Listen Labs scrapped $125M for Salesforce talks

Listen Labs signed a $125 million term sheet and then walked away from it. The AI research startup had a Series C led by Menlo Ventures at a $1.5 billion valuation, papered and ready. It killed the round to keep talking to Salesforce about an acquisition at roughly $2 billion. Founders almost never do this — a signed term sheet is treated as a handshake in venture. Listen Labs decided the exit was worth the reputational hit. If you run automated customer research on their platform, that decision is now yours to plan around.

What actually happened

TechCrunch reported on September 9 that Listen Labs abandoned the signed Menlo-led round while Salesforce discussions were live. The talks are not closed and may not produce a deal.

The company was founded in 2023 by Florian Jüngermann and Alfred Wahlforss. It uses voice AI to run customer interviews at volume — the qualitative research work that used to require a moderator and a calendar. It raised a $69 million Series B at a $500 million valuation in January 2026, led by Ribbit Capital with Sequoia, Conviction, and Pear VC. Annualized revenue is around $30 million. Customers include Microsoft, Canva, Anthropic, and Sweetgreen.

For scale: competitor Simile raised a $200 million Series B at a $2 billion valuation in July 2026 on roughly $10 million annualized. Listen Labs does about three times the revenue at a lower headline number. That gap is why a $2 billion offer looked better than another round.

Why AI vendor acquisition risk matters for your business

A tool at $30 million ARR with Microsoft and Anthropic on the logo wall reads as safe. It is not a comment on the product — it is a comment on the category. Research tooling, CRM-adjacent AI, and anything that produces structured customer data is exactly what the big platforms are shopping for right now.

Three things usually change after an acquisition like this: pricing moves onto the acquirer's per-seat model, the standalone tier gets a sunset date, and the roadmap starts serving the acquirer's install base instead of yours. None of that is bad faith. It is just what integration looks like.

So do the boring work before it happens. Know how you get your data out — not "there's an export button," but a tested export you have actually run and can read. Know what breaks if the API changes. Know what the replacement costs and how long a migration takes. If you can answer those three today, an acquisition is a scheduling problem. If you can't, it's an outage with a press release attached.

Sources: TechCrunch.