Ambiq Reveals 12nm SPOT® Platform Enabling Operation Down to 300mV for Atomiq™
Rhea-AI Impact
(High)
Rhea-AI Sentiment
(Positive)
Tags
Key Terms
SoCtechnical
Standard of care (often abbreviated SOC) is the treatment or management approach that is widely accepted and used by medical professionals for a particular disease or condition. For investors, SOC provides the benchmark against which new therapies, devices, or clinical results are judged—like comparing a new car to the current most popular model; a product that meaningfully outperforms the SOC can win market share and drive revenue, while failure to beat or match it limits commercial potential.
FinFETtechnical
A FinFET is a modern type of transistor used inside microchips where the conducting channel is formed as a thin vertical “fin” rather than a flat surface; that three-dimensional shape gives manufacturers tighter control over electrical flow. For investors, FinFET technology matters because it lets chips become smaller, faster and more energy-efficient, which affects product performance, manufacturing costs, yields and competitive advantage across the semiconductor supply chain—similar to replacing a single gate on a road with a fenced multi-lane control that reduces leakage and improves traffic flow.
edge AItechnical
Edge AI refers to artificial intelligence systems that process data directly on local devices or nearby servers rather than sending information to distant data centers. This allows for faster decision-making and real-time responses, similar to how a home security camera can instantly detect motion without needing to connect to a remote server. For investors, edge AI represents a growing trend toward more efficient, responsive technology that can create new opportunities across various industries.
multiply-accumulate (MAC)technical
A multiply-accumulate (MAC) is a basic computing step that multiplies two numbers and adds the product to an ongoing total; repeating this operation many times is the core of tasks like machine learning, audio/video processing and scientific computing. For investors, MAC performance—how many of these operations a chip can do per second and at what energy cost—serves like an engine’s horsepower for digital workloads, influencing a processor’s competitiveness, power needs and potential market demand.
A single-inductor multiple-output (SIMO) buck converter is a power regulator that uses one energy-storing component to generate several lower, independently controlled voltages from a higher supply, like one water pump feeding multiple taps with valves that time and balance the flow. For investors, SIMO designs can reduce size, parts and manufacturing costs and extend battery life in products, but they also introduce design complexity and potential efficiency or reliability trade-offs that can affect competitiveness and margins.
subthresholdtechnical
Subthreshold describes symptoms, test results, or a condition that fall short of the official cutoff for a formal diagnosis or regulatory approval—think of it like a student who passes most lessons but doesn’t meet the score needed to graduate. For investors, subthreshold findings matter because they can affect a drug’s perceived market size, trial success, labeling, reimbursement and regulatory path; small differences between subthreshold and full-threshold outcomes can materially change commercial prospects.
AUSTIN, Texas--(BUSINESS WIRE)--
Ambiq Micro, Inc. ("Ambiq®"), a leader in ultra-low-power semiconductor solutions for edge AI, today revealed new technical details of its upcoming Atomiq™ SoC. Built on Ambiq's newly designed 12nm SPOT® platform using the TSMC N12e® process, Atomiq introduces a new Ultra-Low-Power (ULP) mode engineered to operate as low as 300mV— the lowest operating voltage in Ambiq's history.
This announcement marks the first planned deployment of SPOT on a FinFET node and reflects continued progress toward initial Atomiq110 samples, with production scheduled for 2027.
FinFET + SPOT: Unlocking a New Power Frontier
By combining Ambiq’s proprietary subthreshold design methodology with TSMC’s advanced 12nm FinFET process, Atomiq enables high-speed sub- and near-threshold performance that was previously impractical in planar CMOS processes.
The new ULP mode targets operation at voltages as low as 300mV while maintaining effective performance, expanding dynamic range across Ultra-Low-Power (ULP), Low-Power (LP), and High-Performance (HP) modes. This advancement positions Ambiq to deliver its lowest operating voltage and highest energy efficiency to date.
Purpose-Built for Edge AI
Atomiq is optimized for AI workloads, which scale efficiently through parallel processing rather than raw clock speed. The previously announced Arm® Ethos™-U85 NPU executes hundreds of multiply-accumulate (MAC) operations simultaneously, delivering tens of billions of operations per second even in ULP mode with clock speeds as low as 100 MHz.
Running inference in ULP mode near 300mV is expected to significantly reduce energy consumption for edge AI applications while using standard development tools and ecosystems. This ultra-low-power feature supports the SoC's subsystems, including the processor, NPU, and multimedia engines.
Achieving operation down to 300 mV required significant architectural innovation. Atomiq incorporates a redesigned on-chip power management unit with a highly configurable, multi-channel Single-Inductor Multiple-Output (SIMO) buck converter. This adaptive power system dynamically adjusts to variations in process, temperature, and workload, delivering six independent voltage rails across load ranges spanning from nanoamps to milliamps.
Achieving such aggressive voltage scaling also required development of new custom foundation IP engineered for reliable ultra-low-voltage operation. The 12nm SPOT platform has been validated on silicon in Ambiq’s labs, marking a significant technical milestone toward commercializing Atomiq.
Executive Quotes
"Atomiq represents years of SPOT refinement and architectural innovation," said Scott Hanson, Founder and CTO of Ambiq. "Bringing SPOT to a 12nm FinFET node allows us to push voltage lower than ever before while delivering the performance required for next-generation edge AI."
“As IoT devices evolve to meet the demands of the AI era, they require cutting-edge semiconductor technologies to achieve faster performance, optimized power efficiency, and minimized power leakage,” said Dr. Lucas Tsai, Vice President of Business Management at TSMC North America. “We continue to collaborate with industry innovators like Ambiq to drive advancements in ultra-low-power semiconductor solutions through our industry-leading ULP IoT platform.”
Looking Ahead
The Atomiq110 SoC is scheduled for production in 2027. Ambiq also plans to announce additional innovations in the coming months to address one of the remaining power challenges in AI systems: memory efficiency.
Headquartered in Austin, Texas, Ambiq's mission is to enable intelligence (artificial intelligence (AI) and beyond) everywhere by delivering the lowest power semiconductor solutions. Ambiq enables its customers to deliver AI compute at the edge, where power consumption challenges are the most severe. Ambiq's technology innovations, built on the patented and proprietary subthreshold power optimized technology (SPOT®), fundamentally deliver a multi-fold improvement in power consumption over traditional semiconductor designs. Ambiq has powered over 290 million devices to date. For more information, visit www.ambiq.com.