Google acquires Mechanize team for $1.5B to boost autonomous AI agents

The AI industry is undergoing another seismic shift as Google completes a landmark talent acquisition deal with automation startup Mechanize, valued at more than US$1.5bn.

Rather than pursuing a full corporate merger, Google has structured the agreement as an acqui-hire, absorbing the startup's core team directly into Google DeepMind while leaving the original company intact.

Mechanize Co-Founder Tamay Besiroglu joins as a Research Scientist, accompanied by more than 12 senior engineers.

The deal marks a clear industry pivot away from simply enlarging language models, redirecting attention toward creating autonomous digital workers capable of independently managing complex enterprise tasks without continuous human oversight.

From passive assistants to autonomous workers

At the heart of Mechanize's work is the development of simulated digital environments — virtual workplaces where AI agents operate much like human professionals.

Unlike traditional AI systems that function as passive tools requiring constant prompting and supervision, Mechanize's agents are designed to work autonomously across extended timeframes.

Inside secure virtual sandboxes, these agents learn to manage complex multi-step workflows, coordinate with other digital workers and reprioritise their objectives in response to unexpected business challenges.

For leaders, this represents a fundamental transformation in how corporate teams may be structured in the near future.

These training environments allow intelligent systems to act, fail, receive automated feedback and retry tasks without exposing real-world corporate assets to risk.

The acquisition gives Google the engineering expertise required to build and scale these robust testing infrastructures, accelerating the deployment of reliable AI in highly regulated enterprise environments.

Automating labour could generate vast abundance, much higher standards of living and new goods and services that we can't even imagine today."

• Google's acqui-hire of Mechanize is valued at more than US$1.5bn
• Mechanize Co-Founder Tamay Besiroglu is joined by more than 12 senior engineers in the move to Google DeepMind
• Mechanize's acquisition valuation is more than three times its estimated worth of US$500m from early 2025
• US workers are paid approximately US$18tn per year in aggregate
• The global annual wage figure is more than three times the US total

Automation targets white-collar knowledge work

Mechanize was built with an explicit mission: the full automation of the modern economy, with a particular focus on white-collar roles across major global industries.

By developing AI agents capable of executing long-horizon tasks, the founding team envisions a wholesale restructuring of the traditional labour market.

As autonomous systems grow more capable of handling repetitive administrative and technical responsibilities, leaders must begin preparing to incorporate non-human agents directly into their organisational hierarchies.

A growing trend in tech talent acquisition

The Mechanize transaction reflects a broader pattern emerging across the technology sector. By licensing intellectual property and onboarding core personnel without acquiring the corporate entity outright, major technology firms can rapidly secure elite talent while sidestepping the antitrust scrutiny that accompanies full corporate takeovers.

The deal joins a series of similar high-profile acqui-hires executed by Google over the past two years, including strategic moves involving Windsurf and Character.AI.

At US$1.5bn, the Mechanize valuation exceeds three times its estimated worth of US$500m from early 2025 — a stark demonstration of the premium placed on specialised AI talent.

The economic case for full labour automation

The founders' vision extends well beyond productivity enhancement. They argue that automating routine work at scale could unlock transformative economic value.

"Automating labour could generate vast abundance, much higher standards of living and new goods and services that we can't even imagine today," says Tamay.

He adds: "Workers in the US are paid around US$18tn per year in aggregate; for the entire world, the number is over three times greater."

By integrating this team into Google DeepMind, Google positions itself at the forefront of the AI workplace revolution, challenging leaders to entirely reimagine the future of human work.

Google's key partners:

Intel: Google partners with Intel to power its cloud infrastructure using high-performance Xeon processors and integrate Gemini AI into semiconductor design workflows.

CEO: Lip-Bu Tan | Headquarters: Santa Clara, California, US

Arm: Google collaborates closely with Arm to co-develop and optimize custom cloud silicon, including Google’s custom Arm-based Axion central processing units.

CEO: Rene Haas | Headquarters: Cambridge, UK

Broadcom: Google relies heavily on Broadcom to co-design and manufacture its custom Tensor Processing Units (TPUs) for artificial intelligence workloads.

CEO: Hock Tan| Headquarters: Palo Alto, California, US