Relay, an AI automation startup, is shutting down — and its team is moving over to Google's Chrome group, marking one more agentic-AI outfit that folded rather than found a path to a standalone business.
According to TechCrunch, Relay is closing operations as a company while its staff joins Chrome, one of Google's most actively developed AI battlegrounds right now. No financial terms, headcount, or founder statements were disclosed alongside the news.
The move fits a pattern that has become familiar in the AI industry over the past two years: a well-funded automation or agent startup runs out of runway or strategic options, and rather than a traditional acquisition or a hard shutdown, its core team resurfaces inside a large platform company weeks later.
Why Chrome, specifically
Chrome is not a neutral landing spot. Google has spent the past two years turning its browser into the delivery vehicle for its AI agent ambitions — bundling Gemini directly into the browser chrome, building AI Mode into search results rendered inside Chrome, and prototyping agentic browsing features that let a model click, fill forms, and complete multi-step tasks on a user's behalf. An "AI automation startup" — by definition, a company building tools that let software carry out multi-step actions on the web — sits squarely on that roadmap.
In our estimation, that overlap explains why Relay's team ended up inside Chrome rather than, say, Google DeepMind or Google Cloud — the destination suggests Google valued the team's experience building agentic workflows more than any specific product Relay had shipped.
An exit route that has become the norm
Relay is not the first agent-adjacent startup to end this way. The industry has seen a recurring shape to these deals:
- Inflection AI wound down as an independent product company in 2024, with its co-founders and much of its research staff moving to Microsoft.
- Character.AI's founders returned to Google the same year, alongside a licensing deal for the underlying technology.
- Adept AI's leadership and several researchers joined Amazon in 2024 after the startup scaled back its own agent product plans.
Each of these deals let the acquiring company absorb talent and know-how without taking on the acquired company's cap table, product liabilities, or customer contracts — a structure that has become the default off-ramp for AI startups that can't clear the bar for a traditional acquisition or an IPO.
What it means for builders in this space
For teams building on top of "AI automation" as a category — browser agents, workflow copilots, task-completion bots — Relay's shutdown is a reminder of two structural pressures:
- Platform gravity is strong. When the browser vendor itself ships agentic features natively, a standalone automation layer has to justify why it should sit on top of, rather than be replaced by, the platform.
- Talent is the exit, not the product. Increasingly, the value large platforms are paying for is the team that knows how to build reliable agent behavior, not the go-to-market or the customer base of the startup itself.
For founders, that changes how a company should be built from day one: a product that layers thin automation on top of the browser is more exposed to this kind of squeeze than one that owns a distinct data source, workflow, or customer relationship the platform has no incentive to replicate.
AiiN's takeaway
Relay's shutdown says less about Relay specifically than about where the AI automation category currently stands: still too early for most independent players to reach durable scale, and adjacent enough to a Big Tech roadmap that "shut down, then get hired" is a viable, even respectable, outcome for a founding team. Expect more of these quiet folds as browser vendors and model labs keep shipping the exact agentic features that startups like Relay were built to provide — and watch which automation companies differentiate on something a platform can't easily absorb.