The News
EnCharge AI, a startup developing a unique type of energy-efficient AI chip, has raised $100 million in Series B funding led by investment firm Tiger Global.
EnCharge aims to tackle the ever-growing demand for energy required to keep artificial intelligence data centers running — which McKinsey projects will amount to nearly 12% of the US’ power demand by 2030, up from just 3.7% in 2023.
The requirements are expected to continue increasing as the industry moves from the “pre-training” phase, when raw AI models are created, to inference, a process that allows existing models to become better at reasoning and draw conclusions. Current compute power is insufficient to maximize AI models’ potential, and meeting power needs is expensive and not environmentally sustainable.
EnCharge’s chips use a still relatively niche technology called analog in-memory computing, which allows artificial intelligence data to be processed directly onto a device rather than on the cloud and through data centers — therefore requiring “up to 20 times less energy compared to leading AI chips available today,” the company said, while also reducing latency and cost by keeping the data in one place.
EnCharge will use the Series B funding announced Thursday to bring to market its AI accelerator chip, at first for client computing, so that it can be deployed closer to the actual consumers of AI, Ram Rangarajan, EnCharge’s SVP of product and strategy, told Semafor.
Know More
EnCharge spun out of the Princeton research work led by professor Naveen Verma — who is co-founder and CEO of the company.
The US Defense Advanced Research Projects Agency and the Department of Defense backed the company’s work early on, while Samsung Ventures, Foxconn, and In-Q-Tel — an investment firm linked to US national security — participated in the latest funding round. The company is also working with TSMC, the Taiwanese leading semiconductor manufacturer, to fabricate its first chips.
Investors hope EnCharge’s chips will more easily adapt to space and power constraints for certain industries, like aerospace and defense, said Dan Ateya, president and managing director of RTX Ventures, another Series B investor.
But the company expects its use cases to be much broader, Rangarajan said: “A lot of companies in the AI space talk about how they are a data center solution, or an edge solution, sort of speaking to market specificity. We view AI deployment as a continuum, which means it is going to be there in the data center; it’s going to be there at the edge; it’s going to be there in the client. And we believe that the AI acceleration solution that we have to offer spans the gamut.”