Highlights:

  • NeuReality Ltd., the startup, has announced that it has received USD 35 million funding to market its NR1 chip, which is meant to accelerate artificial intelligence applications.
  • NeuReality sells the NR1 as part of the NR1-S Inference Server appliance. The appliance has several NR1 chips.

NeuReality Ltd., the startup recently announced that it had received USD 35 million to market its NR1 chip, which is meant to accelerate artificial intelligence applications.

Samsung Ventures, Cardumen Capital, Varana Capital, OurCrowd, and XT Hitech led the Series A fundraising round. Several other institutional backers also participated! Among them was SK Hynix Inc., one of the world’s top memory chip manufacturers.

XT High-Tech Managing Director Yoav Sebba said, “High performance and sustainable Inference computing is critical for growth in day-to-day usage of AI. We are seeing endless AI opportunities developed by software companies, but the existing hardware infrastructure is limiting the deployment of those use cases.”

NeuReality, the Israel-based company, has invented the NR1 system-on-chip to perform AI inference tasks. Running trained neural networks in production is referred to as ‘Inference’. According to NeuReality, the NR1 can power a wide range of AI workloads, including natural language processing models and product recommendation engines.

NeuReality sells the NR1 as part of the NR1-S Inference Server appliance. The appliance has several NR1 chips. NeuReality claims that the NR1-S Inference Server can reduce expenses and energy consumption by a factor of 50 compared to competitor hardware systems.

Additionally, the firm permits consumers to utilize its chip in different ways. It sells the NR1 as part of an accelerator card known as the NR1-M, which may be connected to a server using a PCIe port. The accelerator card lets businesses connect NeuReality’s technology with their current data center servers.

In large-scale AI infrastructure systems, hardware efficiency frequently poses a hurdle. The more the number of chips added to an AI system, the greater the processing power required to manage the chips. This issue reduces the amount of available processing power for machine learning applications.

NeuReality claims that its NR1 chip overcomes the issue by enabling linear scalability. The phrase linear scalability describes a scenario in which adding more chips to a server cluster does not dramatically reduce hardware efficiency. Increasing the effectiveness of AI infrastructure allows businesses to cut hardware expenses as well as power consumption.

In addition to the NR1, NeuReality offers a suite of software tools to facilitate the deployment of AI applications in production. Additionally, the startup’s products promise to simplify application administration. NeuReality’s software portfolio contains, among other components, a so-called AI hypervisor that facilitates the management of machine learning applications implemented on NR1 chips.

Since last May, NeuReality has begun distributing NR1 prototype implementations to partners. With its recently disclosed USD35 million fundraising round, the business intends to roll out its technologies more broadly. NeuReality will hire 20 additional personnel to assist the initiative over the next six months.