma companies, it’s critical to maximize efficiency in preclinical development — the stage after drug discovery and before clinical trials — as poor processes can mask early warning signs of drug candidates that are likely to fail.

While AI can accelerate the path from bench to bedside, the preclinical AI startup landscape remains highly nascent: The space has one-fifth as many startups as drug discovery AI, and half of these companies were founded since 2020, according to CB Insights data.

Within the preclinical stage, formulation development has emerged as a central application for AI. Since 2023, more than half of preclinical AI investment dollars have gone to formulation and tablet development. The concentration of funding reflects the increasing formulation complexity seen in the industry, from complex biologics to poorly soluble small molecules.

See our breakdown of the preclinical AI space here.

3. Clinical trials

Clinical development is one of pharma’s most costly and high-risk investments, with clinical trials averaging $55M each.

CB Insights data shows patient recruitment and data analysis are the 2 most active clinical development AI markets. These functions are drawing attention for their potential to tackle major trial bottlenecks — accelerating patient enrollment and enabling large-scale data analysis.

Notably, Alphabet has established a dominant position in the clinical development space, investing in 6 equity deals since 2020 — more than any single pharma or big tech competitor. It also bolsters its sphere of influence through accelerator programs and Google Cloud infrastructure collaborations.

See our report on the clinical development AI landscape here.

The prognosis

The potential FDA-OpenAI partnership is just one signal of AI's growing role in drug development. 

For pharmaceutical companies, it’s crucial to identify where AI can address specific pipeline bottlenecks — whether in discovery, preclinical, or clinical development — in order to slash the exorbitant costs and lengthy timelines involved with bringing new therapies to market.

Which phase of drug development do you think stands to benefit the most from AI?

1. Drug discovery
2. Pre-clinical development
3. Clinical trials

Reply and let me know what you think.

TLDR — Tech loves drama, right?

• Frenemies: Sam Altman extended a digital olive branch to Elon Musk this weekend, calling for peace in the name of artificial general intelligence. The exchange, sparked by an X thread on Trump, turned from political flashbacks to playful back-and-forths — including a Grok poll where the chatbot ultimately picked Musk as the AI leader humanity needs. Despite Altman’s let’s-be-friends pitch, Musk reposted Grok’s answer with “Asked and answered,” signaling the feud might be cooling… but not quite over.

Source: X

• $13B tension: Microsoft and OpenAI are reportedly locked in “tough negotiations” over the terms of their high-stakes partnership, according to the Financial Times. The talks center on Microsoft’s equity in OpenAI’s restructured for-profit subsidiary — a structure housed under its still-nonprofit parent — and what happens when the current deal expires in 2030. Microsoft, which has invested over $13B in OpenAI, is reportedly willing to trade some of its equity stake for extended access to future OpenAI models. But with OpenAI’s Stargate infrastructure ambitions and enterprise growth increasingly overlapping with Microsoft’s turf, the power dynamic is shifting fast.
  
  

• Kill zone: Y Combinator is backing the DOJ and state attorneys general in their antitrust case against Google. In a new amicus brief, YC argues that Google’s dominance has created a “kill zone” that discourages startup innovation, particularly in AI-powered search and agent tools. The brief calls for sweeping remedies, including forced data access, bans on exclusionary deals, and protections against retaliation. YC frames the issue as a systemic threat to startup funding and internet innovation.

Source: X

• Strapped for upgrades: Whoop is backpedaling after backlash over its confusing (and apparently broken) promise of free device upgrades. When the 5.0 fitness tracker launched, users were told they’d need to pay or extend their subscription to receive the new device — contradicting a previously advertised promise offering free hardware upgrades after six months. Whoop now says that earlier policy was posted in error and has introduced a new standard: users with 12 or more months remaining on their subscription get a free upgrade or refund, while others can extend to qualify. The response has split the community, with some praising the resolution and others criticizing the reversal.

   

• Fined print: Google will pay $1.4B to settle two Texas lawsuits accusing it of secretly tracking users’ location, incognito searches, and biometric data. The suits, filed by AG Ken Paxton in 2022, mark the largest privacy-related payout from a state attorney general to date. Google continues to deny wrongdoing and won’t make any product changes, stating the claims are outdated and have already been addressed in previous settlements. The deal adds to Google’s mounting legal expenses as scrutiny over its data practices grows.

   

• Blocked & loaded: Microsoft has officially banned employees from using the DeepSeek AI app, citing concerns over data privacy and the risk of Chinese state propaganda. During a US Senate hearing, Microsoft President Brad Smith confirmed that the app has been banned internally and removed from the company’s app store. The app reportedly stores user data on servers in China and restricts politically sensitive topics. Microsoft still offers a modified version of DeepSeek’s open-source model on Azure but says it altered the system to remove potentially harmful behavior.

   
• Name game: Social startup Fizz is suing Instacart and Partiful for trademark infringement over their new party delivery app — also named Fizz. The lawsuit argues that the identical name, overlapping Gen Z target market, and event-planning focus create confusion. Fizz says it’s been using the name since 2022 and alleges Instacart registered fizz.‍com in bad faith to capitalize on its brand recognition. The case is headed to court, with Fizz seeking damages and an order blocking further use of the name.
 

I love you.

Anand

@asanwal 

Co-Founder & Exec Chair

P.S. Our analysts recently explored AI’s impact across the entire R&D pipeline in a live briefing. Get the recording for free here.

Get started with CB Insights

CB Insights' emerging technology insights platform provides all the analysis and data from this newsletter. Our data is the easiest way to discover and respond to emerging tech. 

Was this email forwarded to you? Sign up here