Despite the current pandemic-related downturn, the demand for edge AI chips will grow to outstrip demand for cloud AI chips for the first time in 2025, according to a new report from ABI Research. By 2025, the edge AI chip market will reach $12.2bn in revenue, whereas cloud AI chip revenues will reach $11.9bn in the same time frame. While most AI training and inference workloads are handled in the cloud today, ABI Research predicts that growth in the edge AI chipset market will be driven by increasing demands for low latency and data privacy plus the availability of low-cost ultra-low-power chips designed specifically for this application. AI training and inference will be processed in gateways and all types of edge devices, right down to sensor nodes in the next five years. Save Your Seat for: the Upcoming Webinar on "Mind of Engineer" Survey results "By integrating an AI chipset designed to perform high-speed inference and quantized federated learning or collaborative learning models, edge AI brings task automation and augmentation to device and sensor levels across various sectors," said Lian Jye Su, principal analyst at ABI Research.

The firm expects the edge AI chipset market to reach $12.2 billion in revenues, outpacing the cloud AI chipset market, which will reach $11.9 billion in 2025. The transition, says the firm, will be propelled by the increasing focus on low latency, data privacy, and the availability of low-cost and ultra-power-efficient AI chipsets. Currently, the cloud is the center of AI, with most AI training and inference workloads happening in the public and private clouds. Traditionally, the centralization of these workloads in the cloud brings the benefits of flexibility and scalability. However, says the firm, the industry has witnessed a shift in the AI paradigm driven by the need for privacy, cybersecurity, and low latency.

Data-driven experiences are rich, immersive and immediate. Think pizza delivery by drone, video cameras that can record traffic accidents at an intersection, freight trucks that can identify a potential system failure. These kinds of fast-acting activities need lots of data -- quickly. So they can't sustain latency as data travels to and from the cloud. That to-and-fro takes too long.

Data-driven experiences are rich, immersive and immediate. Think pizza delivery by drone, video cameras that can record traffic accidents at an intersection, freight trucks that can identify a potential system failure. These kinds of fast-acting activities need lots of data -- quickly. So they can't sustain latency as data travels to and from the cloud. That to-and-fro takes too long; instead, many of these data-intensive processes must remain localized and processed at the edge and on or near a hardware device.

The opportunity for AI accelerator chips is much-hyped, but how big is the market, and which companies are actually selling chips today? EETimes spoke to the reports' author, Principal Analyst Lian Jye Su, to gain some insight into which companies and technologies are making inroads into this potentially lucrative market. AI in the Cloud The first report, "Cloud AI Chipsets: Market Landscape and Vendor Positioning," highlights how cloud AI inference and training services are growing rapidly. The resulting AI chipset market is expected to grow from US$4.2 billion in 2019 to US$10 billion in 2024. Nvidia and Intel, the current leaders in this space, are being challenged by companies including Cambricon Technologies, Graphcore, Habana Labs and Qualcomm.