AI Resolve, Resilience and Results-despite AI Fragility
Big Tech’s Resolve, the foundation models’ Resilience, and the enterprise Results keep compounding — despite the very real early Fragility of these tools.
Three Key Takes today:
(1) Amazon puts $12B on GlobalStar, with Apple in tow. Second-largest Amazon acquisition ever. Spectrum + 200+ satellites bolt onto Amazon’s LEO cluster. Apple keeps its iPhone/Watch SOS network. Both companies lean into an alternative to Starlink. Three space-AI fronts now in motion: satellite-to-device, AI data centers in space, chips fabricated in space. More also at AI-RTZ #1057 today.
(2) Enterprise token-maxxing is real. Uber’s CTO sees higher AI Compute Budgets: thousands of devs past chatbots into AI agents, meeting and exceeding internal compute budgets. Anthropic and OpenAI raised pricing this week — because demand exceeds their compute supply. Salesforce is standing up AWUs to measure token-generated output. Wave 2 of enterprise AI: not just adoption, now metric measurement and tracking instrumentation.
(3) Samsung S26 pricing up 20-30%. Surface yesterday, Samsung today. Memory shortages favor the biggest buyers. The AI Mainframe → Local Computing transition stays slower than the research cadence — because the RAM to run agentic AI locally costs what it costs.
Plus: the one feature of AI I can’t live without — summarizing the unsummarizable. Read less. Learn more. Then read far wider.
Three things can be true at once: staggering capex, staggering progress, staggering fragility. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)
