Andreessen Horowitz has raised $1.1 billion for a new investment vehicle called Machine Age, aimed at the physical infrastructure underneath AI rather than the models themselves. The firm says it wants to “open the throttle” on hardware: chips, memory, data centers and robots. That framing matters. A venture firm known for backing software companies at high margins is now committing serious capital to industrial, capital-intensive businesses.
Most of the AI funding story so far has centered on model labs racing on capability and reasoning benchmarks. Machine Age instead targets what a16z calls the infrastructure that powers AI, a category the firm defines broadly: chips, memory, interconnects, edge devices, cooling systems, electrical work and real estate. That scope sits closer to industrial investing than the application-software bets that made a16z’s name.
In its announcement, a16z names specific technical bottlenecks the fund intends to address: systems that run quicker on less power, memory that costs less and moves more data at every tier of the hierarchy, and links between nodes and systems that are both faster and easier to scale. It also flags power efficient edge devices as a priority, extending the fund’s remit past data centers into where AI runs in the physical world.
The framing implies a thesis about where AI’s scarcity has moved. When model quality was the constraint, returns went to firms that could out-train or out-tune a foundation model on comparatively modest capital. As frontier labs converge on similar capabilities and the gains from further training narrow, the binding constraint shifts toward whether enough power, chips and physical space exist to run those models at scale.
A fund built around concrete, cooling systems and electrical buildout is a different venture bet than the software-margin thesis that built firms like a16z. It implies longer capital cycles, heavier asset intensity, and returns tied to construction timelines rather than product release cycles. That is a meaningfully different risk profile from funding a small team shipping code, closer to the calculus infrastructure and industrial investors use than the one venture firms typically apply to early-stage software bets.
a16z frames the stakes in sweeping terms, calling AI the “strongest tool ever developed” for solving problems and describing its advancement as a “social and national imperative.” The firm has not named specific portfolio companies or disclosed a target size beyond the $1.1 billion already raised.
For operators tracking where AI capacity gets built next, Machine Age is a signal that at least one major venture firm now treats power and hardware access, not model access, as the resource worth funding ahead of demand.
TechCrunch’s Sean O’Kane reported the Machine Age fund launch on August 28, 2026.
