Good morning.

Elon Musk’s xAI has raised $20 billion in a new funding round, surpassing its earlier $15 billion target.

The investor group includes strategic partners NVIDIA and Cisco, alongside long-time backers Valor Equity Partners, StepStone Group, Fidelity, the Qatar Investment Authority, Abu Dhabi’s MGX, and Baron Capital Group.

NVIDIA and Cisco are not just investors but significant vendors and strategic partners to xAI, underscoring how capital, hardware, and AI infrastructure are tightly intertwined.

Let’s dive into today’s edition.

Vistra Corp. is acquiring the Cogentrix Energy portfolio in a deal valued at approximately $4 billion, significantly expanding its U.S. power-generation footprint. The transaction adds about 5.5 gigawatts of natural gas–fired capacity, strengthening Vistra’s presence in some of the most constrained and strategically important power markets in North America.

Power Play: The assets are concentrated across PJM, ISO New England, and ERCOT, regions facing rising electricity demand, grid congestion, and reliability pressures. The portfolio includes ten relatively modern gas-fired plants, most of which are combined-cycle facilities with strong efficiency metrics, providing Vistra with more flexible, dispatchable power as load growth accelerates.

Big Picture: The deal comes as electricity demand rises from data centers, electrification, and population growth, especially in Texas and the U.S. Northeast. Utilities increasingly see natural gas as a critical, dispatchable backstop to intermittent renewable power.

U.S. political opposition to the AI data center expansion is intensifying, with Bernie Sanders and Ron DeSantis—from opposite ends of the political spectrum—raising concerns about rising electricity bills, grid strain, and community impacts. Sanders has called for a national moratorium on new data centers, while DeSantis unveiled an AI bill of rights that would allow local governments to block projects.

The pushback comes as power systems, especially PJM Interconnection, face acute capacity shortages driven largely by data center demand.

New Battleground: With residential electricity prices expected to continue rising through 2026 and mid-term elections approaching, data centers are shifting from economic engines to political liabilities.

The U.S. government has committed $2.7 billion to expand domestic uranium enrichment capacity amid accelerating electricity demand. The funding, awarded by the U.S. Department of Energy, targets a long-standing weak point in the U.S. energy supply chain amid rising power needs from AI data centers, electrification, and grid strain.

Nuclear Family: The contracts were awarded to Centrus Energy, Orano Federal Services, and General Matter, with each expected to receive roughly $900 million over the next decade. The funding covers production of both low-enriched uranium (LEU) for existing reactors and HALEU fuel needed for advanced and small modular nuclear reactors.

Broader Outlook: The move reflects a wider policy shift to position nuclear power as a stable, carbon-free baseload for AI data centers and other power-intensive technologies, as renewables scale unevenly and grid bottlenecks deepen. It also aims to cut U.S. reliance on foreign uranium enrichment, particularly Russia, while securing a reliable electricity supply amid surging AI-driven demand.

The Federal Reserve System says U.S. data center construction could take sharply different paths over the next two years, with annual investment ranging from $360 billion to $930 billion by late 2027.

Base Scenario: In a base-case scenario, the Fed expects spending to stabilise at around $600 billion annually, up from roughly $100 billion in early 2024, driven by demand for AI and cloud infrastructure.

Fast Buildout: Even the most conservative outlook would still mark an unprecedented wave of capital deployment, cementing data centers as one of the largest drivers of U.S. commercial construction

Big Warning: However, the Fed flagged mounting constraints that could slow momentum. Grid capacity limits, shortages of industrial land, labour bottlenecks, equipment delays, and power availability are already showing signs of strain. The central bank warned that if these supply-side pressures intensify, today’s aggressive expansion plans may prove difficult to sustain

Samsung Electronics is expected to report a 160% jump in fourth-quarter operating profit, fueled by surging memory chip prices as global demand for AI hardware tightens supply. Analysts forecast Q4 operating profit of about 16.9 trillion won ($11.7 billion), compared with 6.5 trillion won a year earlier, marking Samsung’s strongest quarter since 2018.

Samsung’s turnaround follows a period where it lagged rivals such as SK Hynix in supplying advanced chips to Nvidia. Analysts say Samsung is now gaining traction with its next-generation HBM4 memory, positioning it to benefit further as AI infrastructure spending accelerates—though rising chip prices are beginning to squeeze margins in smartphones and other consumer devices.

China’s new export-licensing regime for silver, effective January 1, is emerging as a fresh risk for the AI supply chain. Under the rules, 60–70% of refined silver exports now require government approval, tightening access to a metal critical to AI hardware, data centre power systems, and advanced electronics.

Silver Bullet: Silver is essential for high-performance chips, electrical contacts, servers, and power infrastructure supporting AI data centers, as well as for solar panels that increasingly power AI workloads. With China controlling a dominant share of global silver supply, the move raises the risk of higher input costs and supply bottlenecks for AI infrastructure buildouts.

Critical Choke Points: The restrictions reinforce a broader trend: as AI demand accelerates, critical materials are becoming geopolitical levers. For AI developers and data center operators, silver is no longer just a commodity—it is now a strategic chokepoint that could affect costs, timelines, and long-term capacity planning.

Data centres connected to the largest U.S. power grid added $6.5 billion to electricity procurement costs in the latest auction, intensifying concerns that the AI boom is driving energy inflation. According to PJM Interconnection, total costs associated with data centres have reached $23.1 billion for the June 2025–May 2028 period.

Power Hungry: An independent report by Monitoring Analytics shows that data centers accounted for 49% of the $47.2 billion in costs across three consecutive PJM auctions since mid-2024. The rapid growth of AI-driven electricity demand is increasingly dominating capacity markets, pushing prices higher for all consumers.

Future Outlook: The figures add to mounting scrutiny over how AI infrastructure is reshaping power economics. With grid upgrades lagging and data center demand accelerating,