Artificial intelligence chip startup Blaize Inc. has gotten the funding it needs to support the growing industry demand for its platform after closing on a $71 million late-stage round of funding.

Today’s Series D round was led by new investor Franklin Templeton and existing investor Temasek, with participation from DENSO and other new and existing investors, and brings the company’s total amount raised to $155 million.

Blaize has created a specialized computer chip for high-performance AI workloads. It’s called the Graph Streaming Processor, and it’s based on a unique chip architecture that sees neural networks implemented in the form of a graph data structure.

The GSP is further equipped with optimizations designed to run those graphs. They include a feature that’s designed to save on power by reducing the amount of times data is moved to and from memory while being processed.

This architecture enables Blaize’s GSPs to deliver a very high level of performance. They can run AI models up to 60 times more efficiently than traditional central processing units and graphics cards, for example. Moreover, the GSPs can manage up to 16 trillion operations per second while consuming just seven watts of power, equivalent to what a smart light bulb runs on.

Blaize is pushing for companies to deploy its hardware at the network edge, where AI applications can benefit from having lots of processing power on site. By doing so, that negates the need to send data to a far-flung data center to be processed, cutting down on latency.

To that end, the company last year announced new hardware modules to help deploy its GSPs at the edge. They included a module called Pathfinder P1600 that can be integrated with connected devices such as industrial sensors so they can process data they generate on the spot. So a manufacturer could link the P1600 powered by the Blaize chips to cameras on its factory floor in order to spot defective components during production, for example.

The company’s other modules are designed for desktop computers and servers located at remote locations such as branch offices and retail stores.

JP Scandalios, Franklin Templeton’s senior vice president and portfolio manager, said Blaize has already shown it offers value by helping customers tap into AI at the edge. “Automotive, and numerous edge AI markets, such as retail and metro, hold tremendous potential for Blaize to expand on their early market position as the adoption of AI at the edge accelerates, creating a new wave of industrial systems,” he said.

Blaize Chief Executive Dinakar Mungala said today’s round leaves the company well-positioned to sustain its lead in delivering integrated hardware and software offerings that enable this value creation.

Blaize said it will use the funding to accelerate its product roadmap to meet what it says is growing demand for high-performance, lower-power and low-cost AI hardware and software in segments such as automotive, smart retail and industry.

Image: Blaize