The AI race is no longer just about algorithms. It’s about electricity. The White House has invoked the Defense Production Act to accelerate domestic production of high-voltage circuit breakers and power electronics - critical components for an aging grid now deemed a national security risk. According to The AI Daily Brief, the administration sees the grid as a strategic bottleneck that private industry can’t fix alone.
Energy demand from data centers is on track to double to 11% of total U.S. consumption by 2030. Without intervention, analysts at JP Morgan and Goldman Sachs warn of systemic risk: power shortages could stall the compute race before it finishes. The move signals a fundamental shift - transformers and transmission lines are now treated like weapons systems in a new kind of cold war.
"The race is no longer just about who builds the smartest model, but who provides the cheapest intelligence at scale."
- Matthew Berman, The AI Daily Brief
Ben Horowitz argues on The a16z Show that AI has shattered the old rules of venture capital. Technical leads no longer matter if capital and compute can brute-force replication. The new moats aren’t code - they’re physical: supply chains, sales relationships, and access to gigawatts. Navan, which holds global airline and hotel contracts, won’t be replaced by an AI lab that won’t spend years building real-world logistics.
Jordi Visser, speaking on Bankless, calls this the 'SaaSpocalypse' - a collapse of software value as AI floods the market with zero-cost intelligence. When everything digital becomes infinitely reproducible, investors flee to atoms and math. Bitcoin, he argues, is the purest expression of unreplicable scarcity. Institutional ETFs are now clearing the sell side, marking what Visser calls a 'Bitcoin IPO' - a permanent shift in ownership and valuation.
"AI is the new QE: instead of printing money, it prints productivity. Companies grow while shrinking headcount."
- Jordi Visser, Bankless
Meanwhile, DeepSeek V4 Pro offers near-frontier performance at roughly 1/7th the cost of Anthropic’s Opus 4.6. For most enterprise tasks - coding, summarization, routine logic - the performance gap is negligible. The savings aren’t. Beijing is tightening control, blocking Meta’s acquisition of Manus and instructing firms to reject U.S. capital. The threat isn’t just competitive. It’s strategic: if U.S. companies build on Chinese models, they risk being cut off overnight.
