Episode
228: The Eight Trillion Dollar Elephant: Valuing #Google Without Search
- Podcast
- Deep Dive with Gemini
- Published
- Apr 24, 2026
- Duration seconds
- 3030
- Processing state
not_requested
Actions
POST https://stenobird.com/v1/public/podcasts/deep-dive-with-gemini-7518385/episodes/228-the-eight-trillion-dollar-elephant-valuing-google-without-search/transcription-requests
Idempotently request low-priority transcript generation for this episode.GET https://stenobird.com/podcast/deep-dive-with-gemini-7518385/228-the-eight-trillion-dollar-elephant-valuing-google-without-search.md
Read the agent-friendly Markdown representation of this episode resource.
Summary
Research #TPU vs #GPU the core architectural battle between Google's new custom AI chips, the 8th-generation Tensor Processing Units ( #TPU8t and #TPU8i), and NVIDIA's #VeraRubin GPUs. The competition centers on whether #AI developers prefer Google's specialized, workload-specific chips or #NVIDIA's highly flexible, general-purpose accelerators for training and deploying AI models. #AlphabetValuation This refers to the financial analysis of Google's stock and market capitalization. According to a Sum-of-the-Parts (SOTP) valuation model, Google's individual business segments (such as its AI Hardware, YouTube, and Waymo) could have an implied value of $8.17 trillion, even if its massive Search business was valued at zero. Analysts are closely watching how Google's new hardware will impact its future earnings and stock price. #AgenticEra captures the current shift in the AI industry toward autonomous AI agents capable of multi-step reasoning, planning, and tool execution. The hardware landscape is evolving to meet these needs; for example, Google's inference-specialized TPU 8i is specifically engineered with massive on-chip SRAM to support the memory and latency demands required for these real-time AI agents. #LiquidCooling This represents the extreme physical and thermal infrastructure demands of modern AI data centers. Because next-generation AI accelerators draw immense amounts of power—with the NVIDIA Rubin R100 requiring up to 2.3 kilowatts per GPU—data centers are now forced to transition to advanced, direct-to-chip liquid cooling systems. #AICloud This highlights the booming cloud infrastructure market driven by artificial intelligence. Google is heavily investing in its data center capacity, with projected 2026 capital expenditures reaching between $175 billion an…