JetBrains Air Governance: Cost Control Pitch Meets TCO Reality

JetBrains wants to be the company that keeps AI coding spend under control, even when the agents doing the coding come from someone else. Its new Air platform, announced by CEO Kirill Skrygan on September 22, is built on that premise.1 Independent cost comparisons, however, suggest JetBrains' own AI stack has budget questions to answer before it can credibly sell cost discipline to others.

What JetBrains launched

Air combines three layers under one brand. Air in JetBrains IDEs lets developers run and review coding agents. Air Teams handles shared projects and automated workflows. Air Governance provides organization-wide oversight.1 The notable design choice is openness. Through the Agent Client Protocol (ACP), Air supports rival coding agents, so customers do not have to commit to JetBrains' own agent, Junie, to use it.1

The governance layer has an internal origin. JetBrains has described Air Governance as a commercial version of Central CLI, a tool it built to manage its own AI spending after disclosing a cost problem it characterized as roughly tenfold.3 The product extends cost visibility, policy enforcement and audit controls across providers to organizations running terminal agents such as Claude Code and Codex.3 In effect, the company is selling the solution it built for itself to anyone with the same assortment of tools.

The bet against lock-in

The strategy depends on a specific view of how the market will develop. JetBrains says most of its developers use three to five AI tools each month, and that it chose not to restrict that choice. It contrasts this with Accenture and Uber, which it says took a more restrictive approach.3 The wager is that governance infrastructure will be worth more than lock-in as agentic tooling fragments, particularly in the terminal.3

There is evidence that terminal agents are doing real work. An OpenAI writeup reports that Warp's agents now co-author about 90% of Warp's internal pull requests.3 If that level of adoption spreads, organizations will face more agents, more model providers and more invoices. That is the situation Air Governance is designed to manage.

The positioning is sensible. As RuntimeWire framed it, Air gives JetBrains' code intelligence a place in a multi-vendor workflow for reviewing and governing AI-written software, even when the customer chooses a competitor's agent.1 That is a more defensible position than trying to win a crowded agent market head-on.

Where the cost story gets complicated

The difficulty is that JetBrains is selling cost governance while its own AI offering is not clearly the cheaper option. A 2026 enterprise comparison from Augment Code puts Amazon Q Developer at $3,420 per year for a 15-developer team, all-inclusive, and $4,560 for 20 developers.2 JetBrains AI looks cheaper at $1,500 to $2,000 per year for 15 developers. That figure covers only the AI subscription, however. Teams also need IDE licenses, which the analysis estimates at an additional $2,235 to more than $9,735.2 The same comparison warns that JetBrains AI credits can run out quickly.2

The comparison does credit JetBrains with real strengths:

• Native integration across 11 IDEs
• A choice of Claude, GPT, Gemini and Grok models
• Confirmed SOC 2 Type II certification
• On-premises deployment through its AI Enterprise tier, which Amazon Q does not offer2

It also notes limits. JetBrains AI relies on manual context attachment and RAG rather than automated cross-repository indexing.2

These sources describe different products and should not be treated as a direct contradiction. Augment Code is assessing JetBrains AI Assistant, not Air Governance, and it is itself a vendor in the AI coding market. Still, the two stories overlap. A buyer evaluating JetBrains as a cost-governance partner will reasonably look at how predictable its own pricing is, and credit-based consumption combined with separate license fees is the kind of complexity governance tools are supposed to reduce.

The reading

Air Governance is probably JetBrains' strongest strategic move in AI so far. It turns a vendor-neutral position into something customers can buy, and it rests on a credible internal story: the company had the problem and built a fix. Supporting rival agents through ACP shows real confidence that orchestration and oversight will outlast any single agent.

The cost-governance pitch will work best if it stays clearly separate from JetBrains' own AI subscriptions. Enterprises will want neutral visibility into spending across Claude Code, Codex and others, and they will be skeptical if the governance layer appears to steer them toward JetBrains' credit-metered tools. If JetBrains can show that Air Governance reduces total spend, including on its own products, the pitch holds up. If not, the outside TCO figures will continue to undercut it.