AI tools that died (or got acquired) in 2025–2026

Our AI Graveyard currently contains 196 entries — 95 products that shut down outright, 101 that were acquired and folded into something bigger. Most of those moves happened in the last 18 months.

We keep the Graveyard for two reasons. First: if you used a tool that quietly disappeared, this is the one place online that will tell you where it went. Second: a directory that only celebrates wins isn't a directory, it's a sales brochure. So here's the honest record of 2025–2026.

The big shutdowns

Builder.ai (May 2025) — bankruptcy plus reporting on fabricated customer claims. The cautionary tale of the cohort. We unpublished the listing rather than leave it as a graveyard entry; the brand is poisoned.

Crowdfire (May 2025) — couldn't survive the AI-native social tools wave. A reminder that the social-media-management category got disrupted as fast as anything in 2025.

Sora original (December 2025) — moved to the Graveyard when OpenAI replaced it with Sora 2. Same brand, fresh model, but the original product — and its specific capability set — is genuinely gone.

Bard (February 2024) — rebranded to Gemini. Technically not dead, but the Bard name and product surface ended; we list it for searchers who remember the original.

The shutdown pattern of the year: products that didn't have a defensible position when ChatGPT, Claude, and Gemini consolidated the top.

The big acquisitions

The bigger story of 2025–2026 isn't shutdowns — it's consolidation. Enterprise platforms bought capabilities they couldn't build fast enough, and a wave of mid-tier AI products got rolled up.

The notable enterprise rolls:

• Moveworks → ServiceNow ($2.85B, December 2025)
• Streamlit → Snowflake ($800M, originally 2022, fully integrated 2024)
• Exscientia → Recursion ($688M, August 2024)
• Protect AI → Palo Alto Networks ($500M, August 2025)
• Drift → Salesloft ($500M, February 2024)
• Deci AI → NVIDIA ($300M, May 2024)
• Paperspace → DigitalOcean ($111M, July 2023)
• Flatiron Health → Roche ($1.9B)

The quieter folds (these don't get press releases but they matter for buyers who relied on these products):

• Adept → Amazon (acqui-hire, June 2024)
• Reclaim → Dropbox (August 2024)
• MarketMuse → Siteimprove (2024)
• Wonder Dynamics → Autodesk (May 2024)
• Magnific → Freepik (May 2024)
• Mokker → Photoroom (June 2024)
• Hazy → SAS (June 2024)
• EigenTech → Cohere (September 2024)
• Clipdrop → Jasper (February 2024)
• Aporia → Coralogix (January 2025)
• Pocus → Apollo.io (March 2026)
• Langfuse → ClickHouse (December 2025)
• Weights & Biases → CoreWeave (May 2025)
• Limitless → Meta (December 2025)
• Olive → Waystar (2023)
• Zyro → Hostinger (2023)
• Galileo AI → Google
• Preligens → Safran ($XXM, December 2023)

The rebrands worth knowing

A handful of products didn't die or get acquired — they rebranded hard enough that you might not find them under the old name:

• Codeium → Windsurf
• CodiumAI → Qodo
• Phrasee → Jacquard
• Voicify AI → Jammable
• Foundry → Mithril
• Synthetaic → RAIC Labs

If you're searching for the old name and hitting dead ends, that's where they went.

What this tells us about the market

A few patterns are worth naming:

Consolidation accelerated. 2024 was an "everyone raises a round" year. 2025–2026 was an "everyone gets bought" year. Buyers should assume that mid-tier AI products may not exist as standalone offerings 18 months from now.

The voice / avatar / image-gen layer compressed. Mokker into Photoroom, Magnific into Freepik, Wonder Dynamics into Autodesk, Clipdrop into Jasper — the creative tooling layer is now mostly owned by a few aggregators.

Vertical AI got absorbed by enterprise platforms. ServiceNow, Salesloft, Apollo, Dropbox, DigitalOcean — the buyers were enterprise SaaS incumbents protecting their flank, not other AI startups.

Fraud catches up. The Builder.ai case, the 11x AI customer-claim reporting, the Civitai payments shutdown — buyers got more sophisticated about the difference between "AI-powered" and "actually working."

Lessons for buyers picking tools today

A few practical filters that come out of looking at the graveyard:

1. Pick tools with named customers, not just funding. A round closing tells you a VC believed; a customer logo tells you somebody paid.
2. Check the lifecycle field on a directory before committing. Half the AI tools we list a year ago are now in the Graveyard. The other half, by definition, made it.
3. Be sceptical of tools without independent reviews. G2 with 100+ real reviews is signal. A landing page testimonial isn't.
4. Watch for the bigger-buyer pattern. If a tool's category has been hit by 3+ acquisitions in 18 months, the standalone product is unlikely to stay standalone.

Where to go from here

• Browse the Graveyard — full searchable list with filters by year and cause.
• The Top 100 — our editorial picks of tools that look likely to still be there next year.
• How we review tools — the methodology behind every rating in our directory.

— The ToolDirectory.AI editorial team