At last month’s “Winning the AI Race” summit in DC hosted by the All-In Podcast, one of the key themes was clear: AI isn’t a force for job destruction, it’s a force for job creation
At last month’s “Winning the AI Race” summit in DC hosted by the All-In Podcast, one of the key themes was clear: AI isn’t a force for job destruction, it’s a force for job creation.
We often frame the AI revolution in terms of models, chips, and software. But behind the scenes, it requires a massive buildout of physical infrastructure.
According to McKinsey & Company, companies across the compute power value chain will need to invest $5.2 trillion in data centers by 2030 to meet global AI demand. That number reflects the competitive race underway among hyperscalers and enterprises to build proprietary AI capacity. The result: a boom in data center construction.
But there’s a big problem: we have a shortage of skilled tradespeople.
While everyone’s focused on the AI talent wars or racing to train LLMs, we forgot to train the electricians, welders, HVAC techs, and semiconductor technicians who will build and maintain the physical backbone of the AI economy.
The shortage is already stalling projects.
The U.S. government's new AI Action Plan, released last month, directly addresses this issue. It calls for investment in roles like electricians and advanced HVAC technicians, and calls on agencies (including DOL, DOE, ED, NSF, and DOC) to partner with employers and local governments to create industry-driven training programs that ensure job readiness and direct placement into high-demand roles.
Y Combinator echoed this need in their Fall 2025 Request for Startups list shared last week, calling for a new kind of vocational school tailored to the AI economy.
YC’s thesis aligns closely with ours: founders can use AI to build personalized, scalable training tools that get people job-ready in months, not years.
Multimodal AI could power:
➡️ Voice-based AI coaches guiding trainees through hands-on tasks
➡️ AR/VR environments for practicing in simulation, with AI tutors using vision models to observe and provide real-time feedback
This is a space we’ve been thinking deeply about. Earlier this year, our team published a blog (by Bridget Duru and Ahzam Ahmed - linked below) on building skilled trades talent pipelines and our thesis on companies creating AI-powered co-pilots to support workers on the job. We also led an investment in montamo (Berlin-based), which trains people in just eight weeks for high-demand trades careers.
We're seeing more examples of what this could look like in the US. At the "Winning the AI Race" summit, Chris Power (Founder of Hadrian) shared that Hadrian, which just raised $260M led by Founders Fund and Lux Capital to build automated factories, now trains machinists in 30 days using AI assistants.
If you’re building in this space, we’d love to hear from you!
References:
https://www.whitehouse.gov/wp-content/uploads/2025/07/Americas-AI-Action-Plan.pdf