Why embedded chipmakers are acquiring ML software firms

Large semiconductor suppliers have started acquiring machine learning (ML) software houses in a bid to bolster their artificial intelligence (AI) offerings for embedded systems, and the latest transaction has been signed between Infineon Technologies AG and Imagimob AB, a Stockholm-based startup that provides ML solutions for edge devices. The Swedish company’s toolchain delivers production-grade ML models.

This platform can be used in a variety of sensor and Internet of Things (IoT) use cases like audio event detection, voice control, predictive maintenance, gesture recognition, and signal classification. So, at a time when AI/ML technologies are penetrating almost every embedded system, Infineon aims to leverage this AI edge platform for its sensor and IoT solutions. Moreover, this Tiny Machine Learning technology will boost Infineon’s hardware/software ecosystem for embedded systems.

Figure 1 Infineon’s acquisition is aimed at the adoption of Tiny Machine Learning in IoT applications. Source: Imagimob

Infineon’s European neighbor STMicroelectronics signed a similar deal a couple of years ago when it snapped Cartesiam, a software company developing tools for ML and inferencing on Arm-based microcontrollers. The Toulon, France-based Cartesiam was founded in 2016, and its team included data scientists and embedded signal processing experts.

Cartesiam’s NanoEdge AI Studio enabled embedded systems designers without prior knowledge of AI to rapidly develop specialized libraries and integrate machine-learning algorithms directly into a broad range of applications. At the time of acquisition, the French startup’s AI solution was already in production in connected devices, household appliances, and industrial machines.

At ST, it’d complement the Franco-Italian chipmaker’s STM32Cube.AI toolset, which allows design engineers to map and run pre-trained artificial neural networks on the company’s STM32 microcontrollers. Like Infineon’s acquisition of Imagimob, adding Cartesiam’s ML technology is expected to boost ST’s embedded AI offerings at the edge.

Figure 2 Cartesiam’s ML technology at the edge is expected to complement the STM32Cube.AI platform. Source: STMicroelectronics

These two deals underscore two important trends. First, AI/ML technologies are now a crucial part of hardware and software stacks in embedded system designs. Second, as part of their AI strategy, chipmakers are increasingly offering complementary tools alongside semiconductor devices to address the full spectrum of embedded AI/ML learning needs.

So, expect more deals like these in the future.

Related Content

Getting a Grasp on AI at the Edge
Edge Intelligence Ticks Many Boxes for AI
Generative AI in factory automation and the edge
Adding Low-Power AI/ML Inference to Edge Devices
Factors Affecting the Performance and Efficiency of Edge Artificial Intelligence

<!–

–>

<!–

div-gpt-ad-inread

–>

The post Why embedded chipmakers are acquiring ML software firms appeared first on EDN.