Key facts
- Alphabet plans to raise up to $80 billion in equity for AI infrastructure.
- Berkshire Hathaway purchased $10 billion of Alphabet stock.
- ARK Invest bought over 267,000 shares of Alphabet stock.
- Meta Platforms is considering selling tens of billions of dollars in stock for AI investments.
- Combined AI-related capital expenditures are expected to surpass $700 billion this year.
- US data center construction is not keeping pace with AI demand.
- SoftBank plans to build 3.1 GW of AI data centers in France by 2031.
- Thailand's Sarath Ratanavadi plans to invest up to $4.3 billion over five years in AI infrastructure.
- Xnrgy Climate Systems, a data center parts maker, is exploring a sale valued up to $10 billion.
- Schneider Electric SE is planning to sell €800 million ($930 million) of debt.
- Nvidia's Vera Rubin platform for AI systems has entered full production.
- Mercor's AI token costs now exceed employee salaries.
The artificial intelligence race is driving unprecedented capital investment, with Alphabet planning to raise up to $80 billion in equity to fund its AI infrastructure and compute needs. This includes a $10 billion private deal with Berkshire Hathaway, which purchased Alphabet stock following a dip after the equity offering announcement. ARK Invest also acquired over 267,000 shares. Meta Platforms is reportedly considering selling tens of billions of dollars in stock to finance its own substantial AI infrastructure investments, a move that caused its stock to drop. This trend aligns with a broader industry movement, with combined AI-related capital expenditures expected to surpass $700 billion this year.
The massive investment in AI is creating significant demand for supporting infrastructure, particularly data centers. US data center construction is reportedly lagging behind this demand. SoftBank plans to build 3.1 GW of AI data centers in France by 2031. Thailand's richest individual, Sarath Ratanavadi, plans to invest up to $4.3 billion over five years to expand data centers and AI infrastructure. Companies like Xnrgy Climate Systems, a maker of heating and cooling components for AI data centers, are exploring potential sales valued up to $10 billion due to strong demand. Schneider Electric SE is planning to sell €800 million ($930 million) of debt, benefiting from its involvement in supporting data centers. Dow is showcasing its AI data center cooling technologies, and Daqo New Energy's stock rose 9% on a deal to build an AI data center energy solutions base.
However, the AI infrastructure boom faces several bottlenecks. Copper wiring limitations regarding heat, distance, and power consumption are hindering large-scale GPU connections, with photonics emerging as a potential solution for faster data transfer. Nvidia's CEO, Jensen Huang, stated the company can meet AI chip demand despite supply constraints, and its Vera Rubin platform has entered full production to accelerate AI systems. Mercor's CEO noted that AI token costs now exceed employee salaries, predicting this trend will become common within five years. SambaNova's CEO discussed a new $3.5 billion AI cloud initiative using disaggregated architecture to lower AI inference costs. Layoff data is also beginning to reflect AI's impact on the labor market, with some Amazon engineers criticizing the company's estimated $200 billion AI data center spending amid significant layoffs and supporting a pause on a new Seattle data center.
Discussions around AI's economic impact and tax policy are ongoing, with proposals from Sam Altman, Mark Cuban, and Elizabeth Warren being debated. The author of one piece argues against drastic tax code changes based on speculative future impacts, advocating for established tax principles. OpenAI CEO Sam Altman is also backing Alfred, a stealth robotics startup.