Content About Stocktwits
AI: ‘Missionaries vs Mercenaries’. Battle to Keep AI People. AI-RTZ #1168

AI: ‘Missionaries vs Mercenaries’. Battle to Keep AI People. AI-RTZ #1168

Author
Michael Parekh AI: Reset to Zero
1 day ago
9 MIN READ

This post is a ‘Part 2’ to AI-RTZ #886 from last fall on AI’s revolving doors for talent.

Big Tech companies are starting to bemoan the mercenary nature of the AI Talent War battles this AI Tech Wave. Keeping people longer term vs ‘renting’ them at the high bids and counter-bids. Especially for the coveted AI Researchers. It started in earnest last year with Meta and its founder/CEO Mark Zuckerberg really leaning into it. And has continued revolving since.

Axios lays it out in ‘AI’s talent wars have a loyalty problem’.

“Top AI researchers continue to jump between leading AI labs as competition for elite talent intensifies.”

Their sharper line comes at the end.

“AI companies are finding they can buy someone’s time, but securing lasting allegiance is proving much harder.”

Some notable names of late cited. From Google especially.

Lilian Weng left Thinking Machines Lab citing health, then joined OpenAI two days later to work on recursive self improvement. She is the fourth Thinking Machines co-founder out inside a year.

Illustration of an exit sign on a computer screen

Noam Shazeer left Google for OpenAI in June. Google had paid roughly $2.7 billion to bring him back from Character.AI less than two years before that.

John Jumper, who won the Nobel in chemistry for AlphaFold, left Google DeepMind for Anthropic after nearly nine years.

Daniel Gross left Ilya Sutskever’s Safe Superintelligence for Meta. And researchers Meta recruited into Alexandr Wang’s superintelligence group have since left for OpenAI.

Four moves, four directions, one small cozy pool in Silicon Valley.

And even the other founder/CEOs of key companies are starting to grumble about it.

“A source familiar told Axios that Anthropic CEO Dario Amodei has expressed concern about new talent coming to the firm for the money rather than the mission.”

Money rather than mission. That is worth holding onto, because it has a history.

Jensen Huang has been saying a version of it for years. He wants missionaries, not mercenaries.

I wrote that one up in August 2024 in AI: Nvidia’s ‘Missionary’ Culture. Nvidia was up 3,776% since 2019 at the time. The story was that almost nobody was leaving to go enjoy it.

Sam Altman used the same two words a year later, in a memo to OpenAI researchers while Meta was raiding the place. ‘Missionaries will beat mercenaries.’

And now Anthropic’s CEO is worried about the same thing inside his own hiring.

Three CEOs, one frame, and the slew of announcements continues.

Worth going back a step, because as I mentioned before, Meta kicked this escalation off and the whole sequence is on the record.

  • Scale AI, June 2025. $14.8 billion for 49%, valuing Scale at $28 billion, with the cash going to existing shareholders rather than into the company. Alexandr Wang, then 28, came with it to run a new superintelligence lab. (RTZ #747)

  • The researcher raids, summer 2025. Nine figure signing packages, founder led recruiting, and five Chinese researchers hired out of OpenAI in a single stretch. (RTZ #755, RTZ #770)

    OpenAI Readies Itself for Its Facebook Era

  • Apple’s designers, December 2025. Alan Dye, after nearly two decades running Human Interface at Apple, with Billy Sorrentino, into a new Reality Labs creative studio. (RTZ #926)

  • India, June 2026. $900 million for about 20% of Cred, and Kunal Shah moved from Bangalore to Menlo Park to run WhatsApp and its three billion users. (AI-RTZ #1126)

  • And then the digestion. By October 2025 Meta was already pruning the org it had just bought. Yann LeCun gone. Ahmed Al-Dahle gone to Airbnb. (RTZ #883, RTZ #903)

That is a decade of ordinary org building in most ‘normal tech waves’, run in about twelve months.

The part that got less attention at the time is that it ran both ways. It’s been revolving doors.

When I looked at it last October in AI: OpenAI & Meta’s revolving doors, roughly 630 ex-Meta people were already inside OpenAI. About a fifth of the company.

Fidji Simo at OpenAI running Applications. Vijaye Raji as her CTO. Kate Rouch as CMO. Joaquin Quiñonero Candela running recruiting.

Meta was not simply losing that exchange. It was seeding the culture of the company it was raiding.

Which brings me to the two companies that have generally kept their people over the decades, especially at senior levels.

Nvidia is the clean case. Reported turnover of 2.7% in fiscal 2024 and 2.5% in fiscal 2025, against a semiconductor industry average up near 16%.

Apple is the other one, and there the case is about the senior teams rather than a company wide percentage.

Tim Cook joined in 1998. John Ternus in 2001. Craig Federighi came in with NeXT in 1996. Greg Joswiak has been there since 1986.

Those are careers, not tours. Echoes of Goldman Sachs in its private years. My career alma mater. Different industry, similar culture on their folks being ‘long-term greedy’.

Long Term Greedy: The Triumph of Goldman Sachs

And Apple is exactly where the thesis gets tested, because Apple is now the one being raided.

Meta took designer Alan Dye. OpenAI has hired more than 400 former Apple employees, and Apple sued in July over what came with them. (AI-RTZ #1146)

John Giannandrea was replaced as AI chief. Bloomberg called the stretch Apple’s biggest personnel shake up in decades. (RTZ #929)

Even Nvidia has started buying teams rather than growing them, with the Groq acqui hire last December. (RTZ #947)

So the two companies that kept their senior teams are now the two companies everyone else is shopping at. And doing some shopping of its own.

Here is what makes this wave different from the last few, and it is not the money.

It is the number of domains running at once.

Models. Devices and industrial design. Agents. Chips.

Enterprise sales, where OpenAI and Anthropic went and hired Salesforce’s people. (AI-RTZ #1122)

And now the electricians and the carpenters, where Google, Meta and BlackRock are paying to train the trades that build the data centers. (AI-RTZ #1163)

Every prior wave had a talent war at the model layer of its day. This one is running at every layer at the same time.

I covered these cycles from the research side at Goldman Sachs, through the PC, internet and mobile waves.

I also took General Magic public in 1995, when mobile software agents were the thing everyone was hiring for.

The pattern rhymes each time. The compression is what is new.

One more layer that is easy to miss from here. The flow is not only between American labs.

China was 26% of top tier AI researchers working in the US in 2023, and by some measures produces close to half the world’s top researchers. (RTZ #767)

Tsinghua now sends roughly 20% of its engineering graduates after US PhDs, down from about half before Covid.

Some of the best people are not being hired from one lab to another. They are being hired away from the American market altogether. (RTZ #856)

Jensen has been notably even handed about this, hiring Chinese researchers directly rather than treating the pool as off limits.

My overall Take.

Loyalty is not a soft metric in this wave. It behaves like a cost of capital.

A company that keeps its people compounds. It builds on what it already knows.

A company that rents them restarts, and pays the restart cost again every eighteen months or so.

Nvidia and Apple have kept their senior people across multiple waves.

Everyone else on this list, Meta, Google, OpenAI, Anthropic, Microsoft and Amazon, sits somewhere out on the mercenary end of that spectrum. And the spectrum keeps widening.

That is not a moral point. Mercenaries win battles. Missionaries tend to win long wars, and nobody yet knows which of the two this is.

When the same few hundred people rotate through all of the labs, the labs converge.

Convergence is not what any of them think they are paying for.

It’s going to be interesting to see the divergence of results from these diasporas.

An accelerated version of the talent wars this AI Tech Wave indeed. Stay tuned.

SOURCES

Primary, this post:

For more curious AI readers, MP’s own ‘prior discussed’ threads on this topic:

(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)





Want the latest?

Sign up for Michael Parekh's Newsletter below:


Subscribe Here

More like this

AI: Weekly Summary. RTZ #708
AI: Weekly Summary. RTZ #708
Michael Parekh AI: Reset to Zero
Apple results amidst herculean moves to source beyond China: Apple reported strong Q2 quarterly results, with services revenues reaching new all-time...
1 year ago
4 MIN READ
BROWSE ALL ARTICLES