Future Human
Posts
Formation Bio

Formation Bio

Next-Gen Drugmaking

Andrew Kuzemczak
July 16, 2025

In partnership with

Find out why 1M+ professionals read Superhuman AI daily.

In 2 years you will be working for AI

Or an AI will be working for you

Here's how you can future-proof yourself:

Join the Superhuman AI newsletter – read by 1M+ people at top companies
Master AI tools, tutorials, and news in just 3 minutes a day
Become 10X more productive using AI

Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.

Welcome to the 3 new subscribers who have joined Future Human since our last edition! Join 206 other leaders learning about the future of human health by subscribing here:

TL;DR

Formation Bio is reinventing drug development by combining AI, engineering, and pharma expertise to drastically speed up clinical trials, reduce costs, and bring more treatments to patients faster. Their platform optimizes everything from patient recruitment to trial execution and drug licensing.
Their pipeline targets both common and neglected diseases—like atopic dermatitis, chronic hand eczema, knee osteoarthritis, and rare diseases impacting millions—addressing huge unmet medical needs often overlooked by traditional pharma.
Formation Bio’s AI-driven model tackles the massive inefficiencies in drug development economics, especially the costly and slow clinical trial phase, potentially breaking the industry’s “Eroom’s Law” curse by compressing timelines and costs, and creating a more sustainable, patient-centered future for medicine.

Hi friend,

Welcome back to Future Human! I hope you’re all doing well. For those of you in NYC or anywhere on the East Coast, I hope you haven’t been too affected by the recent thunderstorms. There’s nothing quite like century-old infrastructure to handle modern-day historic flash flooding. Stay safe out there, folks!

I’m just now coming up for air after the third and final week of weddings and settling into my new apartment here in Manhattan. We’re about halfway through the last summer of our lives as M2s—though I’m sure nobody’s counting. This Healthcare Leadership & Management Scholars program continues to be an incredible opportunity. I’ve found myself in rooms I shouldn’t be in, surrounded by people decades my senior who have rebuilt brand-name hospitals, led Ivy League institutions, and founded and exited massive companies. I can’t wait to see what the remaining four weeks have in store. My research in cardiac engineering and CT surgery is also going well—though admittedly less exciting.

Okay, back to science and tech!

A few weeks ago, we wrote about Isomorphic Labs. If you don’t remember, I’m insulted—but no worries. As a reminder, that’s Google DeepMind’s spinoff building on AlphaFold to accelerate drug discovery. Since that deep dive, I’ve remained fascinated by AI’s potential to accelerate discovery, clinical trials, and therapeutic commercialization. Because Isomorphic Labs is affiliated with a giant like Google, we thought it would be timely to also explore some of the more agile startups in drug discovery.

So with that, let me ask you:

What if the biggest breakthrough in medicine isn’t a new drug—but a radically faster, cheaper way to develop them? And if that’s the case, who is best positioned to lead this transformation—the scientists, the technologists, or the fastest-moving startup teams?

The Story

I recently met a Rhodes Scholar here at Weill Cornell, and unsurprisingly, I was impressed by their career path since completing their graduate studies. After that conversation, I was even more excited to discover that the company I was about to dive into was founded by someone with a similar background. While pursuing his doctorate at Oxford as a Rhodes Scholar, Ben Liu was doing what many biotech founders once dreamed of: using machine learning to discover new treatments for neurodegenerative diseases like Parkinson’s and Alzheimer’s. But it wasn’t discovery that held progress back—it was development. Promising therapies were routinely stalled in sluggish clinical trial pipelines. That frustration planted the seed for a company that would become one of the most ambitious attempts to reinvent how new medicines are brought to patients.¹

Liu teamed up with Linhao Zhang, a seasoned engineer from Oscar Health and Salesforce, to build what they call the “pharma company of the future.” What began as TrialSpark—a company focused on digitizing and streamlining clinical trial infrastructure—evolved into Formation Bio, a full-stack pharmaceutical company powered by AI. At first, they built tech tools to accelerate digital patient recruitment and site management. Then, they leveled up to running trials end-to-end for sponsors. Now, they’re acquiring and developing their own drugs, fundamentally rethinking the entire drug development lifecycle.

This storyline is unique and stands apart from other AI-first drug discovery companies. Ben and his team began by tackling the administrative side of drug development and gradually expanded to capture the entire process—unlike most others, who focus narrowly on the scientific side alone. And the need couldn’t be more urgent. As the company puts it, “despite tech advancements, bringing a new drug to market remains a decade-long, billion-dollar endeavor.” That inefficiency leads to higher costs and fewer treatments reaching patients. Formation Bio’s model aims to fix that—by building an “AI-native” pharma company from the ground up, they want to drastically cut timelines and cost across clinical development.

At the core is a belief in convergence: that the best of pharma and the best of tech need to fuse if we’re going to break out of the current system. Formation Bio is not just a tech platform—they’re building a new operating model for pharma that blends top-tier engineering, machine learning, and drug development expertise under one roof.

“Taking on the challenges of modern medicine will require a coordinated pursuit, organized against impossible odds, and synchronized for speed”

Ben Liu, Founder of Formation Bio

Investors are betting big on the idea. Formation Bio has raised over $372 million in Series D financing led by a16z, with participation from Sanofi, Sequoia, Thrive Capital, and Sam Altman, among others. And in late 2024, the company unveiled Muse, a first-of-its-kind AI tool built with OpenAI and Sanofi to optimize patient recruitment in clinical trials—a longstanding pain point in drug development. It’s the first major output of a partnership designed to fuse advanced AI models with real-world pharma expertise to speed breakthrough therapies to patients.²

The Tech

On the technical side, Formation isn’t just layering AI onto traditional pharma workflows—it’s rebuilding the entire drug development model from the ground up. Their core premise is simple: build a pharmaceutical company with the scale and reach of a big pharma organization, but the lean, iterative, and tech-enabled DNA of a Silicon Valley startup. As I mentioned above, instead of focusing narrowly on discovery, Formation Bio uses AI to accelerate and optimize every stage of drug development—from early-stage drug hunting to full clinical trials.¹

At the heart of this model is a proprietary tech platform designed to drive better decisions, faster. Formation Bio licenses or acquires clinical-stage drug candidates, and then runs them through its in-house, AI-enhanced development process. They then aim to reach critical value inflection points more efficiently than traditional players (a lot of big words by me to just say they move faster). With this approach, the company can pursue more assets in parallel—taking more shots on goal and increasing their overall chances of success.

One key example of this in motion is gusacitinib, an oral dual JAK/SYK inhibitor that Formation Bio acquired in 2022. In 2024, they licensed it to Sanofi, who plans to evaluate the drug in a previously unexplored indication through a Phase 1 trial. It’s a clean illustration of Formation Bio’s strategy: find underdeveloped assets, optimize them with tech, and partner at the right moment.³

This AI-enabled development strategy allows Formation Bio to tailor how each drug moves forward—sometimes partnering, sometimes selling, and eventually, they plan to go end-to-end with their own commercial pipeline. As co-founder Ben Liu puts it, their short-term play is partnerships and exits at key milestones, but long-term, they’re aiming to fully develop and commercialize drugs internally, transforming not just the speed of development but the economics and accessibility of new medicines.⁴

Central to this ambition is Muse. Muse targets one of the most stubborn choke points in clinical development: patient recruitment. It works by synthesizing deep research on diseases and demographics, identifying optimal patient profiles, and generating high-quality, IRB-ready recruitment materials tailored to specific subgroups. Strategic assistants conduct research, content-producing agents generate tailored outreach, and validation assistants ensure compliance and clarity.

Behind the scenes, Muse also leverages OpenAI’s Threads API, which simplifies multi-step execution and context tracking. This allows Formation Bio’s team to focus on strategy and science—while the AI handles the heavy lifting of analysis, customization, and deployment.

Pretty slick work.

The Market

The pharmaceutical industry is undergoing a seismic shift, and AI is at the epicenter. The global market for AI in drug discovery is projected to explode from $1.5 billion in 2023 to over $20 billion by 2030, growing at a staggering 29.7% CAGR. That kind of growth isn’t just theoretical—it’s already being driven by breakthroughs from tech giants like Google DeepMind, hybrids like Isomorphic Labs (throwback to newsletter #7) and Insilico Medicine, and a wave of startups aiming to reinvent how drugs are brought to patients.⁵

In this crowded space, Formation Bio has carved out a distinctive niche: not just building AI models to discover new drugs, but reengineering the entire clinical development lifecycle with AI at its core (see above: they are trying to tackle the entire path from molecule to commercialization). While companies like Atomwise, Latent Labs, and Cradle Bio focus primarily on drug discovery—using machine learning to identify molecules or generate new proteins—Formation Bio is focused on the most expensive, operationally complex part of the pipeline: clinical trials. Their work begins after a promising drug candidate is found and focuses on getting it across the finish line faster and more efficiently than traditional players.

The opportunity here is massive. The global clinical trials market was valued at $81.9 billion in 2023 and is projected to nearly double by 2033. Clinical development, especially Phase III trials, remains the most expensive part of bringing a drug to market, with median costs surpassing $19 million per trial. Even more, these late-stage studies require larger, more diverse patient populations, longer durations, and greater logistical coordination. Formation Bio’s model directly addresses this: leveraging AI to streamline patient recruitment, site selection, and protocol optimization—components that are often the main sources of delay and cost overruns.⁶

To put it simply; unlike Isomorphic Labs or Insilico, which are reshaping the front end of drug discovery with predictive biology, Formation Bio operates in the high-stakes middle—where most drugs stumble because development timelines and economics don’t work. Their ability to return value faster makes them one of the few AI-native pharma players integrating across the value chain.

The Sick

In a world where fewer than 22% of the 18,500 recognized diseases have FDA-approved treatments, Formation’s mission is critical. The company’s portfolio reflects this urgency—focusing on diseases that profoundly disrupt daily life, have few to no treatment options, and often fall through the cracks of traditional pharma R&D.⁷^,⁸^,⁹

Take atopic dermatitis, for example—a chronic inflammatory skin disease that affects up to 20% of children and 3% of adults worldwide. The unbearable itch (or pruritus) impacts over 80% of patients and can lead to sleep loss, depression, and social isolation. Formation Bio is advancing ASN008, a sodium channel blocker designed to address this symptom directly, potentially improving quality of life for millions who suffer from relentless itch.

Then there’s notalgia paresthetica, a far less known but surprisingly common neurological skin condition, causing chronic localized itching around the shoulder blades. Affecting as much as 8% of chronic itch patients, this condition has no approved treatment. ASN008 is also being tested here, in a rare instance of a biotech company pursuing an indication the industry typically ignores due to its small size and lack of blockbuster potential.

Formation Bio’s pipeline also includes sprifermin, a biologic therapy for knee osteoarthritis—a condition affecting more than 230 million people globally. Despite its scale and severity, osteoarthritis still lacks any disease-modifying therapies, making sprifermin a potential breakthrough in a disease area that too often defaults to painkillers and joint replacements.

But Formation Bio isn’t stopping at common conditions with high unmet need. Their long-term ambition is to tackle rare diseases—a category that impacts 300–400 million people globally, yet receives only a fraction of research funding. Over 95% of rare diseases have no approved treatment. Nearly one in ten Americans has one, and 50–80% of patients are children, many of whom won’t survive past the age of five. These conditions span every bodily system and include devastating neurological, metabolic, and skeletal disorders—many of which have been completely neglected by the traditional R&D system.⁹^,¹⁰^,¹¹

Formation Bio’s approach—combining fast-moving AI-powered development with a focus on high-need, overlooked conditions—offers a blueprint for what a more equitable future of medicine might look like.

The Economy

The economics of drug development are fundamentally broken. That’s not hyperbole—it’s the thesis behind Eroom’s Law, which we have touched on before. While computing power has gotten exponentially cheaper and faster, the cost to bring a new drug to market doubles roughly every nine years (Eroom’s Law). It now takes over 13 years and $2.6 billion to deliver a single approved treatment. Despite advances in biology, genomics, and AI, we’re still getting worse at translating scientific breakthroughs into accessible, affordable medicine.¹²

The biggest culprit is, in my opinion, clinical trials. More than 75% of drug development costs are incurred during this phase, and much of it is spent on outdated manual processes like in-person monitoring—where human reviewers transcribe paper records and then double-check their own work for errors that appear on virtually every page. It’s labor-intensive, slow, and expensive.

Formation Bio is going straight at this problem with an AI-native approach. Their model isn't about shaving 5% off timelines or budgets—it’s about redefining the cost structure of drug development altogether. Recruitment alone can consume 40% of a trial’s budget, and delays in enrollment are rampant—only 10–17% of studies finish recruitment on time, even internationally. With overheads often hitting $1M/month, even minor lags can result in massive capital burn or lost revenue. Muse’s multi-agent AI system not only identifies optimal patient groups but auto-generates recruitment materials and adapts campaigns in real-time, drastically improving speed and precision.¹⁴^,¹⁵

The economic drag is systemic. Regulation has made it progressively harder and more expensive to bring new drugs to market. R&D budgets balloon, not necessarily because companies are being reckless, but because innovation is stochastic and timelines are long. The result is a “throw money at it” approach, where big companies view budget size as a proxy for productivity.¹³

Formation Bio’s model offers a potential way out of this trap. By reducing time-to-data and cost-to-milestone, they not only unlock more shots on goal, but also reduce the financial risk profile of each program. Their platform could compress multi-year trial phases into single-digit months, enabling a pace of drug development more aligned with the realities of patient needs.¹²

My Thoughts

Well that is honestly not how I expected it all to play out. The behemoth from Google, Isomorphic Labs, remains honed in on one segment of front-end development, while the startup takes on the end-to-end approach. It almost seems backwards when considering resource richness, but I adore the ambition and look forward to supporting it in any way I can. The path to bring life-saving drugs to market just had its biggest break in history in AI. Capitalizing on this moment is essential, and I feel that Ben Liu is well placed with key tailwinds to succeed.

I am wishing them all the best. We will be following closely.

To more lives saved,

Andrew

I always appreciate feedback, questions, and conversation. Feel free to reach out on LinkedIn @andrewkuzemczak.