The long-term answer could be “potential.” When 4G was launched, people expected better smartphone performance, voice over LTE, new apps, but who could have predicted world-changing new services like Uber? Perhaps 5G is best explored as an open door to possibility, rather than enabling or enhancing specific applications.

Machine learning and AI rely on mining massive amounts of data, and an IoT can link thousands of sensors in the field to feed back that data. But even with the increased capacity of 5G, it makes sense to preprocess data closer to the edge. An autonomous vehicle, for example, will be a data center on wheels, making real-time decisions that require minimal latency while sharing environmental and performance data to and from the cloud. AI models developed in a central database will be pushed to the limit for real-world inference applications, while 5G feeds back ongoing data for continuous learning and model improvement.

See more: Trends for the deployment of 5G networks in 2021

Virtual reality and augmented reality require exceptional latency, and when deployed in critical applications such as medical diagnostics or surgical support, security and privacy become paramount. 5G network slicing allows QoS to be tailored to application needs, but optimizing multiple services in a limited and expensive spectrum band is a significant computational challenge only possible through sophisticated AI-informed automation.

How does one prepare for a future that opens up so many possibilities? Decoupling hardware and software to create a general-purpose 5G Telco *plus* AI computing platform offers greater flexibility in the future, but can it match the out-of-the-box performance of custom solutions?

We invited a select panel of industry visionaries to explore the potential that 5G offers in a world of intimate and exclusive connectivity, to discuss the challenges and suggest possible ways forward.