All artificial intelligence (AI) applications have one thing in common: They need a reliable, high-performance network connection. Sometimes it’s for downloading the AI software that runs on the device, such as Apple Intelligence on the iPhone. In other cases, the connection facilitates uploads to the cloud, where hyperscale compute resources enable the AI to analyze and act on data faster than the processor on the device.
5G is ideal for providing that connection, especially for what NVIDIA founder and CEO Jensen Huang recently termed “physical AI”: autonomous vehicles, drones, factory automation, and humanoid robots. Huang is so bullish on physical AI that he expects it to spur the “the first multi-trillion-dollar robotics industry.”
5G has a unique features and capabilities that, individually and collectively, meet physical AI’s demanding, often mission-critical requirements:
- Enhanced Mobile Broadband (eMBB) supports peak download speeds of 20 Gbps and peak uplinks of 10 Gbps. These are fast enough for bandwidth-intensive applications such as real-time machine vision and uploading assembly instructions to manufacturing robots.
- Ultra-Reliable Low Latency Communication (uRLLC) provides 99.99% uptime and latencies as low as 1 millisecond. This is ideal for mission-critical use cases such as autonomous vehicles shuttling raw materials and finished products on the factory floor or around a logistics park.
- Massive Machine Type Communication (mMTC) enables 5G networks to support up to 1 million Internet of Things (IoT) devices per square kilometer. This is a good fit for connecting the sensors, controllers, and other IoT devices that AI will rely on for applications such as Industry 4.0 factory automation.
Why 5G is Ideal for Physical AI
eMBB, uRLLC, and mMTC all far exceed the capabilities of 4G LTE, which is one reason why 5G is a better fit for physical AI. In fact, when 5G standards work began, a major focus was ensuring that it could support IoT and other industrial use cases in ways that 4G couldn’t. 5G was built from the ground up to enable exactly the kinds of applications and devices that comprise physical AI.
Another advantage is future proofing. The initial 4G networks and devices debuted around 2010 — a lifetime ago from a technological perspective. Much of the physical AI equipment being installed today and over the next few years will need to remain in service for a decade or more because they’re major investments that take a long time to recoup. So it makes sense to future proof them with the latest mobile technology rather than a legacy one.
Many factories traditionally used Wi-Fi for connecting fixed and mobile IoT devices. But in a 2021 Analysys Mason survey, 76% of manufacturers said they plan to switch to 5G because it provides greater security and reliability than Wi-Fi. Specifically, respondents plan to use private 5G networks, where they either own the base stations and other infrastructure or get a dedicated virtual slice of a public network.
Great Antennas are Key for Unlocking the Multi-Trillion-Dollar Opportunity
A great antenna is critical for ensuring that 5G lives up to its potential for physical AI applications. Whether they’re embedded, such as in an IoT sensor, or external, such as on an autonomous material handling vehicle, antennas maximize the link budget so physical AI has a consistently reliable, high-performance 5G connection to and from the cloud. That’s true for every 5G iteration: private, public, or hybrid, indoors, outdoors, standalone (SA), mmWave, or sub-6 GHz.
For over five years, Taoglas has been laying the foundation for physical AI. One example is our industry-leading portfolio of 5G new radio (NR) antennas for sub-6GHz, mmWave, embedded, and external applications:
- The new Taoglas Raptor X MA8007 is a low-profile, 7-in-1 combination antenna that’s ideal for routers and gateways in locations such as smart factories.
- The new Taoglas Raptor MAX series is available in configurations ranging from 12-in-1 to 15-in-1.
- The Taoglas Guardian series is available in a variety of configurations, including the 7-in-1 MA997.A.003, the world’s smallest, high-performance combination panel antenna.
- The Taoglas Synergy series includes five new models, such as the MA1557.W.003, a 7-in-1 antenna that’s ideal for mobile applications including autonomous vehicles in factories and logistics yards and cargo container tracking.
This breadth and depth make Taoglas a one-stop shop for device OEMs, systems integrators, and other companies targeting the physical AI opportunity.
Another example is webinars and other resources designed to help companies master 5G technology and then apply it to physical AI. For instance, the “Optimizing Sub-6 GHz 5G RF and Antenna Performance” webinar is a convenient way to get up to speed on aspects such as antenna tuning and integration. And the “Millimeter-Wave Arrays for 5G Applications” case study explores how beam-steering arrays can deliver throughput that rivals fiber.
