Andressen Horowitz created a new $1.1 billion fund called Machine Age. The organization indicated that the new fund would focus its investments on the physical infrastructure of AI in order to accelerate the associated capacity and system development.
Investing in hardware
Unlike a16z, which used to be the more common path of investment in software, this fund is specifically committed to hardware. According to the notes issued by its network of officials, the Fund will focus on a set of bottom facilities that support AI operations, not just models or application-level companies.
Coverage includes areas such as computing chips, storage, data centres and robotics. a16z is of the view that, as the size of the AI system continues to expand, industry requirements for computing, bandwidth, energy consumption and deployment efficiency are rising simultaneously.
Aim at arithmetic and infrastructure bottlenecks
a16z In the note, it is mentioned that the current industry still needs faster and more efficient systems, as well as cheaper and more bandwidth storage programmes. In addition to chips and memory, interconnectivity between nodes and systems is considered to be a key element in AI ' s expansion.
The Agency also mentioned that low-capacity AI equipment for the edges was equally important. This type of equipment is used to allow AI to explore, sense and interact in the real environment, and to apply scenarios to cover terminal equipment and robotic systems.
The data centre package is in the investment range. Internal
In addition to the core computing components, the Fund will also focus on supporting the AI infrastructure landings, including heat dispersion, materials, power systems and real estate construction. a16z ' s statement indicates that its judgement is not limited to the chip itself, but views AI expansion as a system project involving supply chain and physical construction.
This means that the focus of investment around AI is continuing to extend to the infrastructure chain of the heavier assets. As the demand for model training and reasoning increases, funds are increasingly directed to hardware and facilities that can directly support the deployment of computing power.