The Dutch chipmaker NXP Semiconductors N.V. has extended its spending spree by splurging $307 million on a deal to buy Kinara Inc., a California-based startup that develops neural processing units for artificial intelligence workloads at the network edge.

Today’s deal comes just over a month after NXP splashed out $625 million to acquire the autonomous vehicle software company TTTech Auto AG, which was announced just three weeks after it spent $242.5 million on the automotive connectivity systems maker Aviva Links Inc.

The first two acquisitions were clearly designed to enhance NXP’s largest automotive chip business segment, but today’s deal is more about growing its Industrial & IoT group, which encompasses computer chips for the edge.

Though NXP has largely missed the AI boat that has helped Nvidia Corp. to become one of the world’s most valuable companies, it might be a different story at the network edge. NPUs such as Kinara’s are seen as essential for AI workloads in devices such as smart cameras and drones, which operate at the edge of the network.

NXP’s business certainly needs a boost. Today’s acquisition comes just a week after it announced its fourth-quarter financial results, where it saw revenue decline across all four of its major business segments, including automotive, mobile, industrial and the internet of thing, and the division comprising communications, infrastructure and other products. To make matters worse, NXP also forecast first-quarter revenue of between $2.725 billion and $2.925 billion, the midpoint some way short of Wall Street’s $2.92 billion target

The poor results extended a string of less-than-impressive quarters that call into question NXP’s efforts to diversify its chip business, and stories have been circulating that the chipmaker could be about to lay off as many as 1,800 employees.

It remains to be seen what kind of impact Kinara will have on NXP’s bottom line, but the deal does at least give the company access to some viable edge AI products. Kinara’s flagship processors include the Ara-1 NPU, which is designed to handle advanced AI inference at the edge, and the Ara-2 NPU, which is a speedier version of that chip that maxes out at 40 Trillions Operations Per Second.

Kinara also happens to be an existing partner of NXP’s, so that should ensure the speedy integration of its chips with NXP’s platforms once the acquisition closes, likely in the first half of this year. Indeed, the two companies said they will show off their first integrated products next month at Embedded World 2025 in Nuremberg, Germany, next month.

Holger Mueller of Constellation Research Inc. said NXP’s strategy of diversifying into many different markets worked well over the last decade or so, and has only really failed in the last 12 months or so.

“That raises questions about whether or not it’s the wrong strategy, or simply just a case of poor implementation of that strategy,” the analyst said. “This acquisition, together with the others in the last couple of months, suggests NXP believes the diversification strategy is still the right path forward, but it needs to double down on that. If anything, it believes that further diversification is needed, and Kinara and TTTech Auto AG could both help it target a larger total addressable market.”

According to NXP, the acquisition of Kinara will help it to provide “complete and scalable AI platforms, from TinyML to generative AI” by combining that company’s discrete NPUs with its own portfolio of processors, connectivity and security software.

Rafael Sotomayor, NXP’s executive vice president and general manager of secure connected edge, said generative AI can deliver major improvements in efficiency, sustainability, safety and predictability in industrial markets.

“Adding Kinara’s AI capabilities to our broad intelligent edge portfolio creates a scalable platform for new classes of AI-powered systems,” he said.

Image: NXP