If your team runs Amazon Q Developer, you already have a deadline whether you've looked at it yet or not. AWS blocked new Q Developer accounts and subscriptions on 15 May 2026, and support ends outright on 30 April 2027. Existing subscriptions can still add seats until then — you're not forced to move this quarter — but Java upgrades and .NET porting through the IDE plugins stop working the day support ends, full stop, with no replacement feature promised.
We modeled the real per-seat cost of the four realistic replacements, and the numbers are less obvious than the headlines suggest.
The official successor isn't a flat swap
AWS built Kiro as Q's replacement, and it's genuinely different in shape — an agentic IDE and CLI rather than an editor plugin. The pricing model changed too: Q Developer Pro was $19/user/month, flat. Kiro Pro is $20/month for 1,000 credits, with overages at $0.04 each. That's a predictable line item becoming a metered one. A developer running long agentic sessions can blow through 1,000 credits fast, and nobody can quote you a real per-seat number without first measuring your team's actual credit burn.
The three alternatives, and what you're actually paying for
GitHub Copilot Business ($19/seat/month) is the closest like-for-like: a flat per-seat price, plugins across every editor your team already uses, admin controls. If the goal is the least disruption and a predictable invoice, this is the default.
Cursor Teams ($40/seat/month) — twice Copilot's price — buys a stronger multi-file agent and better autocomplete, but your team switches editors to get it. That switching cost is the real number, not the $21/seat delta.
Claude Code Pro ($20/month) works alongside whatever editor people already have rather than replacing it. There's no published per-seat team tier, which is awkward for procurement but cheap for a small team.
At 50 seats, annualized: $11,400 for Copilot Business vs. $24,000 for Cursor Teams. Over three years that gap is roughly $38,000 — worth an actual afternoon of modeling rather than a Slack poll.
One correction to the "wait and see" instinct on Cursor
If you're weighing Cursor Teams, know that the SpaceX–Anysphere story has moved since it was first reported. It wasn't "expected to close in Q3 2026" — the $60B all-stock acquisition of Cursor's parent company actually closed on 14 August 2026, folding Cursor into a new "SpaceXAI" division. Pricing hasn't changed, but you'd be standardizing a team on a tool that just changed owners, not one mid-negotiation.
The timeline that isn't panic
You have about 20 months. Spend the rest of 2026 running Kiro Pro and Copilot Business side by side with a handful of volunteers, and actually record credit burn per developer — that single number decides whether Kiro is a $20/seat tool or a $35/seat tool for your team. Decide and migrate as one cohort in Q1 2027, and close out any in-flight Java/.NET transformations before the April 2027 cutoff, because that capability does not survive it.
Full breakdown — per-seat cost tables at 25/50/100 seats, the complete AWS announcement timeline, and the FAQ: https://devtoolsreview.com/best-for/amazon-q-end-of-support-migration/
Top comments (0)