Google to unify AI coding tools under Antigravity

Antigravity 2.0, launched at Google IO on Tuesday, is the second iteration of Google’s agent-first development platform, and comes with a new desktop app, Antigravity CLI, expanded SDK capabilities, and deeper integration with the Gemini Enterprise Agent Platform. But along with its announcement came the news that Google is beginning to consolidate its existing tools under the Antigravity umbrella.

“Listening to your feedback made one thing clear: we can serve you best by pouring our energy into a single product built for today’s multi-agent reality,” the company wrote in a blog post. To do so, it said, “we’re unifying our efforts into Google Antigravity, our premier agent-first development platform, which includes a powerful server-side harness and a brand-new terminal experience: Antigravity CLI.”

Cleanup of overlapping tools could simplify procurement

While that transition doesn’t mean that Google is immediately shutting down Gemini CLI or Gemini Code Assist for paying enterprise customers, or that there is complete feature parity between the offerings, it signals a wider, sustained effort to unify AI coding assistants, CLIs, agents, and enterprise developer workflows into a single platform, according to analysts and experts.

“Google had too many overlapping tools: Code Assist, Gemini CLI, AI Studio, all doing similar things with no shared backend” said Advait Patel, senior site reliability engineer at Broadcom. “Antigravity is the cleanup. The bet is that the future is not autocomplete in your IDE. It is fleets of agents running refactors, infra changes, and reviews in parallel across desktop, terminal, SDK, and Google Cloud.”

And according to Bhupendra Chopra, chief revenue officer (CRO) at AI, data engineering, and migration consultancy firm Kanerika, the beginning of the unification of these tools will mean future simplification for enterprise decision-makers such as CIOs around procurement. “CIOs have been tracking three overlapping Google products with overlapping pricing, overlapping IAM models, and overlapping support contracts,” he said.

Patel also pointed out that one platform would actually solve the messy governance problem CIOs have been stuck with if they were subscribed to more than one Google tool or solution.

In a similar vein, Abhisekh Satapathy, principal analyst at Avasant, noted that the move is likely to reduce AI tool sprawl, which he sees as one of the biggest challenges presently plaguing CIOs.

Could ease integration hurdles

More broadly though, Satapathy sees the unification as part of Google’s effort to move beyond standalone AI coding assistants toward supporting the entire agentic software development lifecycle, positioning Antigravity as a foundational operating layer for AI-native engineering workflows.

“The disparate products (Gemini CLI, Code Assist) previously behaved as if they were adjacent capabilities. Antigravity moves them toward a shared execution layer where project context, execution history, and agent state persist across coding, testing, debugging, and deployment activities, rather than restarting task by task,” Satapathy said.

“For enterprise software teams, this should reduce integration overhead, duplicated tooling connections, and context switching across development workflows,” he added.

That broader platform strategy also puts Google in more direct competition with rival hyperscalers and LLM providers, including Microsoft, OpenAI, and Anthropic, all of which are increasingly positioning AI coding assistants as enterprise development platforms, although each of them is optimizing for a different gravity well.

“OpenAI is leaning on its model lead and Codex, now past two million weekly active users, to keep developers inside the ChatGPT and API ecosystem. Microsoft is using GitHub Copilot’s installed base and Azure’s enterprise contracts as the distribution wedge,” Chopra said.

But, he added, “Google’s real differentiator is the line connecting Antigravity to Gemini 3.5 Flash, AI Studio, the Gemini API, and the Gemini Enterprise Agent Platform. That stack is harder for rivals to replicate because it spans model, runtime, and managed infrastructure.”

Satapathy believes that distinction is precisely what could appeal to CIOs and enterprise decision-makers. “Small model improvements matter less if teams inherit additional systems, integrations, and support overhead,” he observed.

That emphasis on integrated architecture is also why Patel sees Microsoft, rather than standalone model providers or AWS, as Google’s most significant rival in the enterprise AI development market, since AWS has stronger infrastructure gravity than workflow gravity,.

Pairing integration strategy with pricing incentives

Beyond product integration, Google is also trying to make the Antigravity ecosystem more commercially attractive to enterprise users and developer teams.

The company said its new $100-per-month Google AI Ultra plan will provide five times higher Antigravity usage limits than the Google AI Pro tier, as well as offering temporary bonus credits for developers who exceed their quotas.

For Chopra, the change in pricing reflects Google’s understanding of “what serious agentic workloads actually consume” and its efforts to cater to that reality from a user perspective.

He was even more impressed that the hyperscaler simultaneously dropped the monthly price of its top Ultra tier from $250 to $200. “Google is trying to flatten the upper end and is pushing serious users toward higher Antigravity limits at lower per-dollar cost. The strategy is to make consumption-based usage feel cheaper at scale,” he said.

However, Chopra warned that CIOs evaluating the broader Antigravity strategy will also need to weigh it against the risks of tighter platform dependence and long-term vendor lock-in.

Echoing Chopra, Patel cautioned that CIOs should still “ask hard exit questions before committing at scale.”

Migration risks emerge

Some users, though, may not have much time to weigh those trade-offs.

Google said that starting June 18, 2026, Gemini CLI and Gemini Code Assist IDE extensions will stop serving requests for free individual users, as well as for subscribers on Google AI Pro and Ultra plans, with the company directing users toward Antigravity CLI instead.

The transition will also affect Gemini Code Assist for GitHub, where new installations for GitHub organizations will stop on the same date, before serving of requests is gradually phased out in the following weeks, it added.

For Patel, the short cut-off timeline poses migration risks. “Google has already said [there is] no one-to-one feature parity at launch between the old and new offerings,” he said. “The real risks are in CI/CD pipelines that shell out to Gemini commands, internal plugins that need to be rewritten as Antigravity plugins, and IAM bindings that need to be remapped.”

To prepare, developers should inventory every place Gemini CLI is used, prioritize automation paths, and run both tools in parallel for a few weeks before flipping the switch, Patel said.

Paul Chada, co-founder of agentic AI startup Doozer AI, also warned of a different risk for those who end up migrating: “The thing to flag is where the agent actually runs. The old setup ran on the developer’s machine. The new one runs on Google’s servers, which means your code leaves the building before the agent touches it.”

A brief reprieve for some enterprises

Enterprise customers using Gemini CLI or IDE extensions through Gemini Code Assist Standard or Enterprise licenses, or via Google Cloud integrations, however, will get a longer window to transition to Antigravity 2.0 and the new CLI.

These users will continue to receive support, newer Gemini model access, and ongoing updates for now, Google said, without providing a timeline for its phase out of support. For enterprise users interested in migrating to the new offerings, Antigravity CLI has been made available immediately within its cloud environments, the hyperscaler said, adding that it plans to soon provide migration documentation and video walkthroughs to help developers transition to the new platform.

This article originally appeared on InfoWorld.