A16z’s $1.1 billion Machine Age Fund goes all in on hardware
Andreessen Horowitz has raised $1.1 billion to back the chips, networking, power systems and robotics underpinning AI, making a capital-intensive wager that physical capacity is the industry's next bottleneck.
Andreessen Horowitz raises $1.1B AI infrastructure fund for chips, robots and more
Andreessen Horowitz Raises $1.1 Billion for AI Infrastructure Fund
A16z's new $1.1B fund admits hardware is eating the world, too
Andreessen Horowitz creates $1.1 billion Machine Age fund
Almost 90 new unicorns have been minted so far this year — here they are
Venture capital meets the bottleneck
Andreessen Horowitz has raised $1.1 billion for a new fund focused on the physical systems that make artificial intelligence run. Announced Friday, the vehicle will invest in chips, memory, networking and storage, along with data centers, robotics and AI-enabled home appliances.a16z +1 It is fresh capital rather than a carve-out from the $15 billion the firm announced earlier this year.economictimes
The strategy reflects a widening investment thesis: better models are not enough if power, cooling and data movement cannot keep pace. Andreessen Horowitz says hardware startups now account for more than 20% of the deals it sees, up from a small share only a few years ago.a16z +1
Bigger racks demand bigger checks
The fund’s pitch rests on rapidly rising physical requirements. Compute density has increased 28-fold from Nvidia’s H100 systems to Rubin racks, while power per rack has climbed from roughly 5–10 kilowatts to 100–250 kilowatts; the firm expects it to reach one megawatt within three years.a16z +1 That expansion creates openings beyond semiconductors, including high-bandwidth memory, interconnects, cooling, electrical equipment, materials and real estate.a16z +1
General partners Martin Casado and Raghu Raghuram will lead the strategy, which is expected to back both early- and growth-stage businesses.pitchbook Raghuram, a former VMware chief executive, said hardware companies can require larger initial investments because designs must become working chip or networking prototypes before they reach customers.economictimes
The portfolio behind the thesis is already broad. Recent investments include networking company Nexthop AI and robotics businesses Mind Robotics and Skydio, while earlier bets span Anduril, SpaceX and Waymo.pitchbook +1 That range suggests the vehicle is designed to capture both data-center infrastructure and machines that use AI in the physical world.
A costly hedge against software risk
The move also revives an old venture-capital challenge. Hardware businesses generally need more upfront funding, take longer to build and often carry thinner margins than software companies.pitchbook A dedicated fund can absorb those economics, but it also concentrates exposure to a buildout whose returns depend on sustained AI demand and successful deployment of power-hungry facilities.
For Andreessen Horowitz, the trade-off is strategic. New model features can quickly undercut application startups, encouraging investors to seek assets lower in the technology stack that are harder to replace.pitchbook If computing capacity remains scarce, infrastructure suppliers may gain leverage across the AI market. If demand cools, the same long construction cycles and heavy capital requirements could leave investors waiting longer for exits. The fund is a sizable wager that the industry’s next constraint will be measured in megawatts, memory bandwidth and manufacturing capacity—not merely model quality.