Efficient Computer raises $97M Series B to scale Electron E1 from physical AI to the datacenter

Efficient Computer said Tuesday it entered agreements for more than $97 million in Series B financing at a $650 million valuation, bringing total funding to $173 million, to ship its Electron E1 processor in volume and scale the Fabric architecture toward datacenter-class performance.

AI’s power bill is the bottleneck from a robot’s battery to a datacenter rack. Efficient is betting a general-purpose chip, designed from scratch, can cut energy enough that physical AI and datacenter work stop dying on the power budget.

On Tuesday, 29 September 2026, Efficient Computer said it entered agreements for more than $97 million in Series B financing at a $650 million valuation. The company said that brings the total it has raised to $173 million. Series B is a growth round, new cash after earlier lettered rounds. A valuation is the price this announcement puts on the whole company. More than $97 million means the agreements are above that figure. The wire is datelined Pittsburgh. The page stamps the release Sep 29, 2026, 11:00 ET, which is 11:00 a.m. Eastern. Those figures are Efficient’s, on PR Newswire.

TQ Ventures leads the round. Efficient names the other participants as Eclipse, Union Square Ventures, Giant Ventures, Triatomic Capital, TO Capital, TF Capital, Mana Ventures, Toyota Ventures, Overmatch, and Borderless. That list is the company’s.

Efficient said the Electron E1 is now in volume production, and that the new capital will ship that processor in volume to lead customers. Lead customers, here, means the buyers who get the chip as production ramps. The company does not print their names. The same money is meant to scale the Fabric architecture to datacenter-class performance. Datacenter-class means performance aimed at the large buildings that run AI, not only at a small device. Volume production means the chip is being made in quantity, not only as a lab sample. Those lines are Efficient’s.

The company claims more than a 10 times improvement in energy use versus systems built today, on the datacenter line. More than 10 times means the same work would use less than a tenth of the energy. The same release also says the Fabric brings a 10-to-100-times improvement in energy use for general-purpose computation, including AI. Those multiples are Efficient’s claims. The release does not name the machines in the comparison, and it does not print a price for the chip.

Electron E1 is Efficient’s general-purpose processor on the Fabric design. General-purpose means it is meant to run a whole application, not only one slice of an AI model. The company says customers are adopting it for physical AI and autonomy, for watching critical infrastructure, for space and defense, and for wearable devices. Physical AI means software that has to sense and act in the world, such as a robot or a drone, rather than only answer in a chat window. Efficient says it has scaled production to meet customer interest and will keep scaling volume into 2027. Those lines are Efficient’s.

Efficient says the Fabric stays general-purpose on purpose. It says chips built for only one piece of an AI model do not run most software, and can go stale as models change. The company says its chip supports ordinary software, including C and C++. Those lines are Efficient’s.

The company describes the limit as energy. It says demand for computing, from physical AI to the datacenter, is outrunning the ability to generate, store, and deliver power. It says a robot’s need to think, plan, and act safely can cut how long it runs to minutes instead of hours. It says each future datacenter needs its own power plan, which it calls a strain the grid cannot take. Those lines are Efficient’s account of the bottleneck.

Brandon Lucia, chief executive and co-founder, said: “Every customer we meet has a version of their product they cannot build, because the compute power budget makes the new capabilities they want infeasible.” A power budget is the electricity a product is allowed to use. He said this financing lets Efficient scale the Fabric from the devices shipping today to datacenter scale, and that the company will not stop until energy is no longer a limit on AI and computing. Those sentences are his, in the release.

Andrew Marks, co-founding partner at TQ Ventures, said the firm was convinced by the team’s ability to build both the hardware and the software. He said they have taped out four times and are already shipping chips to customers at volume. A tape-out is the point when a chip design is sent to be manufactured. That quotation is his, on the same release.

In plain terms, a Pittsburgh chip company said it lined up more than $97 million, at a $650 million valuation, led by TQ Ventures. The money is to ship the Electron E1 in volume and to push the same Fabric design toward datacenter performance. The energy savings in the release are the company’s claims.

The picture is Efficient Computer’s Electron E1 development board. The E1 chip and the EFFICIENT name are on the board. It is the company’s product photograph. It is not a shot of a datacenter.