Four industrial models under pressure at once

German carmakers are announcing historic job cuts as Chinese rivals flood European markets, according to the Financial Times. The threat to the industrial model of Europe's largest economy is mounting, with legacy automakers unable to match the cost structure or production velocity of incoming competitors. The same week, Shenzhen expanded its driverless vehicle deployment, leaving gig-economy drivers on the sidelines. Two different automation waves, same displacement story.

Elsewhere in the tech stack, Google capped Meta's access to its Gemini models as surging demand for advanced AI strained available capacity. The FT reports that computing power has become the scarcest commodity in the industry. What was abundant six quarters ago is now rationed between hyperscale partners. The infrastructure that was supposed to democratize model access is instead creating new choke points and new dependencies.

Meanwhile, the U.S. defense sector is developing modular workshops designed to mass-produce cheap missiles during wartime. The FT frames it as America's search for a McDonald's model for missile making—standardized, distributed, scalable under stress. The contrast with legacy defense procurement could not be sharper.

Four stories, four industrial models under revision. The German auto model that defined postwar European exports is shedding labor. The Chinese mobility model is absorbing taxi work into software. The AI cloud model that promised infinite scale is hitting resource walls. And the defense-industrial model is trying to borrow the language of fast casual. Each one is a named Current in motion: the reshoring and remilitarization of supply chains, the margin compression from AI capital intensity, the automation of legacy labor markets.

The through-line is not technology for its own sake. It is the repricing of what used to be stable: employment in exporters, compute access in hyperscale, human drivers in mobility, and Pentagon procurement timelines. The volatility is structural, not cyclical, and it is moving faster than the frameworks built to manage it.