Ambiq Announces New Apollo5 SoC for Efficient AI

By Ken Briodagh

Editor in Chief

Embedded Computing Design

March 27, 2024

Story

Ambiq Announces New Apollo5 SoC for Efficient AI

In a new release, semiconductor company Ambiq has revealed its Apollo510, the first member of the company’s Apollo5 SoC family, designed to drive efficient and low-latency AI processing.

The company says that the Apollo510 MCU leverages the Arm Cortex-M55 CPU with Arm Helium to reach processing speeds up to 250MHz with a 10x reduction in latency and a 2x reduction in energy consumption as compared to Ambiq’s Apollo4 series. With this MCU combining performance and efficiency, Ambiq is setting its customers up with the tools to deploy AI-driven applications for speech, vision, health, and industrial models on battery-powered devices. Reportedly, this is the most efficient semiconductor on the market to operate with the Arm Cortex-M55.

Scott Hanson, Ambiq’s CTO and founder told us that the company is solving the problem of AI at the edge with the Apollo5 series. “How do you shrink a cloud’s worth of material into tiny, low power devices?” The way, he says, is through sub-threshold MCU tech like what’s in the Apollo510. This allows the MCU to run at voltages as low as 0.5 V, which keeps batteries working longer.

“We at Ambiq have pushed our proprietary SPOT platform to optimize power consumption in support of our customers, who are aggressively increasing the intelligence and sophistication of their battery-powered devices year after year,” said Hanson. “Our memory is on-chip to keep power lower and save latency, which translates to efficiency.”

The Apollo510 reportedly can run most endpoint AI calculations, including low-power sensor monitoring, always-on voice commands, and telco-quality audio enhancement. Thanks to its design, Ambiq says that manufacturers of IoT devices for AI/ML inferencing will be able to expand their power budgets and add capabilities to their devices through the SPOT-optimized design. These include next-gen wearables, digital health devices, AR/VR glasses, factory automation, and remote monitoring devices, Ambiq said.

The built-in Arm Helium tech supports up to 8 MACs per cycle as well as half, full, and double precision floating point operations, designed for AI calculations in addition to general signal processing operations. Apollo510 also includes 4 MB of on-chip NVM and 3.75 MB of on-chip SRAM and TCM to give developers have smooth development and application flexibility.

“As applications across health, industrial and smart home continue to advance, the need for secure edge AI is crucial for next generation devices,” said Paul Williamson, SVP and GM, IoT Line of Business at Arm. “Ambiq’s new family of SoCs, built on Arm, will deliver significant performance gains for on-device AI, helping developers and device manufacturers deliver the capabilities required for the AI era.”

Security at the Edge is, naturally, critical. To that end, and using Ambiq’s secureSPOT platform, Apollo510 integrates Arm TrustZone technology with a physical unclonable function (PUF), tamper-resistant OTP, and secure peripherals.

The Apollo510 MCU is already sampling with customers, and the company says it will be generally available in Q4. Take a look in person at Embedded World  in Nuremburg, Germany April 9-11.

Ken Briodagh is a writer and editor with two decades of experience under his belt. He is in love with technology and if he had his druthers, he would beta test everything from shoe phones to flying cars. In previous lives, he’s been a short order cook, telemarketer, medical supply technician, mover of the bodies at a funeral home, pirate, poet, partial alliterist, parent, partner and pretender to various thrones. Most of his exploits are either exaggerated or blatantly false.

More from Ken