Anthropic is in talks to acquire Israeli AI startup Descartes for $6 billion, primarily attracted by its chip optimization technology that enhances the efficiency and cost-effectiveness of running AI models. This acquisition, potentially Anthropic’s largest, aligns with industry trends of securing specialized AI technologies to improve infrastructure and manage rising operational costs amid growing AI adoption.
Anthropic is reportedly in talks to acquire Descartes, an Israeli AI startup, for $6 billion. Descartes specializes in two main areas: developing world models that understand the physical world through video data, going beyond text-based AI, and creating software platforms that optimize chip performance for running AI models. The acquisition is believed to be driven primarily by interest in Descartes’ chip optimization technology, which could help improve the efficiency and economics of running AI workloads.
This potential deal would mark Anthropic’s largest acquisition to date and comes at an interesting time as the company prepares for an IPO, possibly in late September or early October. Despite the sizable $6 billion price tag, some analysts view it as reasonable given Anthropic’s substantial financial resources, having raised over $100 billion recently. The acquisition fits into a broader trend where leading AI labs and hyperscalers use mergers and acquisitions to secure specialized technologies and talent in niche areas of AI development.
There were rumors that SpaceX might also be interested in acquiring Descartes for a similar amount, but Elon Musk has denied these reports, calling them fake news. The appeal of Descartes to companies like Anthropic and potentially SpaceX lies in its software that optimizes chip performance, which is crucial given the current supply constraints and high demand for compute resources needed to run AI models efficiently. This optimization can significantly reduce operational costs and improve the scalability of AI deployments.
The discussion also touched on the economics of running AI models, highlighting that the cost is typically measured in dollars per million input and output tokens. Managing these costs is a growing concern for companies using AI, as consumption-based pricing models can lead to unexpectedly high expenses. CFOs and managers are increasingly scrutinizing AI usage to control budgets, especially as some organizations have struggled with runaway spending on AI compute resources.
Overall, the acquisition talks underscore the strategic importance of optimizing AI infrastructure and managing operational costs in the rapidly evolving AI landscape. Anthropic’s move to acquire Descartes reflects a broader industry focus on enhancing the efficiency of AI hardware and software stacks, which is critical for sustaining growth and competitiveness as AI adoption expands across various sectors.