Data center hardware startup Mueon Corp. today announced that it has raised $65 million in funding. 

Semiconductor-focused fund Maverick Silicon led the Series A round. It was joined by Intel Corp.’s venture capital arm, Oregon Venture Fund, Wellington Capital, IQT and several others.

Mueon was founded last year by former Intel engineers Wilfred Gomes and Jack Hwang. Chief Executive Gomes worked on die stacking technologies during his time at the chip giant. Chief Technology Officer Hwang helped develop the techniques that Intel uses to etch transistors into silicon wafers.

A traditional server’s core components sit next to one another on a motherboard in a kind of two-dimensional configuration. Mueon is developing a three-dimensional server architecture, Cubelet, that stacks computing and memory components atop one another. According to the company, the technology could prove up to be 10 times more power- and cost-efficient than current hardware.

Stacked system designs are more efficient than traditional 2D hardware because they pack server components closer together. That reduces the physical distance data must travel between those components. Shortening data travel times, in turn, lowers the associated power use. It also speeds up computations in the process. 

In a blog post published last year, Mueon investor Oregon Venture Fund described Cubelets as “Legos for the data center.” The post stated that a Cubelet module includes not only computing and storage modules but also the auxiliary hardware they need to operate reliably. That includes interconnect, power delivery and thermal management equipment.

Thermal management is a major challenge in 3D computing systems. One source of complexity is that processors and memory chips often have different maximum operating temperatures. That means a processor operating near its maximum temperature can overheat any memory modules that may be stacked on top of it.

In many servers, chips dissipate a part of the heat they generate by channeling it to the motherboard. Doing so is relatively simple when a chip is placed directly on its motherboard. However, that’s often not the case in 3D systems. The heat from the top layer of a 3D system must go through multiple other components before it reaches the motherboard.

Last year’s Oregon Venture Fund blog post stated that Mueon had found a way of addressing 3D systems’ thermal management issues. The article also detailed that the company’s Cubelet modules can host multiple types of chips. They will support, among others, graphics cards and Arm central processing units.

Mueon stated today that its Cubelets are optimized for artificial intelligence and hyperscale workloads. The company will use the proceeds from its funding round to hire more engineers and improve its system design. In parallel, Mueon is developing software that will enable customers to run popular AI frameworks on its hardware.

“This financing gives us the scale to bring the next generation of compute and memory system-scale integration to market,” Gomes said. “Mueon’s architecture is built to move past today’s rack-level constraints and deliver the kind of density the rest of the industry is only beginning to chart a path toward.”

Image: Unsplash