• Nvidia has published its latest State of AI in Telecommunications report
• It has many positive conclusions about AI adoption in the sector
• But, somehow, there are still some telecom sector companies not engaging with AI tools in any way, according to Nvidia’s survey results
• Nvidia uses the report as its latest clarion call for the deployment of AI-RAN systems by mobile operators

It’s increasingly difficult to find companies with digital technology at the core of their businesses that still claim to be neither using, testing or even planning to use AI applications as part of their operations, but Nvidia has found more than a dozen in the telecom sector, the AI chip giant notes in its latest State of AI in Telecommunications report. 

Late last year, Nvidia surveyed 450 telecom sector professionals (working at telcos, vendors, systems integrators etc) who are “the primary agents responsible for investing, implementing, and delivering AI to their companies” to ask a broad range of AI-related questions. The results are interesting, of course – who in the industry doesn’t want to get a sense of how everyone else is thinking about AI? – and cover the broad range of potential impact that the use of AI could have in the sector. And we’ll get to that in a minute.

But it’s intriguing that, two years after OpenAI’s ChatGPT generative AI (GenAI) application melted the tech sector’s collective minds and changed the way that many individuals and companies work and play, and many years after AI tools started having an impact on analytics and automation in the telecom sector, Nvidia’s survey found that almost 3% of respondents (about 12 individuals) said their companies are “not using or planning to use AI” at all. I find that remarkable, though this is apparently quite an advance from the 10% that said AI wasn’t in their corporate mix a year earlier. 

Now, of course, this means that more than 97% are engaged with AI – 49% are “actively using AI in their operations, up from 41% in 2023,” according to the Nvidia report, while about the same proportion are “in an assessment phase of trials or pilots”. 

Meanwhile, 65% of respondents said their companies plan to increase spending on AI infrastructure in 2025 – it is surprising that number also is not higher. 

There are some notable, though not surprising, trends related to AI investment priorities highlighted in the report, as the chart above shows. According to the report, “investment in AI will be spread across different priorities, but the primary theme is that spending will be concentrated on resources for further development and adoption of AI at scale. The top investment priority in the coming year, cited by 43% of respondents, will be to engage third-party partners, such as independent software vendors, global system integrators, and service delivery partners, to accelerate AI adoption. In addition, 40% of respondents said their companies will be investing in AI training for employees “to take advantage of the productivity companies can gain with AI solutions” – this, according to Nvidia, is also in response to one of the main challenges that telecom sector companies face, namely lack of in-house AI expertise. 

In addition, 35% said AI infrastructure will be an investment priority this year. As you might expect, all these numbers are all higher than the results from a year earlier, with AI training as an investment priority increasing the most notable amount. 

There are plenty of other stats from the telecom sector-specific report, the main ones being: 

• 84% reported AI is helping to increase their company’s annual revenue
• 77% said AI helps reduce annual operating costs
• 60% cited increased employee productivity as their biggest benefit from AI
• 44% are investing in AI for customer experience optimisation, which is the number one area of investment for AI in telecommunications
• 40% are deploying AI into their network planning and operations, including RAN.

You can get a copy of the report here, while Ronnie Vasishta, Nvidia’s senior VP of Telecom, has shared some thoughts on the results in this blog, where one of the topics he touches on is the use of AI in the radio access network (RAN).  

Adding AI into the RAN in multiple, different ways is a significant area of focus not only for mobile network operators but also for Nvidia, which is one of the founding members of the AI-RAN Alliance that was launched at MWC24 – see AI-RAN Alliance launches at #MWC24.

Using AI for network planning and operation (including in the RAN) is already one of the top AI use cases in the telecom sector – Nvidia and its alliance partners would like that number to be higher and for AI to play a much broader role in the RAN too, with integrated AI enabling operators to use the technology and offer AI services to customers too, all from the same IT stack. 

It’s a compelling idea but one that still generates a great many questions about cost, impact on energy efficiency and general feasibility, even if Japan’s SoftBank (another AI-RAN Alliance founder) is leading the charge, in partnership with Nvidia, of course – see SoftBank and Nvidia tie revenue model to new AI-RAN solution.

Expect to hear a great deal of noise about AI-RAN at MWC25 in Barcelona at the beginning of March – the Alliance’s founders and supporters are on a mission to get this architecture adopted by mobile operators as soon as possible but it looks like the barriers to adoption are going to be too high for many (though not all) operators for quite some time yet. 

- Ray Le Maistre, Editorial Director, TelecomTV