Anthropic is in talks to acquire AI startup Decart for $6 billion to boost its AI infrastructure efficiency ahead of a potential IPO, marking a strategic shift in its typically disciplined M&A approach amid strong market demand for AI-related technology. Meanwhile, industry players like Cisco, CoreWeave, and Cerberus face challenges and opportunities in AI hardware and cloud computing, with evolving chip dependencies and market dynamics influencing growth and investment outlooks.
Anthropic is reportedly in talks to acquire AI startup Decart for about $6 billion, marking its largest acquisition to date. Decart specializes in optimizing computing infrastructure efficiency across various chip types, including NVIDIA, TPU, and Amazon chips, which is crucial as AI companies face increasing demand for compute resources. This acquisition aligns with Anthropic’s strategy to enhance cost efficiency and performance ahead of a potential IPO expected as soon as October. Unlike some competitors, Anthropic has maintained a disciplined M&A approach, making this deal a significant shift in its growth strategy.
In the public markets, technology shares, particularly in the chip sector, continue to push record highs, driven by strong demand for AI-related infrastructure. Investors are increasingly looking beyond mega-cap tech giants to mid-cap companies that play critical roles in the AI ecosystem, such as those involved in power, water, and data center infrastructure. These mid-cap firms offer attractive valuations and diversification opportunities, with some leveraging AI to improve their business outcomes, exemplified by companies like Robinhood using AI to enhance customer service.
Cisco recently reported fourth-quarter earnings that disappointed investors despite projecting $7.5 billion in AI-related sales, a figure seen as conservative compared to previous years. The main concern lies in gross margin expectations and the timing of backlog conversion into revenue, issues echoed across other networking and semiconductor players due to supply constraints. Nonetheless, Cisco remains well-positioned in the AI space, particularly in optics and networking silicon, which are experiencing significant growth.
CoreWeave, a cloud computing company heavily reliant on NVIDIA chips, warned investors that diversifying away from exclusive NVIDIA architecture will require substantial time and investment. This signals a potential shift in the AI chip landscape as customers seek alternatives to NVIDIA’s dominance. Despite this, CoreWeave reported strong earnings and margin performance, alleviating some bearish sentiment around neo-cloud providers.
Cerberus, an AI hardware company, saw its shares slide after reporting slower-than-expected growth, though its cloud rental business is expanding rapidly. CEO Andrew Feldman highlighted the challenges of deploying AI compute capacity, including data center availability and manufacturing scale-up, but remains optimistic about tripling revenue next year. The company is also collaborating with AMD to enhance its offerings, aiming to provide customers with a combination of speed and throughput advantages in AI processing. Meanwhile, broader market dynamics include SpaceX’s recent AI model launch and Meta facing a significant legal trial over alleged mental health impacts linked to its platforms.