My inbox blinked with the Bloomberg tip and my Slack filled with speculation. I read the anonymous sourcing and felt the deal wobble. Now Bloomberg says Anthropic walked away.
I’ll be direct: you should care because this was not a routine M&A whisper. I follow these patterns, and when a company like Anthropic chases a firm that doesn’t look like them, the motives tell a story about strategy, cash, and timing.
A Bloomberg tip landed in the queue and the rumor spread — what actually happened
Bloomberg reported that Anthropic weighed buying Decart for $6 billion (≈€5.5 billion) and then, according to anonymous sources, walked away. I called that initial piece a classic anonymous-sourced negotiation signal; you’ve seen that breed of reporting before. The latest readout says due diligence happened, offers may have been exchanged, but the deal stalled.
Did Anthropic buy Decart?
No. According to Bloomberg’s sources, Anthropic “performed due diligence on Decart, but ultimately walked away,” though both companies may still talk about working together. The rumor landed in the public square like a tossed match — loud and brief.
I watched the tech demos — why the fit looked strange at first glance
Decart’s demos are flashy: Lucy edits faces in real-time video and Oasis turns images into playable first-person worlds. I tried both demos; Lucy can replace a streamer’s face with a high-resolution anime avatar, and Oasis spins images into explorable spaces.
That spectacle is why the deal made headlines: Anthropic is an AI-model company spending heavily on compute. The likely prize was DOS, Decart’s so-called Decart Optimization Stack, which claims to shrink the friction between AI hardware and software and trim compute bills. In plain terms, Decart’s optimization stack is a Swiss Army knife for video AI — useful, compact, and full of practical tools.
Why did Anthropic walk away from Decart?
There are three plausible forces I’d put on the table. First, price and valuation: a $6 billion (~€5.5 billion) price tag reshapes capital plans. Second, integration risk: marrying an experimental, demo-driven consumer play with a model-focused infra shop can be tortuous. Third, timing: Anthropic is rumored to be eyeing an IPO around Halloween, according to Reuters, and that pushes a bias toward conservative moves.
I checked the incentives and the people — what this says about strategy
Anthropic spends vast sums on compute and is building toward a public market moment. You should see the math: shaving even a fraction of inference cost scales directly to margins. That’s where DOS would matter. But executives also have to answer how an acquisition would sit with investors watching an IPO calendar.
Bloomberg’s framing — anonymous sources plus a mouthpiece that the two companies could still collaborate — fits a pattern I’ve tracked: negotiations leak to build leverage, then parties explore partnership paths after a full buyout proves messy.
What is Decart’s Lucy and Oasis?
Lucy: real-time face and scene editing for video streams. Oasis: image-to-interactive-space generation that looks like simplified first-person games. Both are demo-first, attention-driven products that can create massive user moments but also raise content-moderation and rights questions fast.
If you’re watching the AI industry, this episode matters more for what it signals than for the check that didn’t clear. Anthropic’s apparent retreat suggests it prefers a lower-risk route before an IPO and might chase partnerships, licensing, or narrower IP buys instead of a headline-grabbing acquisition.
Bloomberg did the thing anonymous-sourced outlets do: it nudged market attention and then reported the fallout. You and I both know that in tech M&A the noise often maps to bargaining chips, not final deals. So where does that leave Decart, Anthropic, and the broader market for optimization tools — are we looking at a pause or the start of a new, quieter arms race?