Ambarella CEO Sees Edge AI Growth Across Security, Autos and Wearables

Ambarella sees expanding edge AI demand across security, automotive, consumer video, wearables and infrastructure, driven by customers needing more processing power for transformer and large language models on devices.

The company’s key differentiator is power efficiency, combining AI inference, video processing and CPU functions in its SoCs.

Its portfolio spans roughly 1 to 1,000 TOPS and is supported by a unified software development kit.

Automotive represents about 30% of revenue and security remains the largest market, while wearables and robotics offer faster but earlier-stage growth.

Ambarella is targeting 10%–15% fiscal 2027 revenue growth, though memory availability and pricing remain risks.

Chief Executive Officer Fermi Wang said the fabless semiconductor company is positioning its system-on-chip portfolio for broader adoption of edge artificial intelligence, with opportunities spanning enterprise security, consumer video, wearables, automotive and edge infrastructure.

Speaking at Citi's TMT Conference, Wang said Ambarella, founded in 2004, initially focused on digital-video applications before spending the past decade developing edge AI processing products.

Its SoCs combine AI inference engines with image-processing pipelines, CPUs and peripheral functions, allowing the company to address complete edge-device systems rather than only a single processing function.

Wang said customer engagement has reinforced Ambarella's view that demand for AI performance at the edge is increasing.

Customers have progressed from convolutional neural networks to transformer-based models and large language models, he said, and are seeking more processing capability in devices.

However, he said the edge AI market does not yet have a single dominant application comparable with the major workloads that drive data-center spending.

Wang identified power efficiency as Ambarella's primary competitive advantage, describing performance per watt across AI inference, video processing and other functions as central to the company's design-win efforts.

He said customers may initially use GPUs because they are broadly available and easy to program, but power consumption becomes a constraint for battery-powered products and systems with limited heat dissipation.

He also pointed to Ambarella's digital-video processing heritage and its Cooper software development kit.

The unified SDK supports the company's 15 AI SoCs, Wang said, enabling customers to move applications between different silicon products with limited software modification.

Among competitors, Wang said NVIDIA offers high performance but GPUs can be less suitable for customers requiring energy-efficient solutions.

He characterized Qualcomm as Ambarella's largest competitor because of greater overlap in their offerings, while describing Synaptics' edge AI coverage as concentrated more in the roughly 1 TOPS to 10 TOPS range.

Ambarella offers products ranging from 1 TOP to 1,000 TOPS, he said.

Enterprise security remains Ambarella's largest addressed application, according to Wang.

He said the role of security cameras is expanding beyond surveillance as retailers and other organizations connect cameras to edge systems that can use large language models to analyze store traffic, customer behavior, purchasing activity and operational conditions.

In consumer video, Wang cited demand f...