Telcos shifting capex to AI infra, says Omdia

By Juan Pedro Tomás April 8, 2026

Collected at: https://www.rcrwireless.com/20260408/carriers/telcos-capex-ai-omdia

Hyperscalers continue to dominate AI infrastructure globally, but regulatory dynamics are opening space for telcos in specific segments, observes Omdia

In sum – what to know:

Demand-led models – Telcos must secure anchor tenants and real workloads early, as AI infrastructure economics depend heavily on utilization and long-term customer pipelines.

Sovereignty gap – Regulatory requirements are opening targeted opportunities for telcos, even as hyperscalers maintain overall dominance in AI infrastructure.

Execution risks – High costs, uncertain demand, and potential trade-offs with core network investment create financial and operational challenges over the next three to five years.

Telecom operators are increasingly redirecting capital toward AI infrastructure, but early monetization will depend less on scale and more on execution, according to an Omdia analyst.

Julia Schindler, principal analyst at Omdia, told RCR Wireless News that operators that succeed in AI infrastructure are those that anchor investments in real demand rather than speculative buildouts.

“Telcos best positioned to monetize AI infrastructure treat it as a demand-led business case. AI data centers, for example, carry high upfront and ongoing costs — power, cooling, operations — which makes it pivotal to secure sufficient occupancy through anchor projects and early customer pipelines,” she said.

Rather than pursuing standalone infrastructure plays, Schindler pointed to the importance of leveraging existing enterprise relationships. “Many telcos leverage existing relationships with public-sector bodies, regulated enterprises, and R&D programs, such as local LLMs, to secure them as anchor tenants.”

Beyond securing tenants, monetization strategies are also evolving. Operators are increasingly embedding AI infrastructure into broader enterprise portfolios. “Monetization not only depends on occupancy but also on how the infrastructure is embedded into the broader telco portfolio. Rather than only reselling GPU capacity, some telcos bundle and cross-sell across their B2B portfolio, including private cloud, security, and business applications, creating a more unique value proposition,” she said

At the same time, hyperscalers continue to dominate AI infrastructure globally, but regulatory dynamics are opening space for telcos in specific segments. “Hyperscalers have dominated AI infrastructure in the past and will continue to do so across many countries and workloads, given the scale and pace of their investment,” Schindler said.

However, the analyst said that sovereignty requirements are reshaping the competitive landscape. “Sovereignty requirements, however, are creating room for telcos to grow, as they can play a meaningful role in sensitive and highly regulated segments. At the same time, hyperscalers have already responded with sovereign offerings and are staying close to telco partners to avoid exclusion from key projects.”

Government policies are emerging as a key variable. “The balance between hyperscaler-led and telco-led deployments will ultimately hinge on the rigor of sovereignty rules: where rules are strict—China is a clear example — local providers are scaling and the market is shifting; in many other markets, sovereignty frameworks are not yet clearly defined or consistently implemented,” she added.

Despite the opportunity, the shift toward AI infrastructure introduces new financial and operational risks. “Shifting capex from traditional connectivity to AI infrastructure is a strategic bet, and outcomes will vary by market conditions, regulatory expectations, and each operator’s approach,” Schindler said.

She warned that uncertainty around demand and cost structures could weigh on returns. “Financially, demand and monetization are still uncertain and fast technology cycles will require ongoing reinvestments, while energy and cost of capital will also influence project economics.”

Operational trade-offs are also a concern. “Operationally, some operators may face trade-offs with core network capabilities that could affect competitiveness, though the extent will be context dependent,” she said, adding that partnerships and phased investments will be critical to managing exposure.

Earlier this year, Omdia had published a report which highlights how operators across regions are scaling multi-year investments in data centers, GPU-as-a-service, and AI-driven network capabilities. The report shows early signs of monetization, with data center and digital infrastructure revenues beginning to contribute a growing share of total operator income, while also emphasizing that there is no single investment model across the industry.