Telecommunications companies are increasing their investments in artificial intelligence as network automation moves ahead of customer service as a leading application, according to NVIDIA’s 2026 State of AI in Telecommunications report.

The survey found that 66 percent of respondents said their organizations were actively using AI, up from 49 percent in 2024 and 41 percent in 2023. Meanwhile, 89 percent said they planned to increase their AI budgets in 2026, compared with 65 percent who reported the same intention in the previous year’s survey.

The findings point to a shift from testing AI applications to deploying them across business operations, with operators pursuing greater network efficiency, lower operating costs and new revenue opportunities.

NVIDIA said the fourth annual survey received more than 1,000 responses worldwide. About a quarter of respondents came from telecommunications operators, including multiple-system operators, while the remainder represented network equipment providers, systems integrators, software vendors and other industry participants.

Network automation takes the lead

Network automation emerged as the leading AI use case for investment, deployment and return on investment, overtaking customer experience and support, which had ranked higher in previous editions of the survey.

Overall, 54 percent of respondents said their organizations were using AI for network automation covering planning, field deployment, operations and optimization, a 17-percentage-point increase from 2024. Customer service and experience optimization followed at 46 percent, while internal process optimization across functions such as information technology, human resources and finance reached 43 percent.

Among telecommunications operators, 65 percent said AI was driving network automation, while 59 percent identified it as one of the leading AI use cases contributing to return on investment.

The report also found that 41 percent of respondents already deploying agentic AI had applied it to network automation. Another 52 percent expected AI agents for network automation to deliver significant returns within the next two years.

These developments are contributing to the industry’s pursuit of autonomous networks, in which software and AI systems increasingly handle network management and optimization with less human intervention.

According to the report, 88 percent of respondents were at Levels 1 to 3 on TM Forum’s Autonomous Network Levels spectrum. The framework ranges from Level 0, which represents no automation, to Level 5, where a network operates with full autonomy.

Among telecommunications operators and multiple-system operators, 35 percent placed their networks at Level 2 and 36 percent at Level 3. Nineteen percent reported Level 1, indicating that most operators remain in the early to intermediate stages of network autonomy.

AI delivers measurable business gains

The growing investment in AI reflects reported improvements in revenue, costs and employee productivity.

Ninety percent of respondents said AI was helping increase annual revenue and reduce annual costs. Specifically, 67 percent reported revenue increases exceeding 5 percent, while 55 percent reported annual cost reductions exceeding 5 percent.

The survey also found that 99 percent of respondents said AI had improved employee productivity, with 26 percent reporting major to significant improvements.

Beyond operational efficiency, telecommunications companies are exploring AI as a source of new business. Forty-three percent of respondents cited opening new business opportunities as a primary objective of their AI initiatives, while 46 percent said AI had helped their businesses by creating additional revenue opportunities.

Potential offerings include AI-as-a-service, AI platforms and graphics processing unit computing services for enterprise customers, governments and industry-specific markets.

For telecommunications providers, these services could extend their commercial role beyond connectivity, allowing them to offer computing infrastructure and AI capabilities to businesses seeking to develop or deploy their own applications.

Generative AI expands as agentic AI gains ground

Generative AI remained the most widely reported AI workload in the survey, with adoption reaching 60 percent, up from 49 percent in 2024 and 43 percent in 2023.

Data processing and analytics ranked second at 58 percent, followed by agentic AI at 48 percent. The latter includes systems designed to carry out tasks through AI-driven workflows and actions, rather than simply generate responses to prompts.

Among respondents from companies with more than 10,000 employees, 80 percent reported using generative AI.

The report said AI applications were expanding across network operations, customer experience and internal processes, including information technology and human resources. Telecommunications operators were also applying AI to wireless network research and development, including work involving 5G, AI-enabled radio access networks and 6G.

However, adoption is not without obstacles. Fifty-four percent of respondents cited data-related issues as their biggest challenge in achieving AI objectives, up from 20 percent in 2024. These issues included privacy, data sovereignty, siloed information, data volume and complexity.

A shortage of AI experts and data scientists was the second-largest challenge, cited by 47 percent of respondents. Another 34 percent said measuring return on investment remained difficult.

The report identified data quality, fragmentation and sensitivity as important considerations as telecommunications companies move AI applications from pilot projects into production.

Computing investment points toward AI-native networks

The survey also indicated growing interest in distributing computing resources closer to network users to support AI inference, the process through which trained AI models generate outputs.

Seventy-eight percent of respondents said their organizations were investing or planning to invest in computing at the network edge to support AI inference. Meanwhile, 77 percent expected AI-native wireless networks to be deployed before the launch of 6G, which the report places around 2030.

AI-native radio access networks, or AI-RAN, are intended to integrate AI capabilities into wireless network architecture. The report said these approaches could improve network utilization, spectral efficiency and energy efficiency while allowing AI and radio access network workloads to share infrastructure.

Among respondents whose organizations were using AI for network research and development, 65 percent cited improved spectral and energy efficiency as a primary objective for investing in AI-native networks.

The survey suggests that the development of AI-native infrastructure may advance alongside existing network generations rather than wait for commercial 6G deployments.

For the Philippine telecommunications industry, these global findings provide a reference point for assessing how operators could apply AI to network operations, enterprise services and future wireless infrastructure. However, the survey does not establish the adoption rate, investment plans or expected returns of individual Philippine operators.

The extent to which these trends translate into local benefits will depend on deployment decisions, available skills, data governance, infrastructure investment and the ability to demonstrate measurable improvements in network performance and customer service.