Be a part of our every day and weekly newsletters for the most recent updates and unique content material on industry-leading AI protection. Be taught Extra
As enterprises proceed to discover other ways to optimize how they deal with totally different workloads within the information heart and on the edge, a brand new startup, Ubitium, has emerged from stealth with an attention-grabbing, cost-saving computing strategy: common processing.
Led by semiconductor {industry} veterans, the startup has developed a microprocessor structure that consolidates all processing duties – be it for AI inferencing or general-purpose duties – right into a single versatile chip.
This, Ubitium says, has the potential to remodel how enterprises strategy computing, saving them the effort of counting on various kinds of processors and processor cores for various specialised workloads. It additionally introduced $3.7 million in funding from a number of enterprise capital corporations.
Ubitium mentioned it’s presently centered on creating common chips that might optimize computing for edge or embedded gadgets, serving to enterprises reduce down deployment prices by an element of as much as 100x. Nonetheless, it emphasised that the structure is extremely scalable and will also be used for information facilities sooner or later.
It’s going up towards some established names within the edge AI compute area reminiscent of Nvidia with its Jetson line of chips and Sima.AI with its Modalix household, displaying how the race to create AI-specific processors is transferring down funnel from giant datacenters to extra discrete gadgets and workloads.
Why an all-in-one chip?
In the present day, relating to powering an edge or embedded system, organizations depend on system-on-chips (SoCs) integrating a number of specialised processing models — CPUs for basic duties, GPUs for graphics and parallel processing, NPUs for accelerated AI workloads, DSPs for sign processing and FPGAs for customizable {hardware} features. These built-in models work in conjunction to make sure that the gadget delivers the anticipated efficiency. instance is the case of smartphones which regularly use NPUs with different processors for environment friendly on-device AI processing whereas sustaining low energy consumption.
Whereas the strategy does the job, it comes on the expense of elevated {hardware} and software program complexity and better manufacturing prices — making adoption tough for enterprises. On prime of it, when there’s a patchwork of parts on the stack, underutilization of sources can grow to be a significant concern. Primarily, when the gadget just isn’t operating an AI perform, the NPU for AI workloads would simply be idling, taking over the silicon space (and vitality).
To repair this hole, Martin Vorbach, who holds over 200 semiconductor patents licensed by main American chip corporations, got here up with the common processing structure. He spent 15 years creating the expertise and ultimately teamed up with CEO Hyun Shin Cho and former Intel exec Peter Weber to commercialize it.
On the core, Shin Cho defined, the microprocessor structure permits the identical transistors of the chip to be reused for various processing duties, thereby enabling a single processor to dynamically adapt to totally different workloads, proper from basic computing required for easy management logic to huge parallel information move processing and AI inferencing.
“As we reuse the same transistors for various workloads, replacing an array of chips and reducing complexity, we lower the overall cost of the system. Depending on the baseline, this is a performance/cost ratio of 10x to 100x…The reuse of transistors for different workloads drastically reduces the overall transistor count in the processor — further saving energy and silicon area,” the CEO added.
Purpose to make superior computing accessible
With the homogeneous, workload-agnostic microprocessing structure, Ubitium hopes it will likely be in a position to substitute typical processors – CPUs, NPUs, GPUs, DSPs, and FPGAs – with a single, versatile chip. The consolidation (resulting in simplified system design and decrease prices) will make superior computing extra accessible, enabling quicker improvement cycles for functions throughout shopper electronics, industrial automation, dwelling automation, healthcare, automotive, area and protection.
The structure can also be totally compliant with RISC-V, the open-source instruction set structure for processor improvement. This makes it simple to make the most of for functions like IoT, human-machine interfaces and robotics.
“By lowering the barrier for high-performance compute deployment and AI capabilities, our technology allows IoT devices to process data locally and make intelligent decisions in real-time. This will also help solve interoperability issues by enabling devices to adapt and communicate seamlessly with diverse systems,” Cho defined.
At this stage, the corporate has 18 patents on the expertise with an FPGA emulation-based prototype and is transferring to develop a portfolio of chips various in array dimension however sharing the identical underlying common structure and software program stack. It plans to launch a multi-project wafer prototype with a improvement equipment within the coming months and ship the primary edge computing chips to prospects in 2026.
In the end, Cho mentioned, the work will permit them to supply scalable computing options for various (and evolving) efficiency wants, from embedded gadgets to large-scale edge computing programs.
“Our workload-agnostic processor will also be able to adapt to new AI developments without hardware modifications. This will enable developers to implement the latest AI models on existing devices, reducing costs and complexity associated with hardware changes.… By separating the hardware and software layers, we aim to establish our processor as a standard computing platform that simplifies development and accelerates innovation across diverse industries,” he added.
