AI Just Designed a Cancer Drug

In partnership with

A company called Isomorphic Labs, spun out of Google DeepMind in London, has spent the last five years doing something that sounds like science fiction: using artificial intelligence to design cancer drugs from scratch. Not helping scientists find drugs. Not speeding up paperwork. Designing the actual molecules — the chemical structures that, if everything goes right, will enter a patient's bloodstream and attack a tumour.

This week, that work is finally moving from computers to human bodies. Isomorphic has announced it is beginning its first human clinical trials for AI-designed oncology drugs — the first time in history that a drug conceived entirely by an AI system will be tested on real patients. The company's president described it simply as "the next big milestone."

Someone just spent $236,000,000 on a painting. Here’s why it matters for your wallet.

Late last year, a Klimt sold for the highest price ever paid for modern art at auction.

An outlier sure, but it wasn't a fluke. U.S. auction sales grew 23.1% in 2025. The $1-5mm segment even grew 40.8% YoY.

Meanwhile, Apollo’s chief economist Torsten Slok said to expect ‘zero in return in the S&P 500 over the coming decade.’

So, that's where it gets interesting.

Each environment is unique, but after dot-com, post war and contemporary art grew about 24% annually for a decade. After 2008, about 11% for 12 years.

It’s also had near-zero correlation with the S&P 500 since ‘95.*

Now, Masterworks lets you invest in shares of artworks featuring legends like Banksy, Basquiat, and Picasso.

• $1.3 billion invested across over 500 artworks.

• 28 sales to date. 

• Net annualized returns on sold works held 12 months+ like 14.6%, 17.6%, and 17.8%.

Shares can sell quickly, but my subscribers can skip the waitlist:

Skip the waitlist

^{*Investing involves risk. Past performance is not indicative of future returns. See important Reg A disclosures at masterworks.com/cd.}

What is Isomorphic Labs?

It was founded in 2021 by Demis Hassabis, a British scientist who won the 2024 Nobel Prize in Chemistry for building AlphaFold — an AI that can predict how proteins fold, which is fundamental to understanding disease. Isomorphic takes that protein knowledge and uses it to design new drugs. Think of it as an architect for molecules, except the architect never sleeps and can test millions of designs before a human enters the lab. The company just raised $2.1 billion in fresh funding, bringing its total to $2.6 billion.

Why this is a bigger deal than it sounds

Developing a drug the traditional way takes, on average, twelve years and costs over a billion dollars. The process involves years of manual laboratory experiments, most of which fail. Of every 10,000 compounds that enter drug discovery, roughly one ever reaches patients. It is one of the most expensive and inefficient processes in all of science.

AI does not eliminate that difficulty — biology is genuinely complex, and no algorithm can fully predict how a human body will respond to a new molecule. But it does change the starting point dramatically. Instead of scientists making educated guesses and running physical experiments for years, AI can screen billions of potential molecules in days, predict which ones are most likely to bind to the right proteins, and flag the most promising candidates before a single test tube is used.

The result, if it works, is a drug pipeline that is faster, cheaper, and — in theory — more precise. Cancer drugs that are tailored not just to a type of tumour but to the specific genetic profile of a patient's cancer. Treatments that reach patients in years rather than decades.

"AI may not cure cancer. But it could be the thing that finally makes the search affordable enough to try a hundred more times."

What this means if you are not a scientist

Most of us will never design a drug. But this story matters for two reasons that go beyond the laboratory.

First, it is the clearest sign yet that AI is moving out of the "productivity tool" phase and into the "changes what is possible" phase. Writing faster, summarising documents, answering customer emails — these are useful. Designing molecules that did not previously exist is categorically different. It is AI not just doing old tasks more efficiently but doing things that humans could not do alone at any cost.

Second, this moment creates genuine opportunities for people who pay attention. The healthcare AI sector raised $4 billion in venture capital in the first quarter of 2026 alone — and the vast majority of that money is not going to frontier drug discovery. It is going to the quieter layer beneath: the tools that help hospitals process data, the apps that help patients manage chronic conditions, the software that helps small clinics in Lagos or Manila do in minutes what used to take days.

Those problems do not require a Nobel Prize. They require someone who understands the local context, builds a narrow solution, and earns the trust of a specific group of people who have no better option. That is exactly the kind of opportunity that one person with a laptop and the right tools can pursue — right now, in almost any country in the world.

Where the real opportunities are hiding

01Patient education, locally. AI avatar doctors are already reducing anxiety and improving understanding for cancer patients before consultations. The technology exists. What does not exist is a version built for patients in Nairobi, Accra, or Cebu — in their language, with their health system's context.

02Clinical admin for small practices. The biggest cost in healthcare is not drugs — it is paperwork. Appointment scheduling, insurance coding, follow-up communications. In most of the world, small clinics still do this by hand. A focused AI tool for one of these tasks, built for a specific market, is a real business.

03Health information people can actually use. AI now writes better patient summaries than doctors, according to a study published this week. A newsletter, a WhatsApp channel, or a simple web tool that translates complex medical information into plain language — for a specific condition, in a specific language — serves a real need that billion-dollar companies have no incentive to address.

04Chronic disease management. Diabetes, hypertension, and asthma affect hundreds of millions of people globally, most of whom receive minimal ongoing support. AI tools that send reminders, track symptoms, and flag warning signs — built for specific populations — are already proving their value in trials. The gap between "exists in a research paper" and "available to a patient in Southeast Asia" is a gap someone can close.

History will probably record Isomorphic's first human trial as the moment AI drug discovery became real rather than theoretical. Whether the drug works or not matters less than what it proves: that the pipeline from AI model to human patient is now open. What flows through it next is still being decided. ■

Top News This Week

Americas

OpenAI just put ads inside ChatGPT — and is already targeting $100 billion in ad revenue by 2030

MarketingProfs · May 21, 2026 · OpenAI launched a self-serve advertising platform that lets brands buy placements directly inside ChatGPT, with partnerships spanning Dentsu, WPP, Omnicom, and Adobe. It is targeting $2.5 billion in ad revenue this year alone. The AI assistant you use to avoid ads is now selling them.

Elon Musk lost his lawsuit against OpenAI. The $1 trillion IPO is back on track.

Reuters / The Guardian · May 20, 2026 · A court ruled against Musk's claim that OpenAI betrayed its founding non-profit mission. With the legal cloud lifted, OpenAI's planned IPO — targeting a near-$1 trillion valuation later this year — moves forward. The company now has 900 million weekly active users and over $20 billion in annualised revenue.

Google launched its most ambitious AI yet at I/O — an agent that works 24/7 even when your laptop is closed

Build Fast With AI · May 20, 2026 · Google's "Gemini Spark," announced at I/O 2026, is a personal AI agent running continuously on cloud servers, capable of acting autonomously across Google Workspace and third-party apps. It also dropped a new $100/month AI Ultra subscription, down from $250. The model wars have moved from benchmarks to who can do the most work while you sleep.

Europe & Middle East

The Pope published an AI encyclical — framing it as the defining moral challenge of our era

Build Fast With AI · May 20, 2026 · Pope Leo XIV signed a formal letter on AI on May 15, exactly 135 years after his predecessor's landmark document on the Industrial Revolution. He frames AI as a comparable disruption demanding ethical governance. It is the first papal document dedicated to AI — and a signal of how far the conversation has moved beyond tech circles.

Saudi Arabia is pouring billions into AI compute as nations race to own their own infrastructure

TechStartups · May 20, 2026 · Saudi Arabia announced major new investment in AI data centres and compute capacity this week, part of a wider pattern of governments treating AI infrastructure as a strategic national asset — on par with ports, power grids, and telecommunications.

APAC

Taiwan has 50 approved AI medical products and a national AI health strategy — most countries have neither

LucidQuest Ventures · May 18, 2026 · Taiwan's health ministry described a digital health system built on AI, cloud, and national health data — including AI-enabled chronic disease management, virtual insurance cards, and over 50 approved AI medical tools. A quiet model for what national healthcare AI infrastructure can look like.

Apple is planning to let third-party AI models run natively on iPhones — including competitors to Siri

MarketingProfs · May 21, 2026 · Apple is reportedly building "Extensions" — a system that lets users choose which AI model handles text, editing, and image tasks on their devices via the App Store. A significant opening for smaller AI providers to reach iPhone users without going through Apple's own models.

This Week's Top AI Funding Rounds