Ad
Skip to content
Read full article about: Anthropic places $21 billion order for Google chips via Broadcom

AI lab Anthropic has placed orders totaling $21 billion with Broadcom for Google's AI chips. Broadcom CEO Hock Tan confirmed that the startup is purchasing "Ironwood Racks" equipped with Google's Tensor Processing Units (TPUs).

The move follows a massive cloud partnership between Anthropic and Google announced in late October. That deal grants Anthropic access to up to one million TPUs and is expected to bring over one gigawatt of new AI compute capacity online by 2026. Anthropic maintains a multi-cloud strategy, spreading its workloads across Google TPUs, Amazon's Trainium chips, and Nvidia GPUs.

Comment Source: CNBC
Read full article about: Trump signs executive order threatening funding cuts over state AI rules

On Thursday, President Donald Trump signed an executive order threatening to withhold federal funding from states whose AI regulations impede American innovation. Speaking to reporters, Trump argued that a patchwork of state-level rules slows down the industry.

The order directs the Secretary of Commerce to review state laws and, in cases of conflict, block access to the $42 billion broadband fund. Previous attempts to enforce such bans had failed.

The directive specifically targets anti-discrimination rules in states like Colorado, which the administration claims lead to "ideological bias." While tech giants like Google and OpenAI welcomed the move toward national standards, Representative Don Beyer warned that the order could create a "lawless Wild West" and potentially violates the Constitution by undermining state-level safety reforms.

Read full article about: Google opens its infrastructure for AI models via MCP

Google is integrating Anthropic's Model Context Protocol (MCP) directly into its cloud infrastructure. MCP serves as a universal standard for connecting AI models with external data and tools, eliminating the need to program new interfaces for every application.

Starting immediately, Google is offering managed servers that give AI agents direct access to services like Google Maps, BigQuery, and both the Compute and Kubernetes Engines. This allows AI to handle tasks like independently managing infrastructure or planning travel routes. Through the Apigee platform, companies can also deploy their own internal APIs as AI tools. Google announced plans to expand support to additional services, such as Cloud Storage and databases, in the near future.