NVIDIA released its third annual State of AI in Telecommunications Report, reporting that 97% of respondents are “assessing or adopting AI with the goals of enhancing customer experiences and employee productivity, improving network operations, reducing costs, and opening new business opportunities.” Half of respondents said that they have already implemented their first generative AI use case.
There’s no reference for the satellite portion of the telecom industry, but even if implementation in our corner of the market lags, many see the potential benefits and many satellite, mobile and terrestrial network operators are exploring options for managing network operations. Unfortunately, the state of our legacy hardware ground systems may slow us down even more. It is simply much harder to implement robust and dynamic optimization and automation for an analog system. There are limits to what you can do when it comes to collecting and fusing data for the AI algorithms to use.
In contrast, virtualizing ground infrastructure with a digital IF architecture at or close to the antenna provides easy access to any data required to enable optimization or automation. Data from every stage of routing and processing is accessible making it far easier to collect, train and deploy AI and ML models.
For example, suppose an operator wanted to optimize network performance and minimize teleport power requirements by automatically adjusting network configuration and amplifier gains based on historical and actual traffic patterns. That requires data from multiple diverse systems that are today managed in different, stovepiped ways. With virtualized and orchestrated signal routing and processing, however, network and teleport infrastructure run in the same compute environment, enabling the AI to access the full range of applicable information. What’s more, it enables incorporation of longer-term trending data that can’t be accessed via hardware today.
According to NVIDIA, integrating AI Into network operations is gaining traction at communications service providers. “Investing in AI solutions for network infrastructure has become a growing priority within the telecom industry. Network planning and operations, including integrating AI into the radio access network, was cited by 37%
of respondents as an investment priority, while another 33% said they’re investing in AI for field operations optimization.”
AI is one more growing surface for satcom integration with terrestrial infrastructure. Combine it with global initiatives uniting the two worlds behind 5G NTN and 6G, and the opportunity for satellite to grow market share looks even brighter. In fact, the NVIDIA report specifically cites using AI to monetize 5G and research and development of 6G networks as key investment areas for telecom companies. But it will require our respective systems to work together better than they do today.
