The Exit

Reid Hoffman is done with boardrooms — at least Microsoft's. The tech giant announced Thursday, June 5, that Hoffman is stepping down from its board of directors, according to TechCrunch.

He's not retiring. He's going back to building.

A Decade of Consequential Seat-Warming

Hoffman joined Microsoft's board in 2016 after the company acquired his company LinkedIn for $26.2 billion.

While he was on that board, Microsoft dropped its first $1 billion into OpenAI in 2019. That investment has since ballooned into one of the most consequential tech bets in history. Hoffman was already an original OpenAI investor before that deal, giving him a rare double-exposure to the AI moment everyone else is still trying to catch up to.

He quietly stepped off OpenAI's board in 2023, citing conflicts of interest — which was the honest thing to do, given how tangled his investments had become.

Also while on Microsoft's board, the company cut a $650 million non-acquisition deal with Inflection AI — Hoffman's other startup. That deal brought Inflection co-founder Mustafa Suleiman into Microsoft as an executive. Hoffman was on both sides of that table.

What He's Running Toward

The destination is Manus, an AI drug discovery company. Hoffman told Microsoft CEO Satya Nadella on his Possible podcast that he realized he needed to "go founder mode" after watching the startup's recent progress.

Manus raised over $50 million across a couple of seed rounds last year, according to TechCrunch. Investors include Hoffman himself and General Catalyst.

Hoffman is listed as co-founder and chairman — NOT the CEO. That role belongs to Dr. Siddhartha Mukherjee, a physician, biologist, and Pulitzer Prize-winning author of the 2011 book The Emperor of All Maladies: A Biography of Cancer. That's a serious person running a serious operation.

The "Move 37" Pitch

Hoffman's stated goal for Manus is what he calls "Move 37 AI" — AI that doesn't just assist human researchers but actually surpasses human creativity in chemistry. Specifically targeted at cancer.

Move 37 is a reference to the famous AlphaGo move against world champion Lee Sedol in 2016 — a move no human would have made, and one that changed how people thought about machine intelligence. Hoffman is saying he wants that moment, but for oncology.

That's a bold claim. It's also a measurable one. Either Manus produces drug discoveries that human researchers couldn't have reached on their own, or it doesn't.

What Mainstream Coverage Is Missing

The tech press is covering this mostly as a feel-good "visionary returns to his roots" story. That's too soft.

First: Hoffman was on Microsoft's board during the Inflection AI deal — a deal that benefited his own co-founder and arguably his own reputation in the AI space. The structure of that deal, a licensing arrangement rather than a direct acquisition, was widely seen as unusual. Microsoft got the talent and the tech without a full acquisition price. Hoffman wasn't recused.

Second: Manus is early-stage. $50 million in seed funding sounds like a lot until you remember that serious drug discovery requires hundreds of millions and takes years before any clinical validation. The "Move 37" framing is compelling, but drug development is where AI hype goes to die. We've heard the AI-will-cure-cancer pitch before. The difference here is Mukherjee — a genuine expert, not a tech bro cosplaying as a scientist.

Third: Hoffman leaving the Microsoft board right now carries timing considerations. Microsoft is navigating a complicated relationship with OpenAI, under pressure to define what it actually owns versus what it's just paying for. Hoffman, as a former OpenAI board member AND a Microsoft board member, was sitting on top of one of the most conflict-ridden positions in all of tech. His departure cleans that up — for everyone.

What This Means for Regular People

If Manus works — genuinely works — it could mean faster, cheaper paths to cancer treatments. AI-driven drug discovery, done right, could compress decade-long research timelines into years.

If it doesn't work, it's a $50 million lesson from one of tech's most connected networkers.

But the pattern Hoffman represents deserves attention: the people who built the current AI infrastructure are now turning it inward, pointing it at hard science. That's either the most important thing happening in tech right now, or the most expensive hype cycle yet.

The data will tell us which. Eventually.