Anthropic in Talks to Buy Decart AI for $6 Billion: Why the Deal Matters for Claude

Anthropic is reportedly in talks to acquire AI startup Decart for about $6 billion, a deal that could give Claude a major boost in computing efficiency, real-time AI and next-generation world models.

The artificial intelligence race is entering a new phase. The biggest AI companies are no longer competing only over who can build the smartest model. They are also racing to secure the technology, computing infrastructure and engineering talent needed to make those models faster, cheaper and more capable.

That is what makes Anthropic’s reported negotiations to acquire Decart AI particularly significant.

According to Reuters, Anthropic, the company behind Claude, is in discussions to buy Decart, an AI startup backed by Nvidia. Bloomberg first reported that the potential transaction could be worth approximately $6 billion. The talks are reportedly still ongoing, meaning there is no guarantee that a deal will ultimately be completed.

If completed, however, the acquisition would represent a major strategic move for Anthropic and potentially one of the most consequential AI deals of 2026.

What Is Decart AI?

Founded in 2023, Decart has built its business around a problem that has become increasingly important as AI models become more sophisticated: how to make AI computing more efficient.

The company operates across multiple layers of the AI technology stack. Its work includes software designed to optimize AI workloads across different processors as well as generative video and real-time “world models.” Decart describes itself as a frontier AI lab focused on building real-time world models and efficient AI systems.

One of its key technologies is the Decart Optimization Stack (DOS). The software is designed to help AI developers run models efficiently across hardware from companies including Nvidia, Amazon and Google, potentially reducing the friction and cost involved in moving AI workloads between different computing platforms.

That capability could be especially valuable to Anthropic.

As AI adoption accelerates, computing capacity has become one of the industry’s biggest constraints. Training frontier models requires enormous amounts of compute, but serving millions of users and AI agents can be just as demanding.

Improving efficiency therefore isn’t simply a technical advantage. It can directly affect the economics of an AI company.

Why Anthropic Would Want Decart

At first glance, spending roughly $6 billion on a startup known partly for infrastructure optimization might seem surprising.

Anthropic already has access to enormous computing resources through partnerships with major cloud and chip companies. Amazon, for example, has invested heavily in Anthropic and works with the company on AI infrastructure, including access to Amazon’s Trainium chips.

But Anthropic’s challenge is changing.

The company says demand for Claude is growing rapidly. In May, Anthropic announced a $65 billion Series H funding round at a $965 billion post-money valuation, saying the money would be used to expand compute capacity, advance research and scale Claude products and partnerships. Anthropic also said its run-rate revenue had crossed $47 billion earlier that month.

The implication is straightforward: the more successful Claude becomes, the more computing power Anthropic needs.

Decart could potentially help Anthropic attack that problem from the software side.

Instead of simply buying more hardware, better optimization can allow existing hardware to do more work. That could translate into lower inference costs, improved performance and greater flexibility when Anthropic deploys Claude across different computing environments.

The Chip-Optimization Angle Could Be the Biggest Prize

Decart’s infrastructure technology may be particularly attractive because AI companies are increasingly looking beyond a single type of accelerator.

Nvidia remains the dominant force in AI chips, but Amazon, Google and other companies are developing their own alternatives. Switching AI workloads between these platforms can be complicated because different processors require different software optimization.

Decart’s DOS technology is designed to make that process easier. The company has said its technology can improve the efficiency of AI models across Nvidia, Amazon and Google hardware.

For Anthropic, that could provide strategic leverage.

Rather than being overly dependent on one hardware ecosystem, Anthropic could potentially optimize Claude for a wider range of processors and use whichever infrastructure provides the best combination of performance, availability and cost.

That matters at a time when the economics of AI are becoming increasingly important.

The future of generative AI may not be determined solely by who has the biggest model. It could increasingly depend on who can deliver intelligence at the lowest cost and highest reliability.

Decart Is More Than an AI Infrastructure Company

There is another reason the reported acquisition is intriguing: Decart has also been developing technology that goes well beyond traditional language models.

The company has invested heavily in real-time generative video and world models.

Its Oasis family of models is designed to generate interactive environments in real time. In June, Decart introduced Oasis 3, a world model capable of generating photorealistic driving environments for autonomous-vehicle testing. The company was also targeting robotics and other physical-AI applications.

Decart’s Lucy models have similarly focused on real-time video generation and transformation.

That technology could open doors to applications in gaming, robotics, simulation, e-commerce, entertainment and autonomous systems. Decart says its systems are designed to generate interactive environments that respond to actions in real time.

For Anthropic, this could potentially broaden the technological foundation around Claude.

Claude is primarily associated with language, reasoning, coding and enterprise AI. Decart brings expertise in a different but increasingly important category: AI systems that understand and generate dynamic environments.

That combination could become increasingly valuable as the industry moves from chatbots toward AI agents that interact with the physical and digital world.

The Deal Would Also Be a Massive Premium

The reported $6 billion price tag is striking because Decart was only recently valued at nearly $4 billion.

In May, Decart raised $300 million in a funding round that valued the company at nearly $4 billion. Nvidia participated in that round alongside other investors, while Decart continued expanding its infrastructure and world-model businesses.

A roughly $6 billion acquisition would therefore represent a substantial premium over that recent valuation.

But Anthropic can afford to think differently about the price.

After raising $65 billion in May, the company has access to an extraordinary amount of capital. Its latest valuation approached $1 trillion, reflecting investor expectations that Claude and Anthropic’s enterprise AI business could become foundational to the technology industry.

If Decart can materially reduce Anthropic’s computing costs or accelerate its ability to deploy new AI capabilities, the acquisition could pay for itself over time.

What It Could Mean for Claude

For ordinary Claude users, the most important question is simple: Would this make Claude better?

Potentially, yes—but probably not in the way a new model release does.

The most immediate benefit could be improved efficiency and scalability.

If Decart’s optimization technology helps Anthropic extract more performance from its computing infrastructure, Claude could potentially handle more workloads without requiring a proportional increase in hardware.

That could support faster responses, greater capacity and potentially lower operating costs.

It could also help Anthropic experiment more aggressively with increasingly compute-intensive AI systems.

And Decart’s expertise in real-time world models could eventually contribute to capabilities that extend beyond text.

Imagine Claude being able to reason about continuously changing visual environments, interact with simulated physical worlds or serve as an intelligence layer for robots and autonomous systems. Those capabilities are still a significant distance from today’s mainstream Claude experience, but Decart’s technology points toward that broader direction.

The acquisition would therefore be less about adding one feature to Claude and more about expanding the technological foundation on which future versions of Claude could be built.

Anthropic Is Preparing for an Even Bigger AI Battle

The reported Decart talks also reveal something important about Anthropic’s strategy.

The company is no longer operating like a small AI research laboratory.

Anthropic has raised enormous amounts of capital, expanded its enterprise presence and invested heavily in computing infrastructure. Its February Series G round raised $30 billion at a $380 billion valuation, before the company nearly tripled that valuation in its May funding round.

At the same time, competitors are pursuing similarly aggressive strategies.

OpenAI, Google and other AI companies are investing billions in chips, data centers, models and acquisitions. The result is an industry in which technological advantages can disappear quickly.

Buying Decart would allow Anthropic to acquire capabilities rather than spending years developing them internally.

It would also give the company access to a team that has already demonstrated expertise in solving some of the hardest problems surrounding efficient AI deployment.

Nvidia’s Role Makes the Story Even More Interesting

There is an additional twist.

Nvidia is an investor in Decart.

That creates an unusual dynamic if Anthropic ultimately buys the startup. Decart’s optimization technology is designed to make it easier for AI developers to use different types of processors, including chips that compete with Nvidia’s products. Nvidia nevertheless participated in Decart’s $300 million funding round.

The situation highlights how complicated the AI ecosystem has become.

Chipmakers, cloud providers, model developers and AI startups are increasingly interconnected. A company can benefit from Nvidia hardware while simultaneously developing software that makes alternative hardware more viable.

For Anthropic, having technology that can efficiently operate across multiple hardware platforms could be strategically valuable as the AI chip market becomes more competitive.

Why the $6 Billion Figure Matters

The headline number itself says something about where the AI industry is heading.

A few years ago, a $6 billion acquisition of an AI startup would have been extraordinary. In today’s market, it is becoming part of a broader pattern in which frontier AI companies are willing to spend billions to acquire specialized technology and talent.

The important question is no longer simply whether a startup has a great model.

It is whether that startup can solve a bottleneck that could limit the growth of a much larger AI platform.

Decart appears to fit that description.

Its infrastructure technology targets the cost and complexity of AI computing, while its world models address the emerging frontier of interactive and physical AI. Both areas could become increasingly important as AI systems move from generating answers to continuously operating in the real world.

A Potential Turning Point for Claude’s Next Chapter

It is important to emphasize that Anthropic has not announced an acquisition. The companies are reportedly in talks, and negotiations can change or fall apart before a transaction is completed. Reuters confirmed the discussions but noted that the reported $6 billion figure came from Bloomberg’s reporting.

Still, the possibility tells us where Anthropic sees the next major bottlenecks.

The AI race is becoming less about simply making models larger and more about building complete systems that can reason, operate efficiently, interact with the world and serve enormous numbers of users.

If Anthropic does acquire Decart, the headline may eventually be remembered as more than a $6 billion technology deal.

It could mark the moment when Anthropic decided that the future of Claude would depend not only on how intelligent the model is, but how efficiently that intelligence can operate at global scale.