SAN JOSE, California —Arm CEO Simon Segars said that the company is adding a new feature to its processors that will allow system-on-chip (SoC) makers to design custom chips by adding their own instructions for specific embedded IoT applications.

“We are, for the very first time, enabling people to add custom instructions into Arm CPUs,” Segars said during the Arm TechCon keynote today. “We have never done that before — we are a company famed for standardization.”

And while that standardization has enabled a “powerful and mature ecosystem,” ensuring that “software written to run on one Arm processor will run on pretty much any Arm processor,” it also limits flexibility, and makes it difficult to optimize chips for emerging applications, Segars added. “What we’ve seen with the evolution of workloads people want to run on their devices: there’s a flexibility with what people want to do with the chips. With customization of the processors, you can put that right into the design of the chip.”

The Arm Custom Instructions will initially be available in Arm Cortex-M33 CPUs beginning in the in the first half of 2020, and it will be free to new and existing licensees.

The new application-specific features work like this: modifications to the CPU reserve encoding space for designers to add custom datapath extensions while maintaining the integrity of the existing software ecosystem. This, together with the existing co-processor interface, enable the CPUs to be extended with accelerators optimized for edge compute use cases including machine learning and artificial intelligence (AI).

“This is big change for us,” Segars said. “We’re thinking about how we evolve continually and how we enable this fifth wave of computing.”

The intersection of AI, 5G, and IoT enables this fifth wave of computing, according to Segars.

“What’s driving the growth of these technologies is the ubiquity of intelligence semiconductors,” he added. And, of course, that’s where Arm comes into play.

Arm’s semiconductor partners — the company licenses its IP to companies that embed that into their chips — have now shipped 150 billion devices based on Arm technology during the company’s history, Segars said. “It took 50 years to ship the first 50 billion, and we’re anticipating the next 50 billion will ship in two years alone,” he added. The sheer volume of data created by the booming number of connected devices will drive this demand, and “the next winners and losers are going to be the companies who are determined by the ability to gain insight from all of that data.”

This requires AI. And while AI traditionally happens in the cloud, it needs to move to the edge of the network and into the endpoints to enable rapid processing of all of this data.

For example: a small water sensor that can be attached to pipes to monitor water flow and check for anomalies such as leaks and blockages. “The company that built this has found some real resonance deploying it in healthcare, hospitals,” Segars said.

The way this device is currently built requires the sensor monitoring the water flow to connect via Bluetooth to a gateway that transfers this data to the cloud. “That’s fine, but it’s actually very, very inefficient,” Segars said. “Transferring all that data — that’s not scalable.”

The next generation of this device, he added will enable machine learning workloads on the device itself at the point where the data is captured.

“But even then, we can go further by directly connecting this into the network,” Segars said. “The problem is, of course you can’t do that with today’s networks and that’s why 5G is so important.”

However, as these anticipated 1 trillion smart devices connect directly onto a cellular network, security becomes even more of a concern.

To this end, Arm earlier announced what it calls Platform Security Architecture (PSA) framework, and at MCW in Barcelona earlier this year it announced PSA Certified, which provides independent security certification. PSA-certified devices are now coming to market, Segars said.

Additionally, the company is adding to its secure enclave technology, which isolates applications from each other. This means if an attacker does hack into one application, he or she can’t access data from another application, thus limiting the scope of the breach. Segars teased some announcements on this front coming soon: “Over the next couple of weeks we’ll have some exciting news to share with you.”