Gourav Introduction
I’ll be honest—when I first heard about Mustafa Suleyman joining Microsoft, I thought it was just another tech executive changing companies. But boy, was I wrong. The more I dug into this story, the more fascinating it became. We’re talking about a DeepMind co-founder essentially raiding his former company to build what might be the most powerful AI division in the tech world.
When I went through LinkedIn and tech forums recently, I noticed a pattern of comments like “Microsoft is basically rebuilding DeepMind inside their walls” and “Google must be panicking right now.” And honestly? After researching this, I can see why people are saying that.
Who Is Mustafa Suleyman and Why Does This Matter?
Before we dive into the hiring spree, let me give you some context. Mustafa Suleyman isn’t just some random tech executive. This guy co-founded DeepMind back in 2010—you know, that AI company Google bought for £400 million in 2014. He spent years there building some of the most advanced AI systems on the planet before leaving to start Inflection AI in 2022.
Then in March 2024, Microsoft basically said “we want you and your entire team” and created a whole new division called Microsoft AI just for him. He now reports directly to CEO Satya Nadella, which tells you how serious Microsoft is about this.
In my experience following tech news, when a company creates an entire division for one person, something big is happening.
The Great AI Talent Migration: Numbers Don’t Lie
Here’s where things get really interesting. According to CNBC’s reporting, Microsoft hired around two dozen employees from Google DeepMind in recent months. That’s not a coincidence—that’s a strategic talent grab.
Key Hires That Caught My Attention
| Name | Previous Role at Google | New Role at Microsoft | Significance |
|---|---|---|---|
| Amar Subramanya | VP of Engineering (Gemini assistant) – 16 years at Google | Corporate Vice President, Microsoft AI | Direct competitor to his previous work |
| Adam Sadovsky | Distinguished Software Engineer, Senior Director at DeepMind – 18 years at Google | Senior role at Microsoft AI | Deep technical expertise |
| Sonal Gupta | Engineering Lead at Google DeepMind | Member of Technical Staff | Building core AI capabilities |
| Jonas Rothfuss | Research Scientist at Google DeepMind | Technical Staff at Microsoft AI | Research-focused role |
| Marco Tagliasacchi | Google DeepMind Scientist | Microsoft AI (Zurich office) | Co-creator of Audio Overviews feature |
| Zalán Borsos | Google DeepMind Scientist | Microsoft AI (Zurich office) | AI podcasting tool specialist |
| Matthias Minderer | Research Scientist at DeepMind (PhD from Harvard) | Microsoft AI | Vision capabilities expert |
| Dominic King | Head of DeepMind’s Health Sector | Microsoft’s Consumer AI Health Division | Medical AI diagnostics |
| Christopher Kelly | Clinical Research Scientist at DeepMind | Microsoft’s Consumer AI Health Division | Healthcare AI applications |
When I scrolled through Reddit tech communities, I saw discussions about how Tagliasacchi and Borsos are particularly significant because they built Google’s “Audio Overviews” feature. Microsoft essentially poached the people who created one of Google’s coolest AI tools.
What Stood Out to Me: The Strategy Behind the Hiring
What I personally feel is most interesting here isn’t just that Microsoft is hiring from Google—it’s how strategically they’re doing it. Let me break down what I noticed:
H3: They’re Not Just Hiring Random Talent
Microsoft isn’t throwing money at every AI researcher. They’re specifically targeting people who:
- Worked on consumer-facing AI products (like Gemini)
- Have deep expertise in specific domains (health, audio, vision)
- Already know how to ship products, not just do research
H3: Building Entire Teams, Not Just Individuals
Here’s something that really struck me: When Suleyman joined Microsoft, he brought several team members from Inflection AI with him, including co-founder Karen Simonyan as Chief Scientist. Then he continued this pattern by hiring groups of people who already worked together at DeepMind.
In a Bloomberg interview, Suleyman explicitly said he doesn’t believe in Meta’s approach of “hiring a lot of individuals rather than maybe creating a team.” He’s building cohesive units, not collecting star researchers.
H3: Geographic Expansion
Microsoft even opened a new Zurich office specifically for some of these hires. That shows they’re willing to build infrastructure around talent, not just shove everyone into Redmond.
The Culture vs. Cash Debate
While scrolling through X (formerly Twitter), I noticed a heated debate about Microsoft’s hiring philosophy. Apparently, Meta has been offering AI researchers packages worth up to $300 million over four years, with more than $100 million paid in the first year.
Suleyman’s response? “I don’t think anyone’s matching those things.”
Instead, Microsoft is betting on:
- Long-term culture fit
- Team cohesion
- Sustainable growth
- Alignment with company values
I personally feel this is a smarter approach. Sure, you might lose a few people to massive paychecks, but if you’re building something that requires deep collaboration (which advanced AI definitely does), you want people who actually want to work together.
What This Means for Microsoft’s AI Future
Based on everything I’ve researched, here’s what I think is happening:
H4: Microsoft Is Building AI Independence
For years, Microsoft relied heavily on OpenAI. But Suleyman has been crystal clear about wanting “AI self-sufficiency.” He’s building Microsoft’s own foundation models that can compete at the frontier level.
According to Fortune’s reporting, Microsoft formed the “MAI Superintelligence Team” in November 2025, focused on what Suleyman calls “Humanist Superintelligence (HSI).”
H4: They’re Going After Specific Product Categories
The hires tell a story:
- Audio/podcasting (Tagliasacchi, Borsos): Competing with tools like NotebookLM
- Healthcare AI (King, Kelly): Building diagnostic systems
- Vision capabilities (Minderer): Understanding images and video
- Consumer assistants (Subramanya): Making Copilot better than Gemini
H4: The Bold Automation Prediction
Here’s where things get controversial. In a recent Financial Times interview, Suleyman predicted that most white-collar work will be fully automated within 12-18 months.
He specifically named:
- Lawyers
- Accountants
- Project managers
- Marketing professionals
When I saw this circulating on Instagram, the comments were split between “this is terrifying” and “finally, I can automate the boring stuff.” I’m somewhere in the middle—I think he’s being optimistic about the timeline, but the direction is clear.
The Competitive Landscape: Who Else Is Doing This?
To be fair, Microsoft isn’t alone in this talent war. While researching, I found:
- Meta: Offering $100 million signing bonuses (according to OpenAI CEO Sam Altman)
- Google: Spent $2.4 billion acquiring Windsurf’s leadership team
- OpenAI: Matching Meta’s offers to retain staff
But what makes Microsoft’s approach different is the focus. They’re not just accumulating talent—they’re building specific capabilities with specific teams for specific products.
My Personal Take on All This
After spending hours researching this story, here’s what I honestly think:
The Good:
- Microsoft is playing the long game with culture-first hiring
- They’re building actual products, not just hoarding researchers
- The focus on “Humanist Superintelligence” suggests they’re thinking about safety and alignment
The Concerning:
- This level of talent concentration in a few companies could stifle innovation elsewhere
- Google is losing some of its best people to a competitor
- The aggressive automation timeline could create real economic disruption
The Fascinating:
- We’re watching AI talent become more valuable than almost any other resource
- The relationship between Suleyman and his former DeepMind colleagues (including current Google DeepMind head Demis Hassabis) makes this personal
- Microsoft went from being an OpenAI partner to building its own competitive AI stack in less than two years
Comparison: Microsoft’s Approach vs. Competitors
| Company | Hiring Strategy | Focus Area | Investment Style |
|---|---|---|---|
| Microsoft | Culture-first, team-based | Product integration, HSI | Strategic, measured |
| Meta | Mega bonuses, individual stars | Infrastructure, internal tools | Aggressive, high-spend |
| Acquisitions, defensive retention | Maintaining lead | Reactive, protective | |
| OpenAI | Matching offers, mission-driven | AGI research | Defensive, competitive |
Looking Ahead: What to Watch
Based on my research, here are the things I’ll be paying attention to:
- Microsoft’s 2026 model releases: Suleyman hinted these could launch this year
- Copilot evolution: Will it actually become better than ChatGPT and Gemini?
- Google’s response: Will they increase retention packages or double down on their own AI development?
- The automation reality check: Will Suleyman’s 18-month prediction hold up?
Conclusion
So, did Mustafa Suleyman hire AI experts from Google for Microsoft’s AI advancements? Absolutely, yes. But it’s more nuanced than just “poaching talent.”
He’s systematically building teams of people who already know how to work together, focusing on specific product categories, and betting that culture and mission will win over massive signing bonuses in the long run.
Whether this strategy works remains to be seen. But one thing is clear: the AI talent wars are real, they’re intense, and Microsoft is playing to win.
