Amazon in Talks to Buy AI Startup Decart for $7 Billion, Report Says
The discussions would deepen an existing relationship between AWS and Decart, whose software is designed to make AI workloads run efficiently across competing chips
Topics
News
- Fired OpenAI Researchers Warn Dismissals Could Chill Safety Work
- ARTEX Developer Closes AI Hacking Tool After Bank Attacks
- AI Redraws Global Race for Quant Talent as Coding Tools Lower Entry Barriers
- AI Dispatch | India Draws Billions Into AI
- AI Bots Now Make Up Half of Traffic to US Fed’s FRED Database
- Musk Blames ‘Oligarchs’ For Delay in Starlink’s India launch
Amazon is in advanced talks to acquire artificial-intelligence startup Decart for about $7 billion, according to the Wall Street Journal, in what would be one of the technology group’s largest acquisitions and another push to strengthen its position in AI infrastructure.
The discussions have not been finalized and could still fall apart or draw competing bidders, the Journal reported.
Decart develops software and AI models aimed at running demanding workloads with lower latency and across different types of processors, including Nvidia GPUs, Google TPUs and Amazon Web Services’ Trainium chips.
The company describes its Decart Optimization Stack as a layer that can extract more performance from different AI accelerators rather than tying developers to one hardware platform.
That technology is closely aligned with Amazon’s efforts to build an alternative to Nvidia’s dominant AI-computing ecosystem.
AWS has spent years developing its Trainium family of AI accelerators. Its Trainium3 chips, launched last year, are designed for training and inference workloads including large language models, reasoning systems and real-time generative video.
Decart is already a prominent Trainium customer.
AWS says Decart’s real-time video models achieved as much as four times the frame throughput and twice the cost efficiency of leading GPUs in early Trainium testing, while latency fell from about 40 milliseconds to 10 milliseconds.
Decart has also attracted substantial venture funding.
The company announced a $300 million round in May led by Radical Ventures, after raising $100 million in August 2025 and an earlier $21 million seed round. Its May financing valued the business at close to $4 billion, according to the Journal.
Decart’s products stretch beyond chip optimization. Its Lucy models generate and alter video in real time, while Oasis is designed to create interactive simulated environments that can be used for robotics, autonomous vehicles, manufacturing and other forms of “physical AI.”
The potential acquisition would also give Amazon more control over software that can make its own chips competitive with Nvidia hardware.
Amazon has been trying to reduce the cost of AI computing for AWS customers while limiting its dependence on Nvidia. Trainium has become a central part of that strategy, with customers including Anthropic, OpenAI and Decart using the chips for AI workloads.
For Decart, an Amazon takeover would follow earlier interest from another major AI company.
Anthropic explored buying the startup earlier this year in a separate transaction before those discussions ended. Amazon’s talks are therefore a new potential deal rather than a continuation of the earlier approach.
The companies have not publicly confirmed the current negotiations.
