Nvidia and AI: Transforming revenue in telecommunications

Table of Contents

• 1. Report findings
• 2. Impact of artificial intelligence in telecommunications
• 3. Key results of the Nvidia survey
  • 3.1 Positive effects on revenue and cost management
  • 3.2 Expectations for AI-native networks
• 4. Adoption of generative AI in operations
• 5. Increase in investment in artificial intelligence
  • 5.1 Budget growth projections
  • 5.2 Investment in network infrastructure and edge computing
• 6. Transformation of network functions through automation
• 7. Transform your customer experience with Suricata Cx
  • 7.1 Cost optimization and operational efficiency
  • 7.2 Improvement in customer satisfaction
  • 7.3 Scalability and adaptation to future needs

self-configure, self-heal, and self-optimize with minimal human intervention, a shift that affects everything from day-to-day operations to readiness for next-generation services.

AI-native core automation
– “Autonomous network”: operation with automation loops that detect conditions (failures, congestion, degradation), execute actions (reconfigure, balance, isolate), and verify the outcome.
– “AI-native”: AI is not an “add-on”; it is designed to be integrated into operational workflows (observability, assurance, optimization) and into the data/telemetry architecture.
– Why it displaces CX as a priority: when you automate the core (fewer outages, better performance, less manual work), the effect propagates to service quality and, by extension, the customer experience.

In parallel, return on investment (ROI) is associated with levers such as autonomous networks, improvements in customer service, and optimization of internal processes. The underlying takeaway is clear: AI is being integrated as a cross-cutting “layer” to reduce operational friction and enable new capabilities, while operators prepare for the technological evolution that precedes the 6G era.

Key results of the Nvidia survey

The Nvidia report, based on responses from operators and ecosystem participants, sketches an adoption map that combines business impact, technical priorities, and expectations for network evolution. Two signals stand out: the perception of benefits already materializing (revenue and costs) and the conviction that “AI-native” networks will arrive before commercial 6G.

Responsible interpretation of results
How to read these results without overinterpreting them:
– Perception vs. measurement: several findings are “what respondents report” (impact/ROI), not necessarily financial audits or standardized KPIs.
– Expectations vs. actual deployment: “expecting AI-native before 6G” reflects intent/planning; it does not imply that the entire network already operates autonomously.
– Who responds matters: by including “operators and ecosystem participants,” responses may mix priorities from operations, vendors, and partners.
– Useful for decisions: treat the percentages as a trend signal (priorities and investment direction) and validate internally with your metrics (OPEX, failures, energy, resolution times).

The survey also suggests a reshuffling of the agenda: AI is no longer limited to “improving” existing functions, but is being used to redesign how networks and processes are operated. In that context, network automation appears as the main

destination for efforts, beyond the best-known front-office uses.

Positive effects on revenue and cost management

The report’s most compelling finding is the breadth of consensus: approximately nine out of ten respondents state that AI has had a positive impact on both annual revenue and cost management. In other words, the report reflects participants’ stated perception of business impact and efficiency, along with the areas to which they attribute that return. In a sector where returns are often demanded quickly, this majority perception serves as an indicator of maturity: AI is being evaluated by results, not by novelty.

Respondents attribute ROI mainly to three drivers. First, autonomous networks, which promise operational efficiencies by automating repetitive tasks and reducing manual intervention. Second, improvements in customer service, where AI can speed up responses and resolve inquiries with less friction. Third, optimization of internal processes, a broad category aimed at automating workflows and operational decisions.

Overall, the message is that AI is impacting the P&L through two simultaneous paths: growth (revenue) and efficiency (costs). And it is doing so not only at the organization’s “edge,” but at the core of how the network is operated and processes are executed.

Expectations for AI-native networks

The survey captures a relevant timing expectation: around 77% of operators say they expect AI-native networks to be launched before commercial 6G. The implication is strategic: the industry is not waiting for the next generation of radio standard to transform its operations; it is pushing a prior evolution based on automation and integrated AI capabilities.

This approach aligns with the idea of networks managed with greater autonomy: systems that detect conditions, adjust parameters, and optimize performance with less reliance on human intervention. In terms of readiness for future services, AI “nativeness” suggests that intelligence will not be an add-on, but part of the operational design.

The survey also mentions that open-source models and software are considered critical components of operators’ AI strategies. This points to an ecosystem where technological flexibility and adaptability matter as much as performance: telcos are looking to build AI capabilities that can evolve quickly, integrate with existing systems, and be sustained over time.

Adoption of generative AI in operations

generative in operations

Generative AI is gaining ground in telecommunications operations, not as an isolated experiment, but as a tool under evaluation or in active use. Nvidia’s report indicates that 60% of respondents say they are already using or evaluating generative AI in their operations, up from 49% in 2024. The jump suggests an acceleration in deployment intent and, above all, a normalization of the conversation: generative moves from “possibility” to “roadmap.”

From Evaluation to AI Operations
A short path to move from “evaluating” to “operating” generative AI (with checkpoints):
1) Select 1–2 high-volume use cases (e.g., ticket summarization, agent assist, knowledge base search).
– Checkpoint: is there enough data and a process “owner” (Ops/CX/IT)?
2) Pilot with simple metrics (resolution time, recontacts, escalation-to-human rate, response quality).
– Checkpoint: define what “better” means before testing.
3) Minimal integration with systems (tickets/CRM/KB) to prevent the pilot from being just “chat.”
– Checkpoint: does the flow leave traceability (what was answered and why)?
4) Operational controls (human-in-the-loop, action limits, monitoring of errors and biases).
– Checkpoint: what happens when the model doesn’t know or makes a mistake?
5) Phased move to production (by reason, by channel, by segment) and monthly review of metrics.

Although the report does not detail specific use cases by operator, it does frame the return in areas where generative can fit: customer service improvements, optimization of internal processes, and support for automation initiatives. In telecom, where repetitive interactions and standardizable procedures abound, generative is shaping up as a complement to speed up flows, summarize operational information, or assist in support tasks.

Adoption, however, coexists with a shift in focus: network automation is positioned ahead of customer experience as an AI priority. This does not eliminate the customer service front, but it does suggest that the industry is moving part of the effort toward the operational “backbone”: if the network becomes more autonomous, the benefits can propagate across the entire chain, including service quality and, by extension, user experience.

In that context, generative AI appears as one more piece within a broader strategy: building more automated operations, with less friction and the capacity for continuous adaptation.

Increase in investment in artificial intelligence

Investment accompanies the shift in priorities. Nvidia’s report notes that almost

nine out of ten telecommunications companies plan to increase their AI budgets in 2026. Beyond the percentage, the figure reflects a conviction: AI is no longer treated as discretionary spending, but as a strategic line to sustain competitiveness, efficiency, and technological readiness.

Key AI Investment Decisions
Typical trade-offs when the AI budget increases (and why they matter):
– Capex vs Opex: more owned infrastructure can lower unit costs at scale, but increases upfront investment and operational burden.
– Edge vs Cloud: edge helps with latency/data sovereignty and local resilience; cloud accelerates deployment and elasticity. Many telcos end up hybrid.
– Open source models vs closed vendor: open source provides flexibility and control (and requires internal capability); a vendor reduces operational effort, but can increase dependency and recurring costs.
– Network automation vs CX: automating the core often delivers ROI through continuity/energy/OPEX; CX tends to impact NPS/retention. The priority depends on where it “hurts” most today.

The survey also links investment with architecture: operators are strengthening capabilities to run AI in environments that require low latency and proximity to the user, which drives decisions about network infrastructure, computing, and edge deployments. In parallel, the report highlights the role of open source models and software as critical components, suggesting strategies that combine infrastructure investment with technology stack decisions.

Budget growth projections

The budget increase is not marginal. According to the survey, 35% of operators expect AI spending growth above 10% year over year. In an environment where telcos often balance investments among spectrum, network, modernization, and operations, this expectation indicates that AI is competing strongly for resources.

The figure also aligns with the shift toward network automation: if the goal is to enable more autonomous networks—capable of self-managing—spending is not limited to licenses or pilots, but can encompass platforms, integration, and compute capabilities.

In addition, the emphasis on open source as a critical component suggests that part of the budget may be directed toward adapting, integrating, and operating models and software with greater control, rather than relying exclusively on closed solutions. In strategic terms, this can be interpreted as a search for flexibility and a technological foundation that enables rapid iteration as needs and services change.

Investment in network infrastructure and edge computing

The report also points to a specific investment destination: edge computing and AI-native wireless infrastructure. The logic is operational and commercial: bringing processing closer to users helps support emerging applications and meet latency and performance requirements, while AI embedded in the infrastructure can enable finer-grained automation and optimization.

Industry contributors cited in the report describe autonomous networks as a path to fast ROI due to their ability to reduce outages, lower energy consumption, and automate repetitive workflows. Those benefits, although not quantified in the note, explain why investment is directed to deep layers of the network: if the incidence of failures is reduced and energy is optimized, the impact is reflected in operating costs and service continuity.

Overall, investment in edge and in AI-native infrastructure suggests a transition: it’s not just about “using AI,” but about building a network and an operation prepared for AI to be part of day-to-day functioning.

Transformation of network functions through automation

Network automation emerges as the axis of the transformation. Nvidia’s report indicates that this approach has surpassed traditional customer experience applications as the top priority for AI deployment, with an emphasis on networks capable of self-configuring, self-healing, and self-optimizing with minimal human intervention. The shift is significant: it means moving intelligence to the heart of operations, where configuration, maintenance, optimization, and incident response are managed.

Network Autonomy by Function
Quick map: “self-configure / self-heal / self-optimize” grounded in real network functions
– Configuration (self-configure): service provisioning, parameter changes, policies; goal: fewer human errors and more consistent deployments.
– Assurance (self-heal): anomaly detection, alarm correlation, root-cause isolation; goal: lower MTTR and fewer outages.
– Optimization (self-optimize): load balancing, capacity tuning, traffic prioritization; goal: better performance with the same resources.
– Energy (continuous optimization): intelligent shutdown/hibernation, dynamic tuning; goal: lower consumption without degrading SLAs.
Practical takeaway: the more mature telemetry and automation are, the more “autonomous” the network can be without losing control.

The promise of networks

autonomous—according to what is stated in the report—is to operate with minimal human intervention, with self-configuration, self-healing, and self-optimization capabilities. In practical terms, this translates into fewer repetitive manual tasks and more consistent operation, especially in complex environments where the network combines multiple layers and technologies.

The expected benefits tie into the ROI mentioned by respondents: fewer outages, lower energy consumption, and workflow automation. Although the report does not go into metrics, it does make the intent of the bet clear: to automate core functions to free up operational capacity, reduce costs, and sustain service levels.

The transformation is also linked to the technology timeline. If most expect AI-native networks before commercial 6G, automation becomes a “bridge” to the next stage: a smarter network not only because of speed or standard, but because of its ability to manage itself and adapt continuously.

Transform your customer experience with Suricata Cx

Cost optimization and operational efficiency

Suricata Cx is an AI-powered omnichannel customer experience platform designed specifically for ISPs and telecommunications operators. Its approach combines automation with “human-in-the-loop” workflows, aiming to reduce cost per interaction and speed up response and resolution times, especially for high-volume repetitive inquiries.

Improvement in customer satisfaction

By unifying channels such as WhatsApp, webchat, and social media into a traceable operation, Suricata Cx aims to reduce fragmentation and improve conversational continuity. Automation is complemented by controlled escalation to human agents when judgment or validation is required, maintaining operational control.

Scalability and adaptation to future needs

In a context where the industry is accelerating AI adoption—including generative AI—Suricata Cx is positioned as a foundation for scaling operations without relying exclusively on human models. The combination of automation, oversight, and operational metrics makes it possible to adjust operations as volumes and complexity grow.

Integration of advanced technology in telecommunications

Suricata Cx is geared toward deep operational integrations, with an API-first architecture and real-time synchronization with the operator’s systems (customers, debts, services, tickets). In addition, it incorporates payment and collections flows with Pagoralia, a gateway specialized in ISPs and telecom, to enable

reminders, conversational collections, and service reactivation after payment.

A customer-centric approach to sustainable growth

The operating thesis is hybrid: automate what’s predictable, keep humans in control where it matters, and measure performance with operational indicators (response and resolution times, recontacts, SLAs by reason). In a sector where AI is already associated with commercial impact, this approach seeks to turn efficiency and operational consistency into retention, growth, and a more reliable customer experience.

Key implementation metrics
What “operational evidence” is worth looking at (and reporting) when implementing Suricata Cx in an ISP/telco:
– Before / after (contact KPIs): first response time, resolution time, recontact rate, % of conversations resolved without an agent.
– Quality and control: % escalation to a human by reason, reasons with the highest error, conversation auditing (traceability).
– Cost impact: cost per interaction, agent hours freed up on repetitive reasons, SLA compliance by channel.
– Revenue/collections impact (if Pagoralia applies): promise-to-pay rate, actual payment rate, time to reactivation after payment.
Examples of typical (measurable) use cases:
– Support: guided reboot/diagnostics, incident status, ticket follow-up.
– Account and billing: balance, billing cycle date, invoice copy, arrangements.
– Conversational collections: reminders, payment links, confirmation and reactivation.

“Nvidia and AI: Transforming revenue in telecommunications” confirms that the impact is already being measured in P&L, with network automation and more efficient operations as real priorities. From Suricata Cx, this same logic translates into bringing AI applied to omnichannel operations with deep integrations and human control, to turn efficiency and consistency into a more reliable and sustainable customer experience.

This analysis is framed within Suricata Cx’s approach to AI applied to telecom and ISP operations: automate what’s predictable, maintain human control where it matters, and prioritize operational integrations so that the impact is reflected in service and efficiency metrics.

The percentages and conclusions on adoption and investment reflect responses and expectations stated by participants and may not represent actual results. In practice, results may vary depending on the market, network maturity, data availability, and integration with existing systems. This information is based on public data available at the time of writing and may change as plans and disclosures are updated.