Andreessen Horowitz co-founded in 2009 by Marc Andreessen and Ben Horowitz has raised $1.1 billion for a new artificial intelligence infrastructure fund, expanding its investment focus beyond software and into the physical systems increasingly required to support AI development.
Called the Machine Age Fund, the new pool of capital will invest across computing infrastructure, including chips, memory, networking and storage, while also looking at areas such as data centers, robotics and home appliances.
The fund is notable not only for its size, but also for how Andreessen Horowitz has chosen to structure it. Rather than allocating money from the $15 billion venture fund the firm raised earlier this year, the investor created a separate fund specifically for this category.
Andreessen Horowitz general partner Raghu Raghuram said the decision reflects the different economics involved in backing companies that build physical technology. These businesses can require larger amounts of capital at an earlier stage and often need a different approach to underwriting than software startups.
Early-stage companies will be the main focus
Most of the Machine Age Fund is expected to go into early-stage startups, although part of the capital will be available for more mature companies.
That investment strategy reflects one of the central challenges facing hardware-focused businesses. A semiconductor or networking startup does not stop at designing a product. It must eventually turn that design into a working prototype, a process that can require substantial funding before a commercial product reaches the market.
Raghuram said chip companies, for example, may need more capital at inception than businesses developing AI models because of the cost involved in building prototypes of chips and networking systems.
That requirement helps explain why Andreessen Horowitz has created a dedicated vehicle instead of treating AI infrastructure as another segment within a broader venture portfolio.
Computing capacity has become a constraint
Behind the new fund is Andreessen Horowitz’s view that rapid growth in AI workloads is putting pressure on existing computing infrastructure and supply chains.
The firm believes demand for AI computing is increasing fast enough to create bottlenecks across hardware systems. Raghuram said companies in San Francisco frequently identify limited access to AI computing capacity as one of the factors restricting their growth.
The infrastructure challenge stretches across several layers of the technology stack. More powerful AI systems require not only processors, but also the memory, networking equipment, storage and data-center capacity needed to operate those processors at scale.
That gives the Machine Age Fund a broader investment mandate than simply backing semiconductor companies.
Andreessen Horowitz sees opportunity despite higher capital needs
Hardware businesses can take significant capital to reach the prototype and production stages, but Andreessen Horowitz also sees evidence that successful infrastructure companies can still find viable exit opportunities.
Raghuram pointed to Groq’s reported $20 billion licensing deal with Nvidia and the public debut of Cerebras Systems when discussing the potential paths available to successful technology companies despite their heavier funding requirements.
The wider demand picture is also supporting the firm’s investment thesis. Nvidia CEO Jensen Huang said during an earnings call that demand for the company’s chips was growing by more than 70% in fiscal 2028 while remaining constrained by supply.
A software investor putting more weight behind hardware
Andreessen Horowitz is widely associated with software investing, but the firm says hardware has long been part of its strategy.
Its recent investments include Unconventional AI and Mind Robotics, while it has previously invested in SpaceX. The Machine Age Fund now gives that part of the portfolio a dedicated source of capital at a time when AI development is placing greater demands on physical computing systems.
The $1.1 billion raise therefore represents more than another AI-focused fund. It is a bet that one of the next major investment opportunities in artificial intelligence will sit beneath the models themselves, in the chips, networks, memory, storage and physical infrastructure required to keep them running.
