Anthropic is reportedly considering a $6 billion acquisition of Nvidia-backed Decart AI. This move could further strengthen the Claude developer’s infrastructure capabilities as competition across the artificial intelligence (AI) market continues to ramp up.

The potential deal would see one of the world’s leading AI model developers coming together with a rapidly growing startup that specialises in infrastructure designed to make AI training and deduction more efficient.

Bloomberg news was the first to report on these talks and now Reuters backs this up with a source intimately familiar with the matter. Neither company has formally announced an agreement, which means there’s still a chance that negotiations could change or even end without a deal.

em360tech image

The massive $6 billion discussions highlight the growing importance of the infrastructure underpinning increasingly powerful and computationally demanding AI models like Claude.

What is Decart AI?

Founded in 2023, Decart is rapidly emerging as a key player in the increasingly competitive AI infrastructure market.

The company develops technology aimed at improving the efficiency of AI workloads, including DOS, its optimised inference and training stack. This Decart technology enables AI agents and reasoning models to operate faster and more efficiently.

Its portfolio also extends beyond infrastructure. The company has developed Lucy, a real-time world model designed for immersive experiences, and Oasis, a world model targeting physical AI applications such as robotics, autonomous vehicles, manufacturing and drones.

During its relatively short existence, Decart has attracted significant investor interest. In May 2026, the company announced a $300 million funding round led by Radical Ventures, bringing its total funding to more than $450 million. Nvidia was among the strategic investors participating in this round.

This followed a $100 million funding round announced in August 2025 that ended up with the startup being valued at approximately $3.1 billion.

Anthropic Looks Beyond AI Models

For Anthropic, acquiring Decart could provide access to technology capable of further improving the efficiency of the infrastructure required to train and operate increasingly sophisticated AI systems.

The company behind the Claude family of AI models has grown rapidly alongside rivals including OpenAI and Google, as enterprises increasingly adopt generative AI across software development, customer service, research and other business functions.

But developing more capable AI models requires enormous amounts of computing power, making it a competitive advantage to have access to chips, data centres and efficient AI infrastructure.

The rumoured Decart talks point to an additional dimension of the AI race. Leading AI companies are no longer competing solely over who’s got the better model performance, but also over the infrastructure required to develop and deploy those models at scale.

Bringing Decart's optimisation technology in-house could potentially allow Anthropic to improve how efficiently it uses computing resources as demand for Claude continues to increase.

AI Infrastructure Becomes a Strategic Battleground

The potential acquisition arrives as spending on AI infrastructure continues to accelerate across the technology industry.

Data centres, networking equipment, advanced semiconductors and energy capacity have all become critical components of the generative AI boom, as companies race to secure enough computing resources to train and operate increasingly complex models.

This demand is creating new opportunities for the companies that supply AI infrastructure. Nvidia, the semiconductor giant, remains at the centre of the market with its AI chips, while server manufacturers, cloud providers and data centre operators have similarly expanded their AI capabilities.

The trend is particularly relevant for companies such as Anthropic because reducing the computational cost of AI workloads can translate into significant financial and operational advantages.

Decart's focus on low-latency AI and optimised inference could complement Anthropic's existing model development efforts while potentially reducing its dependence on continually expanding raw computing capacity.

Are you enjoying the content so far?

Anthropic Prepares for Its Next Phase of Growth

The acquisition talks also come amid Anthropic’s rapid, continued growth. In May of 2026, the company raised $65 billion at a reported $965 billion post-money valuation. This was potentially the company’s final major private capital raise.

Reuters reports that the potential Decart acquisition could help Anthropic strengthen its technology and meet growing demand as the company considers a future public listing.

If completed, the deal would also show just how much leading AI companies are willing to invest in technologies that could give them a competitive advantage, setting the benchmark for future deals.

The Bigger Picture

Anthropic's reported interest in Decart shows how the AI race has evolved into an infrastructure race.

While early competition in generative AI focused more on who could build the most powerful AI model, companies are now paying greater attention to the behind the scenes technology needed to run these models efficiently and at scale.

As enterprise AI adoption grows, developers also need to figure out how to manage the high costs that come with computing power, data centres and energy expenditure. This makes technology that can improve AI efficiency increasingly valuable.

A potential $6 billion deal also shows how quickly AI infrastructure companies can grow in value. Decart has gone from raising $21 million in seed funding to potentially becoming a multibillion-dollar company in less than two years.

Whether or not this deal goes ahead, it still reflects a wider industry shift. AI companies are increasingly competing not only on their models but also on the infrastructure needed to power them.