Billion Models Military
A $965 Billion Bet: AI's Infrastructure Financing Reaches Utility Scale
Anthropic closed a $65 billion funding round, valuing the company at $965 billion and placing it among the most valuable private companies in history. The capital requirements that follow are commensurate: Apollo and Blackstone are reportedly shopping a $36 billion debt deal structured specifically to fund purchases of Google's custom TPUs for Anthropic's model training operations — a financing arrangement that resembles the buildout of power plants or telecommunications networks more than a conventional venture capital deal.
European Union officials traveled to San Francisco for discussions with Anthropic described as 'Mythos talks,' signaling that European regulators are departing from their traditional pattern of attempting post-hoc regulation of American technology platforms. Instead, they are engaging directly with AI companies during the development phase, before market dominance is established.
IBM announced a $10 billion quantum computing investment alongside a separate $5 billion open-source security initiative, positioning the company to address both the near-term security requirements of enterprise AI deployment and the longer-term potential of quantum acceleration. Microsoft is expected to announce new in-house AI models at its Build conference, a move that would reduce its dependence on OpenAI and give it greater control over AI integration across its Office, Azure, and Windows platforms.
Cisco released research finding that no frontier AI model is currently safe from multi-turn attacks — a class of threat in which seemingly innocuous initial queries are followed by increasingly sophisticated prompts that gradually manipulate AI systems into producing harmful or unauthorized outputs. Unlike conventional cybersecurity threats, multi-turn attacks use normal conversational patterns, making them exceptionally difficult to defend against with traditional methods.
Vice President JD Vance warned Air Force graduates against ceding military decisions to artificial intelligence, reflecting growing concern about autonomous weapons systems in high-stakes contexts. The warning marks a shift from political rhetoric that has focused primarily on AI's economic implications, toward the question of machine judgment in situations where errors could result in loss of life or international incidents.