Capital & Industrial Strategy

Anthropic is considering going public as soon as October 2026 according to Bloomberg, racing rival OpenAI to public markets even as it fights the Trump administration in federal court. A judge granted Anthropic a preliminary injunction blocking the Pentagon's supply chain risk designation, which would have barred government agencies from using Claude. The judge stated the measures appeared designed to punish Anthropic rather than protect national security, per The Wall Street Journal. The designation threatened billions in potential lost revenue according to Anthropic's court filings. T-Rex has already filed for leveraged Anthropic ETFs ahead of the anticipated IPO, per Reuters, indicating capital markets expect a near-term offering despite the regulatory turbulence.

OpenAI's US advertising pilot surpassed $100 million in annualised recurring revenue in under two months, per CNBC and Reuters. This marks a critical strategic shift for a company previously dependent on enterprise API revenue and consumer subscriptions. The Financial Times reported OpenAI has made a 'Code Red' turn in strategy, ditching its Sora video app and plans for an erotic chatbot to focus on revenue-generating products. The advertising success demonstrates that search-like monetisation models can work at scale for conversational AI, diversifying OpenAI's revenue base ahead of a potential IPO and reducing dependency on enterprise deals that face longer sales cycles.

Meta increased its investment in an El Paso, Texas data center to over $10 billion from a prior $1.5 billion estimate, per Bloomberg and CNBC, targeting 1 gigawatt of capacity by 2028. Separately, Carlyle and KKR won a $2 billion US Army contract to build data centers, with Army Secretary stating the Iran war underscores the need to adapt to AI's role in modern warfare, per Financial Times. The Meta increase represents a more than sixfold expansion at a single facility, signalling hyperscalers are doubling down on AI infrastructure despite macro headwinds. The private equity entry via the Army contract shows institutional capital is now flowing into defence-focused AI infrastructure, a market previously dominated by government contracting firms. Senator Mark Warner proposed taxing data centers to fund job transition programs as AI-driven job loss fears grow, per TechCrunch, indicating political backlash is building even as capital flows accelerate.

Shield AI reached a $12.7 billion valuation in its latest funding round, per Reuters, positioning the defence technology startup among the most valuable private AI companies. This valuation comes as the Carlyle-KKR Army data center contract demonstrates growing Department of Defence commitment to AI infrastructure. The convergence of high private valuations and large government contracts indicates that defence-focused AI is developing as a parallel track to commercial generative AI, with different customer concentration, regulatory dynamics, and strategic logic. Unlike consumer AI companies competing on marginal cost per query, defence AI providers sell mission-critical capabilities to a customer willing to pay premium prices for reliability and security. The valuation suggests investors see limited correlation between commercial AI margin compression and defence AI profitability.

South Korea committed $166 million in public funding to domestic AI chip startup Rebellions, per Reuters, while the US House advanced legislation requiring the Commerce Department to prevent AI chip smuggling to China following the Super Micro indictment, per Bloomberg. The South Korean investment represents direct state capitalization of a semiconductor startup competing with Nvidia and domestic players, signalling that governments now view AI chip capability as strategic infrastructure worth direct subsidy. The US legislation moves enforcement from export controls to active supply chain monitoring, indicating recognition that existing controls have failed to prevent diversion. Google's announcement of a breakthrough memory algorithm that could reduce storage needs for AI development triggered a selloff in memory chip stocks, per Bloomberg, showing how algorithmic advances can disrupt hardware investment theses overnight.