This Week in AI: This Week in AI: GPT 5.4, Anthropic Fires Back...

OpenAI made two major moves this week that set the AI industry abuzz. First, it launched GPT-5.4, featuring native computer-use abilities, a 1-million-token context window, and new integrations with Microsoft Excel and Google Sheets — a clear push toward autonomous enterprise workflows. The model is designed to handle tasks like navigating spreadsheets and coding without human hand-holding, signaling a shift from chat-based AI to agentic assistants. Building on this momentum, OpenAI also closed a massive $110 billion funding round at a $730 billion pre-money valuation, with backing from SoftBank, Nvidia, and Amazon. The company now serves 900 million weekly users and boasts 3 gigawatts of dedicated inference capacity, underscoring that frontier AI has become a capital-intensive infrastructure race where compute and distribution are the defining moats.

However, the week also saw fierce backlash against OpenAI's $200 million Department of Defense contract. Anthropic CEO Dario Amodei publicly accused OpenAI of spreading “straight up lies” about Anthropic’s military stance, while the Pentagon reportedly pressured labs over model restrictions. The controversy triggered a 295% spike in ChatGPT uninstalls and open letters from AI employees urging limits on military deployment. As the governance battle escalated, Nvidia made a quieter but equally strategic move, investing $2 billion each in Lumentum and Coherent to secure advanced optical components for next-gen AI data centers. This underlines that the AI race is as much about physical infrastructure — networking and photonics — as it is about algorithms.