What Do Enterprises Believe the Future Holds for AI in Networking?

_{WHAT DO ENTERPRISES BELIEVE THE FUTURE HOLDS FOR AI IN NETWORKING?}

Which AI technologies will most transform networking in the next 3–5 years?

50.4%

Predictive and simulation technologies (analytics,
digital twins)

49.6%

Autonomous networking (self-healing, self-optimizing)

48.7%

Generative AI (network design, policy generation)

42.0%

AI for wireless and IoT (5G/6G, sensor networks)

32.7%

Edge AI and distributed
intelligence

29.3%

AI for security
and zero trust

_{WHAT DO ENTERPRISES BELIEVE THE FUTURE HOLDS FOR AI IN NETWORKING?}

By 2028, what percentage of your network operations do you expect will be:

Respondents believe predictive and simulation technologies (50.4%) and autonomous networking (49.6%) will be the most transformative AI technologies for networking. AI-powered analytics and digital twins shift network management from a reactive to a proactive model, while self-healing, self-optimizing networks have the potential to lift the heavy burden of management and amplify network efficiency and performance.

Other technologies in the spotlight include GenAI for tasks such as network design and policy generation (48.7%), and AI for wireless and IoT (42.0%).

When it comes to autonomous network operations, however, enterprises are hedging their bets about the near future. By 2028, respondents anticipate 38.4% of their network operations will be AI-assisted, 32.2% will be managed by humans in the traditional way, and only 29.4% will be autonomous.

_{WHAT DO ENTERPRISES BELIEVE THE FUTURE HOLDS FOR AI IN NETWORKING?}

^{Predictive and Simulation Technologies and Autonomous Networking Will Be Transformative}

What financial impact do you expect from AI in network operations by 2028?

12% increase

By 2028, enterprises expect AI in network operations to increase costs by an average of 12%.

This might seem surprising given that one of the early drivers of AI adoption was efficiency and cost reduction. However, this picture has begun to change. A recent study by MIT, for example, found that 95% of GenAI pilots don’t deliver measurable financial impact.

So why do our respondents expect AI-driven networks to cost more than traditional ones?

Most enterprises are implementing AI across the business to be more productive and competitive. But AI is more demanding of the network, which means it needs greater capacity and better performance. And as we’ve seen, network planning and optimization are two of the top four AI in networking use cases that enterprises are piloting or implementing.

Deploying these AI network models, however, has costs associated, like additional computing power, storage, data engineering, integration, and attracting and retaining skilled professionals. However, without AI-driven networking solutions, it’s likely that the excessive cost of greater capacity and poorly utilized bandwidth would be even higher.

_{WHAT DO ENTERPRISES BELIEVE THE FUTURE HOLDS FOR AI IN NETWORKING?}

^{Enterprises Expect Cloud Hyperscalers To Become Major AI-In-Telecoms Players}

44.9% of respondents expect cloud hyperscalers to lead the AI-in-telecoms race by 2030, reflecting the significant investments these massive companies are making in their own terrestrial and satellite communications infrastructure. It remains to be seen, though, if these companies will be able to provide the data sovereignty and local presence that many enterprises need for compliance and for solutions that truly fit their unique business needs.

This is followed by specialized AI/managed intelligence providers (22.2%) and network equipment vendors (13.8%). Only 9.1% of respondents believe that traditional telecoms operators will be at the forefront of the AI-in-telecoms arena.

Who do you think will lead the AI-in-telecom race by 2030?

Enterprises see AI in networking operations offering significant potential value as other AI use cases place an increasing load on their infrastructure, but they appear to be struggling with conflicted feelings about security and strategic approaches.

CONCLUSION

Enterprises see AI in networking operations offering significant potential value as other AI use cases place an increasing load on their infrastructure, but they appear to be struggling with conflicted feelings about security and strategic approaches.

Almost all organizations are having trouble recruiting people with the skills and experience to implement AI safely and effectively, so they’re likely to seek out external partners to help. Security and compliance fears, however, could be a barrier to finding vendors and providers that they feel confident working with.

_{^{This report draws on a survey of}}_{^{490 IT and networking professionals}}_{^{— including CIOs, CISOs, network operations leaders, and enterprise architects — conducted across two research panels to ensure breadth of perspective. Respondents represent organizations ranging from 1,000 to 25,000+ employees, with all findings reported as a consolidated summary across the full sample.}}