2026年及未来：边缘人工智能（AI）变革之路

Edge AI is the fastest growing part of the AI wave, with a 37%CAGR through 20301compared to 28% CAGR2of the overall AI market. It has gained significant traction in the last 5+ years - even asthe world has begun a transformation based on our collective “ChatGPTmoment” in 2022. By collaborating with leading technology developers, academia andresearchers and industry leaders, we can explore what lies ahead in 2026for this transformative technology. Let’s understand what edge AI is, why itis a game changing technology of our times, and how we can apply it in2026 and beyond.. Introduction For the past decade, “cloud-first”has been the default mantra intechnology. The cloud gave usscale, flexibility, and global reach,but here is the truth: cloud-onlyarchitectures are starting to showtheir limits, and businesses thatrely on speed, resilience, andprivacy are noticing. When meeting with customers, weoften hear “We love the power of AI (meaning the cloud), but we can’t afford the latency.” When you need tostop a robotaxi for a wandering pedestrian, or a conveyor belt to avoid aproduct defect, or detect a safety risk in real time, waiting for round trips tothe cloud is not good enough. The same goes for creating truly seamlesspersonal experiences in the smart home - seeing your light turn onseconds after the voice command feels pretty rocky. Edge AI has now entered thediscussion at scale,challenging the previousassumption of “pusheverything to the cloud.” Thisdoes not mean the cloud goesaway, but rather it means theedge is leveraged where real-time decisions and low-power implementations matter most. Thefuture is not cloud versus edge, but a new balance – where the edge canlead as the forefront of intelligence, with the cloud providing scale,retraining & oversight. While there are many contributing factors to the increasing popularity ofedge AI, three drivers stand out to be the main enablers.The first driverbeing an explosion of available data:The sheer amount of data that isbeing generated by connected devices (75% of data is created at theedge3) and edge deployed devices and sensors continue to grow. Edge AIprovides a viable path for making use of this data. For reference, there willbe a projected 40.6 billion IoT devices in the world by 20344. The second driver is the surge in compute performance:Neural Processing Units (NPUs) bring a tremendous increase in Machine Learningperformance to the Edge - all the way from MPUs to extremely constraineddevices, like Microcontrollers (MCUs). This paradigm shift enables amassive increase in edge learning use cases. The market for edge AIdevices is projected to hit close to 5.7 billion edge devices by 20315,presenting a huge market opportunity, and likely not yet projecting fullmarket potential when comparing it to overall edge devices. The third driver is an unsustainable energy and resource path for cloud AI. Much has been written about the dramatic strain that datacenters areputting on the energy grids and systems of communities around the world.Data centers can consume 40% of a community electricity budget. Inaddition, a single large-scale data center can consume up to 5,000,000gallons of water per day.6Finally, data centers are like the heavy metalconcerts of the digital world, with average noise levels cranking up to adeafening 92-96 dB(A). Moving AI workloads to the edge, where the data iscreated, lowers energy, lowers cost and increases impact. What is edge AI and why should we care? Edge AI means the deployment of AI algorithms and models directly onlocal edge devices to enable real-time data processing and responseswithout constant reliance on cloud infrastructure. Many names have beenused for this under the aegis of edge AI such as physical AI, embodied AI, tinyML and more. This can range from sensor to server, although wetypically define “edge” as computing outside the multi-tenant data center.In layman’s terms, it’s “AIin the real world.” When is edge AI the best choice in a use case? There is a fantastic, easy toremember acronym that Daniel Situnayake & Jenny Plunkett (Speelman)put out in their book " Answer: Do the BLERP check! It is short for: B = Bandwidth: Reduced need for constant cloud connectivity, minimizingbandwidth requirements and costsL = Latency:Real-time processing and analysis enables fasterdecision-making in applications where time is criticalE = Economics:Optimized resource utilization, reduced energyconsumption, and connecting only when it mattersR = Reliability:Reliable operation, even in areas with limited or noconnectivity (like the jungle, or the German periphery)P = Privacy:Data never leaves your device and is processed securely &privately - a game changer for sensitive use cases around the home orhealth If any of the above are relevant to the application at hand, edge AI is likelythe better solution to the problem. An example for where “E”, “R”, and “P”matter are smart wearables - typically in a wearable we care aboutb