AI Tools That Shut Down in June 2026: The Zero-Casualty Report
June 2026 closed with no confirmed AI tool shutdowns, acqui-hires or wind-downs in the AI Tool Graveyard database — a flat month after eighteen casualties in the preceding year. This closed-book report explains the three-signal verification behind the zero, the four forces that produced a quiet month, and eight categories scored on consolidation risk for the second half of 2026, with a six-step vendor audit, guidance for founders and investors, and eight detailed FAQs.
Quick answer: which AI tools shut down in June 2026
None. June 2026 closed with no confirmed AI tool shutdowns, acqui-hires, or public wind-downs in the AI Tool Graveyard database. The month is closed and verified.
That is rare. In the twelve months ending June 2026 we logged eighteen notable casualties, and every single month in that window recorded at least one — until June. If you searched for "AI tools that shut down in June 2026" expecting a body count, the honest answer is zero. The useful answer is what a flat month actually tells you, and it is the rest of this report.
Below: how the zero was verified, why the month was quiet, which categories are carrying the pressure into the second half of the year, and the audit worth running while nothing is on fire.
How we verified a zero instead of assuming one
A quiet month is easy to fabricate by simply not looking hard enough. We publish a zero only when three independent signals all come back empty.
Signal one: automated product health. Every tool on the watch list gets a rolling probe: DNS resolution, TLS certificate validity, and homepage HTTP status. We also compare returned HTML against a stored fingerprint, so a parked domain or a "thanks for everything" splash page cannot pass as a healthy product. In June, every previously active tool returned a 200 with recognizable product markup.
Signal two: first-party notices. We watch vendor blogs, RSS feeds, status pages, changelogs, and official accounts for sunset language — "end of life", "sunsetting", "migrate your data by". No vendor we track published one in June 2026.
Signal three: trade reporting. We cross-read TechCrunch AI, Reuters technology, and Bloomberg Technology for talent absorptions and asset sales. Nothing AI-specific crossed our reporting bar, and nothing appeared in the Crunchbase record for the month either.
A tool enters the graveyard only when two of the three signals agree. Three empty signals is a genuine zero, not a coverage gap. The same pipeline output is visible publicly on the death-watch leaderboard and across the tools directory.
Why June 2026 was quiet
Four forces stacked. None of them is "the shakeout is over".
The fragile field is mostly gone already. The 2024–2025 consolidation wave cleared out the thinnest players. Rockset went to OpenAI. Multi went to OpenAI. Adept went to Amazon. Inflection's team went to Microsoft. What remains is a smaller, better-funded field with fewer monthly failure candidates. Fewer candidates, fewer funerals.
Announcement timing clusters elsewhere. Acquirers hold talent deals until they can fold them into an earnings narrative, and big-tech results land in late July. A silent June is often a loaded August.
Mid-year budget cycles freeze decisions. Enterprise renewals concentrate at quarter boundaries. Vendors on the edge push the hard call one quarter out rather than announce into a renewal window they might still win.
Runway extension replaced the clean shutdown. Since 2025, the standard response to a bad quarter is a down round, a bridge note, or a quiet headcount cut. Tools fade instead of dying loudly. We track that as abandoned rather than shut down — and abandoned is where the real user pain now lives. Kite and Peppertype.ai both went that way before their status changed officially.
The categories carrying the pressure
Quiet does not mean safe. We score every tracked category on four inputs: platform encroachment, defensible differentiation, funding pressure, and switching cost.
Standalone voice AI (85). The most exposed board on the chart. Play.ht was absorbed by Meta, and the remaining independents compete against speech features bundled free into frontier models. Our Play.ht alternatives ranking covers where its users actually went, and is Play.ht dead documents the zombie-product pattern that follows an acquisition. Murf AI faces the same squeeze from a different starting point.
AI wearables and pins (80). After the Humane collapse and the muted Rabbit R1 reception, ambient hardware still has no durable winner. Hardware burn is unforgiving, and the phone already in your pocket keeps improving at the same job. See best tools like Rabbit R1 for the software-first options.
Meeting notetakers (73). Fireflies.ai and Otter.ai sell a capability Zoom, Teams, and Google Meet now include by default. Our tools like Fireflies.ai comparison shows how thin the remaining moat is.
Autonomous coding agents (64). A real, growing category — but the independents are boxed between GitHub Copilot, Claude Code, and OpenAI's coding tooling. Devin AI is the reference case; Devin AI alternatives ranks who is genuinely shipping, and Windsurf shows how quickly a favourite can wobble.
AI presentation tools (60). Slide generation now appears natively inside Google Workspace and Microsoft 365. Tome already pivoted away from the consumer deck product.
Legal AI point tools (51). Casetext was absorbed by Thomson Reuters and the pattern is repeating around Harvey. Incumbent legal-research vendors have distribution a startup cannot buy. Harvey AI alternatives covers the practical shortlist.
AI writing assistants (47). Consolidating, not collapsing. QuillBot and Peppertype.ai show both ends of the range — one absorbed into a larger education suite, one quietly folded into a parent brand.
Frontier chat assistants (15). ChatGPT and Claude are the acquirers, not the targets. Risk here is competitive, not existential — which is exactly why the recurring "is ChatGPT dead" searches keep returning no.
What the year before June actually looked like
Eighteen casualties in twelve months fall into four repeatable patterns. Knowing the pattern tells you what warning you will get.
| Pattern | What you see first | Notice period | Typical example |
|---|---|---|---|
| Clean acqui-hire | Founder post about "joining forces" | 30–90 days | Rockset |
| Asset sale | New owner, unchanged marketing site | 0–30 days | Multi |
| Slow fade | Changelog stops, support lags | None | Kite |
| Hard shutdown | Billing stops, then a sunset email | 7–30 days | Forefront AI |
The slow fade is the dangerous one because it never generates a headline. There is no email, no migration window, and no press coverage — just a product that stops improving while you keep paying for it. Our monthly detection sweep exists specifically to catch that class.
The quiet-month audit: what to do with a flat June
A month with no casualties is the cheapest possible time to prepare for the next one. Six steps, roughly a day and a half of total effort across three people.
- Inventory every AI tool actually in use. Pull the list from your card statements and SSO logs, not from memory. Shadow subscriptions are where surprise migrations hurt.
- Check each vendor against public signals. Last changelog date, last funding event, whether the product still ships. Start with the leaderboard and the relevant category pages.
- Confirm the data-export clause. Find it in the contract before you need it. "We'll help you export" in a support chat is not a clause.
- Run one real export per vendor. Not a test of the button — an actual file you open and check. Broken exports are discovered at the worst possible moment.
- Name a documented backup per category. One named alternative, one owner, one rough cost. Our comparison pages exist for exactly this shortlist work.
- Diary the next review 30 days out. Vendor risk is a subscription, not a project.
If you only do one of these, do step four. Every migration horror story we have documented — from Play.ht to Peppertype.ai — starts with an export that nobody had tested.
What a quiet month means for founders
If you build an AI product, a flat month is not permission to relax. It is a window.
Watch the bundle, not the competitor. Independent tools rarely lose to a rival startup. They lose when a platform ships 80% of the feature for free. If your core value fits in a platform release note, assume it will.
Price for the value that survives bundling. Workflow depth, compliance posture, and data ownership survive. A thin wrapper around a model API does not.
Publish your durability. Buyers now check whether a vendor looks alive before signing. A visible changelog, a real status page, and a documented export path convert better than another feature page. That is a direct lesson from every tool in the graveyard.
What a quiet month means for investors
Zero shutdowns in a month is not a signal of sector health, and treating it as one is a mistake we see repeatedly.
Failure rate is a lagging indicator. Bridge financing and headcount cuts push the visible failure six to twelve months out, which means today's calm reflects last year's capital, not this year's fundamentals. The leading indicators are shipping velocity, gross-margin trend against inference costs, and net revenue retention in accounts where a platform has already bundled a competing feature.
For the sector view, Stanford HAI's AI Index remains the best public baseline on funding and adoption. Pair it with the pattern table above, which tells you what the exit will actually look like when it comes.
Method, scope, and corrections
We track shutdowns, acqui-hires, asset sales, and abandonments for AI products with meaningful public usage. We exclude private betas that never launched, feature deprecations inside a still-healthy product, and rebrands where the service continues unchanged.
Every entry carries a confidence score and a last-checked date, both visible on the tool's page. When a signal changes — a product returns, an acquisition completes, a sunset date slips — we update the record and note the change rather than quietly editing history. If you know something we have missed for June 2026, report it and we will verify it against the same three signals.
Next in the series: the July 2026 roundup, which shows whether the lull held.
Frequently Asked Questions
Which AI tools shut down in June 2026?
None. June 2026 closed with zero confirmed AI tool shutdowns, acqui-hires, or public wind-downs in the AI Tool Graveyard database, verified against automated product-health probes, first-party sunset notices, and trade reporting.
Does a zero-shutdown month mean the AI shakeout is over?
No. Failure rate is a lagging indicator. Bridge financing, down rounds, and headcount cuts push a visible failure six to twelve months into the future, so one flat month reflects last year's capital rather than this year's fundamentals.
How do you confirm that no AI tool shut down in a given month?
Three independent checks must all come back empty: an automated probe of DNS, TLS, HTTP status and page fingerprint for every tracked tool; a sweep of vendor blogs, status pages and changelogs for sunset language; and a cross-read of major technology press for acqui-hires and asset sales.
Which AI categories are most likely to see the next shutdown?
Standalone voice AI scores highest on our consolidation risk board, followed by AI wearables and meeting notetakers. All three sell capabilities that frontier models and video platforms now bundle for free, which erodes pricing power fastest.
What is the difference between an abandoned tool and a shut-down tool?
A shut-down tool has stopped serving users, usually with a sunset notice. An abandoned tool still runs but is no longer maintained: the changelog stops, support slows, and pricing freezes. Abandoned tools cause more user pain because there is no migration window.
How much warning do users usually get before an AI tool shuts down?
It depends on the pattern. Clean acqui-hires typically give thirty to ninety days. Asset sales give zero to thirty. Hard shutdowns give seven to thirty. Slow fades give none at all, which is why testing an export before you need it matters more than watching for an email.
What should I do during a month with no AI shutdowns?
Run the six-step audit: inventory the tools actually in use, check each vendor's public health signals, confirm the data-export clause in the contract, run one real export per vendor, name a documented backup per category, and diary the next review thirty days out.
Where can I follow AI tool shutdowns as they happen?
The death-watch leaderboard ranks tracked tools by risk signals and updates as probes run, and each monthly trend report closes the book on a period once every signal has been verified.