Amazon Q Developer Is Shutting Down: What Your Team Moves To (2026)

If your engineering team runs Amazon Q Developer, you have a migration on a deadline. This page is the decision, not the news: what AWS actually announced, what it costs to move, and what the successor’s pricing model changes about your budget.

What AWS announced, with dates

From the Amazon Q Developer end-of-support announcement, published 30 April 2026:

Date	What happens
15 May 2026	New Q Developer accounts and new subscriptions blocked. Existing subscriptions can still add users.
29 May 2026	Opus 4.6 removed from Q Developer Pro. Newer models are Kiro-only.
30 April 2027	End of support. IDE plugins stop being supported; Java upgrades and .NET porting stop being accessible.

Two details matter more than the headline.

Existing subscriptions can still add seats. The May 2026 cutoff blocks new accounts and subscriptions, not new users on a subscription you already have. If you are already paying, you can keep onboarding people — you are not forced to move this quarter.

Code transformation dies with it. If you use Q for Java language upgrades or .NET porting, those run through the IDE plugins and require a paid subscription. After 30 April 2027 they are gone, and no successor feature is promised. Finish in-progress transformations before that date.

What happens if you do nothing

Nothing, until April 2027 — then your team loses tooling on a fixed date with no extension. The real risk is not the deadline; it is that you spend 20 months building workflows on something with a published expiry, and pay the retraining cost anyway at a worse moment.

Your four real options

Kiro is the official path, but it is not the only one, and “official” is not a reason on its own.

Kiro — the successor AWS names

AWS built Kiro as the replacement and is migrating customers to it. It is an agentic IDE and CLI rather than a plugin, which is the point: AWS’s stated view is that plugin-shaped assistants were the wrong shape.

Pricing, from kiro.dev/pricing as of August 2026:

Plan	Monthly	Included
Free	$0	50 credits
Pro	$20	1,000 credits
Pro+	$40	2,000 credits
Pro Max	$100	5,000 credits
Power	$200	10,000 credits
Enterprise	Custom	SSO, centralised billing, usage analytics

Read that table as a budget owner, not a developer. Q Developer Pro was $19 per user per month, flat. Kiro Pro is $20 per user per month for 1,000 credits, with overages at $0.04 each. A developer who runs long agentic tasks can exceed 1,000 credits, and at $0.04 the difference between a light and heavy month is not small. You are trading a predictable line item for a metered one.

Before committing a team, run a month on Pro and measure actual credit burn per developer. That number, not the sticker price, is your real per-seat cost.

GitHub Copilot Business — the closest like-for-like

At $19 per seat per month, Copilot Business is the nearest thing to what you already had: a flat per-seat price, plugins in the IDEs your team already uses, and admin controls. If your priority is replacing Q with the least disruption and a predictable invoice, this is the default answer.

It also runs in every editor your team is likely using — VS Code, JetBrains, Visual Studio, Neovim, Eclipse — which matters because Q was a plugin and Kiro is a separate environment.

Try GitHub Copilot

Cursor Teams — if you want the agent, not the plugin

At $40 per seat per month, Cursor is twice Copilot Business. What you buy is a stronger multi-file agent and a better autocomplete, in an editor your team has to actually switch to. That switching cost is the real number, not the $21/seat difference.

Worth knowing before you standardise on it: SpaceX agreed to acquire Cursor’s parent Anysphere for $60B in an all-stock deal expected to close in Q3 2026. Pricing is unchanged so far, but a tool you are about to build team workflows around is mid-acquisition.

Claude Code — if your team lives in the terminal

$20/month via Pro, and it works alongside whatever editor people already use rather than replacing it. There is no per-seat team plan in the published pricing, which makes it awkward for procurement but cheap for a small team. Strongest of the four at whole-task work; weakest at inline completion.

What it actually costs

Per-seat monthly, at the plan a team would realistically buy:

Team size	Kiro Pro	Copilot Business	Cursor Teams	Claude Code Pro
25	$500 + overages	$475	$1,000	$500
50	$1,000 + overages	$950	$2,000	$1,000
100	$2,000 + overages	$1,900	$4,000	$2,000

Annualised at 50 seats that is $11,400 for Copilot Business against $24,000 for Cursor Teams. The gap over three years is roughly $38,000 — enough that the decision deserves an afternoon rather than a Slack poll.

Kiro’s column carries ”+ overages” because it genuinely cannot be quoted without your team’s credit burn. Anyone who quotes you a flat Kiro number has not read the pricing page.

How the four compare

A migration timeline that is not panic

You have about 20 months. Spread it:

Now → end of 2026. Measure. Run Kiro Pro and Copilot Business side by side with five volunteers for one month. Record actual Kiro credit consumption per developer. That single number decides whether Kiro is $20/seat or $35/seat for you.

Q1 2027. Decide and buy. Move the team in one cohort rather than trickling, so you support one workflow rather than two.

Before 30 April 2027. Finish any Java or .NET transformations still running through Q. That capability does not survive the date.

Frequently asked questions

---

Prices verified August 2026 against AWS’s own announcement and kiro.dev/pricing. We earn nothing from any link on this page and are not enrolled in an affiliate programme for any tool here — full disclosure.

Rolling this out across a team? Our $299 tool audit models your actual per-seat cost at your headcount, including Kiro credit burn, and gives you a written recommendation you can take to finance.