# Google has absorbed Mechanize’s coding-eval team into DeepMind after $1.5B+ deal talks

Public profiles show Mechanize co-founder Tamay Besiroglu and more than a dozen former staff have moved into Google, many into AI midtraining. The final terms are undisclosed, but earlier talks reportedly valued the technology-and-talent package above $1.5 billion.

Google appears to have completed a talent-focused Mechanize deal: the startup still exists, but much of the team that builds coding-agent training environments and evaluations has moved into Google’s model-development work.

- Status: Active
- Published: 2026-09-12T21:18:16+12:00
- Updated: 2026-09-12T21:18:16+12:00
- Categories: Artificial Intelligence, Web Development, Indie Business, AI Models, AI Coding, Product Economics
- Tags: AI evaluations, Coding agents, Google DeepMind, M&A
- Canonical HTML: https://beyondthe.news/dossiers/google-mechanize-deepmind-coding-eval-team-talent-deal

## What changed

Google appears to have completed a talent-focused deal with Mechanize, the startup that builds training environments and evaluations for frontier coding agents. Business Insider reports that Mechanize co-founder and former CEO Tamay Besiroglu is now a research scientist at Google DeepMind and that more than a dozen former Mechanize employees have joined Google, with many working on model midtraining. The final financial terms have not been disclosed. Earlier reporting said Google was discussing a package worth more than $1.5 billion for Mechanize technology and talent. Mechanize itself still has an active website and hiring pages, and its former chief of staff Guive Assadi is now described as CEO in public profiles, so this is not a conventional whole-company acquisition.

## Why it matters

Mechanize is not primarily a consumer coding assistant. Its product is the harder-to-see layer behind coding models: environments, long-horizon software-engineering tasks, graders and reinforcement-learning signals. Moving much of that team into Google suggests that coding-agent progress is increasingly constrained by the quality of training tasks and evaluations, not just model scale. For small AI infrastructure companies, the reported economics are also striking: the market can value a compact team that knows how to make frontier coding systems measurably better far above normal SaaS multiples, even when the buyer does not acquire the entire company.

## The people moved; the company did not simply disappear

Business Insider reports that Tamay Besiroglu has joined Google DeepMind as a research scientist and that more than a dozen former Mechanize employees have moved to Google, many into midtraining work. Google and Besiroglu declined to comment on the final deal terms. Mechanize’s public site remains live and still describes the company as building coding-agent environments and evaluations, which makes the transaction look more like a talent-and-technology arrangement than a standard acquisition.

## Mechanize works on the training substrate behind coding agents

Mechanize says it builds environments in which models perform real software-engineering work such as implementing features, deploying applications and debugging unfamiliar codebases. Graders score those attempts and the resulting signals can be used for reinforcement learning and evaluation. One public example, GBA Eval, asks an agent to build a Game Boy Advance emulator in Rust over a long task horizon. That expertise maps directly onto model midtraining and coding-agent improvement.

## The reported price is about scarce know-how, not normal SaaS revenue

Earlier reporting put the talks at more than $1.5 billion for Mechanize technology and talent, though the final amount is still undisclosed. Mechanize had reportedly raised only $9.1 million earlier in 2026 at a $500 million valuation. Even without a confirmed closing price, the gap illustrates how aggressively frontier labs are pricing small teams that can create useful training environments, evaluation systems and feedback loops for coding models.

## Google has used this transaction shape before

The structure resembles other AI talent-and-licensing deals where a large lab hires key staff and gains technology access without buying the whole legal entity. That can move scarce researchers and infrastructure into a model lab faster than a conventional acquisition, while leaving a smaller operating company behind. The exact Mechanize licensing terms have not been publicly confirmed, so the durable value Google receives beyond the staff moves remains uncertain.

## Key details

- Mechanize co-founder and former CEO Tamay Besiroglu is reported to have joined Google DeepMind as a research scientist.
- More than a dozen former Mechanize employees are reported to have joined Google, with many working on AI model midtraining.
- Earlier reporting said the contemplated technology-and-talent package was worth more than $1.5 billion, but final terms remain undisclosed.
- Mechanize’s public site says it builds environments and evaluations for frontier coding agents, including long-horizon software-engineering tasks and graders used for reinforcement-learning and evaluation signals.
- Mechanize remains publicly active rather than presenting itself as fully acquired or shut down.

## Builder takeaways

- Watch the training-environment and evaluation layer, not only headline model releases. Better tasks, graders and reinforcement-learning infrastructure can be a major source of coding-agent capability gains.
- For teams building agent infrastructure, defensibility may sit in proprietary task environments, evaluation design and feedback loops rather than in another chat interface.
- Do not treat the reported $1.5B+ figure as a confirmed acquisition price. The useful signal is that Google moved a meaningful part of the team into DeepMind after talks reportedly valued the technology-and-talent package at that level.
- If you depend on a small AI infrastructure vendor, talent-focused deals are an operational risk even when the vendor itself remains alive; key people and roadmap momentum can move without a normal acquisition.

## What to watch

- Whether Google or Mechanize publicly confirms the final transaction structure and licensing terms.
- Whether Mechanize continues independently under its new leadership and what products or research it retains.
- Whether future Google coding-model releases explicitly use Mechanize-style environments, benchmarks or midtraining methods.
- Whether regulators increase scrutiny of talent-and-licensing transactions that transfer most of an AI startup’s technical team without a conventional acquisition.

## Uncertainties

- The final transaction value has not been disclosed; the $1.5 billion-plus number comes from earlier reporting on negotiations.
- Google and Besiroglu have not publicly detailed the technology licence, compensation structure or exact number of employees transferred.
- Public profiles establish personnel moves but do not by themselves reveal how much Mechanize intellectual property or infrastructure Google obtained.

## Timeline

- **2026-04-24 — Mechanize announces seed funding:** Mechanize reportedly raises $9.1 million at a $500 million valuation while focusing on environments and evaluations for coding agents.
- **2026-08-05 — Google-Mechanize talks are reported:** Reporting says Google is discussing a technology-and-talent deal valued above $1.5 billion.
- **2026-09-11 — Public profiles show the team move has happened:** Business Insider reports that Besiroglu and more than a dozen former Mechanize employees have joined Google, many in model midtraining roles.

## Sources

- [Google just completed its $1.5 billion-plus talent deal for AI startup Mechanize](https://www.businessinsider.com/google-completes-deal-for-ai-agents-startup-mechanize-2026-9) — Business Insider · independent reporting · 2026-09-11T00:00:00+12:00. Reports the completed personnel moves into Google/DeepMind, the midtraining roles, prior $1.5B+ negotiations and undisclosed final terms.
- [Mechanize, Inc.](https://www.mechanize.work/) — Mechanize · official company site. Primary description of Mechanize’s work building coding-agent environments, graders and evaluation/training signals; also shows the company remains publicly active.