January 11th

Anthropic is finalizing a $10 billion funding round led by Coatue and GIC, propelling its valuation to $350 billion—a nearly six-fold increase since March 2025. This capital injection provides flexible growth funds separate from its recent $15 billion compute-tied deal with Microsoft and NVIDIA, solidifying Anthropic's position as a top-tier frontier lab.

OpenAI and SoftBank invested $1 billion into SB Energy to develop 1.2 GW of data center capacity in Texas, part of the broader $500 billion Stargate infrastructure initiative. This strategic partnership bypasses energy bottlenecks by integrating power generation directly with data center development.

Zhipu AI raised $558 million in its Hong Kong IPO, valuing the Beijing-backed firm at HK$4.3 billion ($553 million) as it becomes the first of China's "AI tigers" to go public. Despite US sanctions, Zhipu's public listing signals the maturity of China's domestic LLM ecosystem.

Bosch committed €2.9 billion to AI investment by 2027, focusing on manufacturing supply chains and perception systems rather than consumer chatbots, moving AI from pilot phases to core operations through edge computing for defect detection and equipment failure prediction.

NVIDIA launched its Vera Rubin computing platform, featuring a GPU that delivers 5x the training performance of the Blackwell generation and is already in full production. The architecture introduces a new memory tier to manage "KV cache," specifically designed to resolve bottlenecks in agentic AI workflows requiring massive context retention.

Meta signed agreements with Vistra, Oklo, and TerraPower to secure over 6 gigawatts of nuclear energy capacity, with some reactors expected online by the early 2030s. This underscores the tech sector's pivot to nuclear fission and small modular reactors (SMRs) to guarantee 24/7 baseload power required by gigawatt-scale AI clusters.

AMD unveiled the Ryzen AI 400 Series and the MI455X GPU, aiming to deliver up to 10x more performance across varying workloads compared to previous generations. By targeting both the data center with its "Helios" rack-scale platform and the consumer PC market, AMD is positioning itself as the only full-stack alternative to NVIDIA.

Fermi America broke ground on "Project Matador," a 5,800-acre campus in Texas designed to host 11 gigawatts of power capacity for AI, backed by legacy nuclear tech, highlighting the extreme scale of US infrastructure ambitions as a national security asset in the AI race against China.

X (formerly Twitter) restricted Grok's image generation tools to paying subscribers after the model generated approximately 6,700 non-consensual sexualized images per hour during independent testing. The EU ordered data retention for investigation and the UK threatened potential bans under the Online Safety Act.

Democratic Senators Wyden, Lujan, and Markey formally requested that Apple and Google remove X from their app stores, citing Grok's generation of illegal content as a violation of distribution terms, escalating the liability debate from platform immunity to app store gatekeeper responsibilities.

California State Senator Steve Padilla introduced a bill to ban AI chatbots in toys for children under 18 for four years, citing risks of manipulation and inappropriate content, reflecting growing legislative skepticism regarding the safety of integrating LLMs into products for vulnerable populations.

The FDA announced it will relax regulations on AI-enabled clinical decision support software, allowing tools that provide recommendations to enter the market without pre-market review, aiming to accelerate innovation but raising concerns about deployment of unverified algorithms in critical healthcare settings.

Boston Dynamics, backed by Hyundai, deployed its fully electric Atlas humanoid robot into active factory work, autonomously sorting parts in a Savannah, Georgia plant. This marks a transition from R&D to commercial viability for humanoid robotics, with Goldman Sachs predicting a $38 billion market within the decade.

Anthropic secured a partnership with Allianz to deploy Claude Code to the insurer's workforce, focusing on automating complex coding and workflow tasks, highlighting a shift in enterprise adoption where AI handles multi-step regulatory and operational workflows rather than just content generation.

Grab acquired robotics startup Infermove to internalize development of delivery robots for unstructured urban environments, aiming to lower delivery costs and mitigate labor shortages in dense Southeast Asian markets by owning the full technology stack.

Google is overhauling the Gmail experience with an experimental "AI Inbox" that prioritizes emails and summarizes threads, while also bringing generative video (Veo) and image capabilities to Google TV, signaling Google's strategy to make its Gemini model the default interface for both productivity and entertainment ecosystems.

Ford announced a native AI assistant arriving in vehicles by 2027, capable of handling complex, vehicle-specific queries like load capacity, leveraging off-the-shelf LLMs via Google Cloud to compete with rivals like Rivian and Tesla who have already deployed similar in-car agents.

Microsoft integrated "buy buttons" directly into Copilot, allowing users to purchase products like sneakers without leaving the chat interface, moving AI chatbots from informational assistants to transactional agents and directly challenging the traditional search-to-commerce funnel.