Industry Overview - The AI ecosystem has solidified with five major players emerging as finalists in the race for the biggest model: Microsoft/OpenAI, Amazon/Anthropic, Google, Meta, and xAI [2][4] - The supply chain has shifted into high gear with Nvidia's Blackwell chip shipping, TSMC building new fab capacity, and Broadcom working on custom AI chips [2] - Data centers planned in early 2024 are entering full build mode, with physical data centers rising across America [3] Key Players and Strategies - Google's advantage is vertical integration, owning its own TPU chips, data centers, and research team [8] - OpenAI has the strongest brand in AI, with revenue reportedly over $3.6B, and may continue to widen the gap with rivals [9] - Anthropic has gained research talent, including key hires from OpenAI, positioning itself as a favored destination for AI scientists [10][11] - xAI is the pace setter for data center scaling, with its 100k GPU Colossus cluster and plans for larger clusters [12] - Meta is all-in on open source, leveraging its Llama models and distribution advantages through Instagram, WhatsApp, and Facebook [13][14] AI Search as a Killer App - AI search is emerging as a powerful reinvention of internet search, with the potential to fragment the monolithic market [18][19] - Domain-specific AI search engines will proliferate, tailored to the needs of different professions such as doctors, lawyers, and accountants [19][22] - OpenAI launched ChatGPT Search in October, reaching 10M monthly active users, challenging Google's dominance [17] CapEx and Industry Dynamics - Big Tech companies have locked arms around the AI revolution, controlling data centers and owning equity stakes in big model companies [25] - 2025 is expected to be a stabilization year for AI CapEx, with companies focusing on completing projects and selling installed capacity [26][27] - Quarterly CapEx figures show a trendline starting to stabilize, with Microsoft and Google leading the way [27][28] - New data center capacity coming online in 2025 will continue the epic decline in AI compute prices, benefiting startups and incentivizing innovation [29] Future Outlook - The AI building blocks are firmly in place, with 2025 set to be a year of leveraging these capabilities to create incredible new applications [30] - The competitive landscape has solidified, and 2025 will reveal which strategies prove prescient and which prove ill-fated [15]