Anthropic Acquires Vercept AI as Meta Poaches Co-Founder in High-Stakes Talent War
Anthropic is acquiring Seattle AI startup Vercept, folding its desktop “computer use” technology and team into Claude as the race to build AI agents that can operate software intensifies. This move highlights the escalating competition in agentic AI—a field where autonomous systems complete multi-step tasks without constant human oversight. What makes this story even more intriguing is the timing: just as Anthropic acquires Vercept AI, Meta reportedly offered co-founder Matt Deitke $250 million over four years to join its Superintelligence Lab, creating a fascinating scenario where tech giants are simultaneously competing for companies and individual engineers.
The Anthropic Vercept deal represents more than a simple acquisition. It’s a strategic play in what industry insiders call the “agentic AI arms race.” The Vercept team—including co-founders Kiana Ehsani, Luca Weihs, and Ross Girshick—have spent years thinking carefully about how AI systems can see and act within the same software humans use every day. That expertise maps directly onto some of the hardest problems we’re working on at Anthropic.
But here’s where things get messy. While Anthropic buys Vercept and brings most of the team on board, Meta successfully lured away one of its key founders before the ink dried. This creates a split scenario where both tech giants walk away with pieces of the same pie—and signals just how fierce the battle for AI talent has become in 2026.
Understanding the Anthropic Vercept Deal: What’s Really at Stake
Computer use enables Claude to do all of that inside live applications, the way a person at a keyboard would. This capability represents a fundamental shift in how we think about AI assistants. Rather than chatbots that answer questions, we’re moving toward agents that actually complete work.
The agentic AI startup acquisition makes perfect sense for Anthropic AI. Vercept closed a $16 million seed round in January 2025, valuing the company at $67 million post-money, according to Pitchbook data. However, Vercept CEO Kiana Ehsani revealed on LinkedIn that the total funding reached around $50 million, suggesting additional capital beyond the initial seed round.
The investor list reads like a who’s who of tech royalty. The angel list was notable: former Google CEO Eric Schmidt, Google DeepMind chief scientist Jeff Dean, Cruise founder Kyle Vogt, and Dropbox co-founder Arash Ferdowsi all participated. When this caliber of investors backs a startup, you know something special is brewing.
Anthropic acquires Vercept AI specifically for its desktop agent technology called Vy. Vercept’s flagship product, Vy, let an AI control a remote MacBook to complete multi-step tasks across apps — the sort of end-to-end work that APIs alone can’t cover. Think about what this means: instead of building individual integrations for every software tool, you create an AI that can literally see your screen and operate any program like a human would.
Meta Poaches Vercept Co-Founder: The $250 Million Question
The drama intensified when Meta’s aggressive talent recruitment strategy snatched Matt Deitke from Vercept’s founding team. Matt Deitke, co-founder of Seattle-based AI startup Vercept, has joined Meta’s Superintelligence Lab after receiving a direct offer from CEO Mark Zuckerberg. Deitke described the initiative as “the most exciting bet I’ve seen in tech history.”
The compensation package reportedly reached stratospheric levels. In 2025, Meta successfully recruited Vercept co-founder Matt Deitke for its Superintelligence Lab, reportedly offering a compensation package worth approximately $250 million. To put this in perspective, most Silicon Valley engineers earn comfortable six-figure salaries. This offer represents generational wealth for a single individual—and demonstrates how desperately Meta wants to catch up in the AI agent race.
Meta poaches Vercept co-founder despite Vercept’s successful trajectory. She said Vercept had a “comfortable runway and a successful product” when the opportunity to join Anthropic emerged. This wasn’t a struggling startup looking for an exit. Instead, it was a well-funded company with momentum that faced an impossible choice when two tech giants came knocking.
Not everyone celebrated the outcome. Contacted via phone this morning, Etzioni elaborated, “I’m pleased to have gotten a positive return but obviously disappointed that after just a little over a year with so much traction, and such a fantastic team, we’re basically throwing in the towel.” Oren Etzioni, another Vercept co-founder and prominent AI researcher, publicly expressed mixed feelings about the rapid exit.
The Impact of Anthropic Vercept Acquisition on Computer Use AI
Let’s talk about what makes computer use technology so valuable. This acquisition follows the recent launch of Claude Sonnet 4.6, which shows a major improvement in computer use skills: on OSWorld, a widely-used evaluation for AI computer use, our Sonnet models went from under 15% in late 2024, when we first released computer use, to 72.5% today. Sonnet 4.6 is now approaching human-level performance on tasks like navigating complex spreadsheets and completing web forms across browser tabs.
The jump from 15% to 72.5% performance in roughly a year is staggering. We’re witnessing exponential improvement in AI’s ability to actually operate computers rather than just generate text. The impact of Anthropic Vercept acquisition accelerates this trajectory even further.
Anthropic buys Vercept precisely because their proprietary model outperformed competitors. The company said VyUI outperformed models from OpenAI, Google, and Anthropic on UI grounding benchmarks. That’s right—Vercept’s technology beat Anthropic’s own models at understanding user interfaces. Rather than compete, Anthropic simply acquired the superior technology.
The competitive landscape extends beyond just Anthropic and Meta. Perplexity AI Inc. today introduced a new agentic automation tool, Perplexity Computer, that can summarize documents, generate code and perform other tasks. Some of its features are powered by Anthropic’s flagship Claude 4.6 Opus model. Even as companies compete, they’re also partnering and building on each other’s platforms.
The Vercept deal is Anthropic’s second acquisition since launch. In December, it bought the startup behind Bun, an open-source JavaScript runtime. Anthropic is clearly pursuing an acquisition-driven growth strategy, bringing in specialized teams rather than building every capability in-house.
Agentic AI Startup Acquisition Trends: A Broader Pattern
The Anthropic acquires Vercept AI transaction fits into a much larger trend. The Agentic List identifies the 120 most promising private companies building enterprise-grade agentic AI — autonomous systems that make decisions and take actions by themselves and collaborate with human teams. We identify companies that are trending up in the last 12 months, making significant impact in large enterprises and key industries. Over 5,000 nominations were received, and nearly 2,000 private companies were screened across product maturity, enterprise adoption, competitive differentiation, growth momentum, and funding trajectory.
Think about those numbers for a moment. Thousands of companies are now building in this space. The agentic AI startup acquisition wave is just beginning.
Other major players are making similar moves. Meta acquired AI agent startup Manus in late 2025 for over $2 billion, demonstrating their commitment to catching up in this space. With Meta Platforms bringing Manus’ platform and incredible slate of talent aboard, the company might have jumped several places ahead in the AI race, especially as 2026 looks to become an even bigger year for the adoption of agentic AI.
Meanwhile, HCLSoftware announced plans to acquire Belgium-based Wobby, an agentic AI startup focused on data analytics. These deals signal that agentic AI has moved from experimental technology to must-have capability.
The talent war extends beyond acquisitions. Anthropic AI faces constant pressure to retain employees as competitors dangle massive compensation packages. But Vercept was operating in a crowded and fast-moving field. Open-source projects like OpenClaw — the viral AI agent that automates tasks through messaging apps like WhatsApp and Telegram — have exploded in popularity.
What Anthropic Gains Beyond Technology
When you strip away the headlines, the Anthropic Vercept deal brings three critical assets:
First, proven technology that already works. Vercept’s Vy demonstrated 92% accuracy on automation benchmarks, far surpassing OpenAI’s 18.3% on similar tasks, providing Anthropic with proven technology for vision-based control that bypasses rigid API dependencies. This isn’t research—it’s production-ready software that real users were paying for.
Second, specialized talent that’s incredibly rare. Co-founders Kiana Ehsani and Luca Weihs, alongside researcher Ross Girshick, are joining, bringing deep Allen Institute roots and applied experience in screen-grounded agents. These individuals spent years at the Allen Institute for AI, one of the premier AI research organizations globally. You can’t simply hire this expertise on the open market.
Third, accelerated timelines. Buying specialist teams shortens the path from research demos to enterprise-grade features. Building computer use capabilities from scratch might take years. Acquiring Vercept compressed that timeline to months.
Anthropic buys Vercept knowing that Vercept’s desktop application, Vy, will shut down in 30 days as part of the transition, according to the startup’s message to users, which encouraged them to try Anthropic’s Claude tools as an alternative while the service winds down. Existing Vercept customers face forced migration to Claude, instantly expanding Anthropic’s user base.
The Messy Reality: When Acquisitions Split Teams
Let’s address the elephant in the room. The Anthropic Vercept deal didn’t bring the entire team together. Along with Deitke, he is also not joining Anthropic, and was vocally less pleased about the acqui-hire. He posted on LinkedIn: “After a little bit more than a year, Vercept is throwing in the towel and giving their customers 30 days to get off the platform.”
The public disagreement between co-founders and investors adds another layer of complexity. On Etzioni’s LinkedIn post, he accused Bannon, the Vercept lead investor, of being “partly responsible” for Vercept not hiring the correct business people. These tensions reveal that not every acquisition is a happy ending.
However, the founders who did join Anthropic expressed genuine excitement. “The choices were clear: we could build independently and work toward the same vision as two separate versions of it, or join forces with an incredible team and accelerate that vision into reality. The decision became an easy choice,” she said of joining Anthropic.
Meanwhile, Meta poaches Vercept co-founder and gets his expertise for their Superintelligence Lab. Despite leaving, Deitke expressed full confidence in Vercept’s continued success, calling the team “absolutely amazing” and telling GeekWire that the direction they are building remains “very exciting.” He publicly congratulated his former colleagues on social media, maintaining positive relationships despite going separate ways.
Market Reactions and Industry Implications
The immediate market response highlighted concerns about disruption. UiPath Inc. (NYSE: PATH) falls after Anthropic acquires Vercept to enhance Claude’s computer interaction+falls+after+Anthropic+acquires+Vercept+to+enhance+Claudes+computer+interaction/26058051.html) as investors worried that improved computer use capabilities threaten traditional robotic process automation companies.
A recent update to Claude Cowork caused a major selloff in enterprise software shares. In the days that followed the launch, software companies with products focused on those tasks lost billions of dollars in market value. When AI agents can automate software workflows directly, companies selling point solutions face an existential threat.
The impact of Anthropic Vercept acquisition extends to how we think about software development itself. The broader industry is converging on this idea. Gartner has forecast rapid enterprise GenAI adoption in the next few years, and McKinsey has estimated GenAI could unlock trillions in annual value, much of it from knowledge-work automation.
What This Means for the Future of Work
Let me share what keeps me up at night about these developments. Computer use AI doesn’t just automate repetitive tasks—it fundamentally changes what it means to interact with technology.
People are using Claude for increasingly complex work—writing and running code across entire repositories, synthesizing research from dozens of sources, and managing workflows that span multiple tools and teams. That means Claude can take on multi-step tasks in live applications, and solve problems impossible with code alone.
We’re approaching an inflection point. Industry predictions suggest 2026 will be the year of agentic AI, with enterprise adoption accelerating rapidly. The Anthropic acquires Vercept AI transaction positions them at the forefront of this shift.
But challenges remain. Vulnerabilities like jailbreaking or prompt injection may persist across frontier AI systems, including the beta computer use API. In some circumstances, Claude will follow commands found in content, sometimes even in conflict with the user’s instructions. For example, Claude instructions on webpages or contained in images may override instructions or cause Claude to make mistakes. Security concerns could slow adoption in sensitive industries.
The Talent War Continues: What’s Next
Meta poaches Vercept co-founder as part of a broader strategy. Meta has already brought on top researchers, including OpenAI’s Lucas Beyer, Apple’s Ruoming Pang, former Scale AI CEO Alexandr Wang, ex-GitHub CEO Nat Friedman, and Daniel Gross of Safe Superintelligence. They’re assembling what amounts to an AI dream team.
The compensation levels signal desperation masked as confidence. When companies pay quarter-billion-dollar packages for individual researchers, they’re not just hiring talent—they’re preventing competitors from having that talent. It’s defensive and offensive simultaneously.
Anthropic AI must now focus on retention. Having acquired Vercept’s team, they face the challenge of keeping them engaged and preventing further poaching. The agentic AI startup acquisition solved one problem but created another.
Practical Implications for Businesses
If you’re running a business today, here’s what the Anthropic Vercept deal means for you:
Software vendors face pressure to justify their existence. If an AI agent can navigate your existing tools and automate workflows without custom integrations, why would companies pay for specialized point solutions?
Knowledge workers should prepare for partnership with AI agents rather than replacement. The most valuable employees will be those who can effectively direct and oversee AI agents completing complex tasks.
Security teams need to develop new frameworks. Consider the following: a. Limiting computer use to trusted environments such as virtual machines or containers with minimal privileges b. Avoiding giving computer use access to sensitive accounts or data without strict oversight becomes essential as AI agents gain more autonomy.
Looking Ahead: The Next Phase
The Anthropic buys Vercept story represents one chapter in a much longer book. We’re witnessing the early days of a fundamental transformation in how humans and computers interact.
Vercept is the latest team we’ve brought into Anthropic, following the acquisition of Bun. Expect more acquisitions as leading AI companies race to assemble the capabilities needed for truly autonomous agents.
Meta continues building its Superintelligence Lab, poaching talent wherever they find it. The competition between Anthropic, Meta, OpenAI, Google, and others will only intensify.
Meanwhile, the impact of Anthropic Vercept acquisition will unfold over months and years. Computer use capabilities will improve, more applications will emerge, and we’ll collectively figure out what this technology should and shouldn’t do.
The ultimate winners might not be the companies with the best technology—but those who figure out how to deploy it responsibly while moving fast enough to maintain competitive advantage. That’s the razor’s edge where Anthropic now finds itself, armed with Vercept’s technology and most of its team, racing toward a future where AI agents don’t just chat—they work.
Frequently Asked Questions
What is the Anthropic Vercept deal about?
Anthropic acquired Seattle-based AI startup Vercept to enhance Claude’s computer use capabilities, bringing most of the founding team on board. The deal gives Anthropic access to Vercept’s proprietary desktop agent technology that allows AI to control computers like humans do, navigating applications and completing multi-step tasks without custom integrations.
How much did Meta pay to poach Vercept’s co-founder?
Meta reportedly offered Vercept co-founder Matt Deitke approximately $250 million over four years to join their Superintelligence Lab. This extraordinary compensation package reflects the intense competition for elite AI talent and Meta’s aggressive strategy to catch up in agentic AI development.
What is agentic AI and why does it matter?
Agentic AI refers to autonomous systems that can make decisions and complete multi-step tasks independently, rather than just responding to prompts. Unlike traditional chatbots, agentic AI can navigate software, automate workflows, and solve problems across multiple applications—potentially transforming knowledge work and enterprise automation.
What was Vercept’s technology before the acquisition?
Vercept developed Vy, a desktop agent that could control a remote MacBook to complete complex tasks across multiple applications. Their VyUI model achieved 92% accuracy on automation benchmarks, significantly outperforming competing models from OpenAI, Google, and even Anthropic’s own technology at understanding and interacting with user interfaces.
How does the Anthropic Vercept acquisition affect existing Vercept users?
Vercept is shutting down its Vy application within 30 days of the acquisition announcement. Existing users are being encouraged to migrate to Anthropic’s Claude tools as an alternative. This rapid sunset reflects Anthropic’s intention to integrate Vercept’s technology directly into Claude rather than maintaining separate products.
What are the broader implications for the AI industry?
The acquisition signals accelerating consolidation in agentic AI, with major players acquiring specialized startups and poaching individual researchers at unprecedented compensation levels. Traditional software automation companies face disruption as AI agents can potentially replace specialized tools, while enterprises gain access to more powerful automation capabilities through general-purpose AI systems.
Who are the key people joining Anthropic from Vercept?
Anthropic is bringing on Vercept co-founders Kiana Ehsani, Luca Weihs, and Ross Girshick, along with additional team members. These individuals bring deep expertise from the Allen Institute for AI and years of experience building AI systems that can perceive and interact with software interfaces. However, co-founder Matt Deitke joined Meta instead, and co-founder Oren Etzioni is not joining Anthropic either.
