T-Mobile's 2026 AI and 5G Capex Strategy for Network Leadership

T-Mobile US: AI integration and 2026 capex implications for 5G network leadership

1) AI integration plans (what T-Mobile says it will put into the network)

A. “Network-native AI” starting with live translation (early-stage rollout)

• T-Mobile says it is “excited” to roll out live translation on beta soon, describing it as its first network-native AI application demonstrated in February. ¹
• Management frames live translation as using language learning models embedded into T-Mobile’s core and translating voice into 80 languages anywhere in the world using a single connected T-Mobile phone. ¹

B. Building AI capabilities directly into the 5G network core (longer-term integration path)

• T-Mobile explicitly links the live-translation pilot to a longer buildout: it says the initial step is building AI capabilities directly into the network core. ¹
• The company also describes a longer-term vision where the network becomes “connective tissue for physical AI” and supports inferencing at the edge. ¹

C. Using 5G Advanced network as an “AI + edge compute” platform for physical AI

• T-Mobile’s CEO characterizes its “big advantage” as having already rolled out 5G Advanced nationwide, and argues that it’s built for physical AI with features aimed at uplink capacity (important for edge/robotics use cases). ²
• Specific network innovations cited include improvements to spectral efficiencies and capacity, especially for the uplink, including uplink transmit switching, higher transmit power, and uplink MIMO. ²
• In response to questions about inference at the edge and whether it requires GPUs, management says T-Mobile has “already started introducing large amounts of AI into our network,” and as it moves “closer towards AI RAN,” it will build compute into the network. ³
• Management describes generating “fallow compute” especially at the edge (in analogy to FWA), and emphasizes that low latency + fallow compute creates an opportunity for physical AI and automation/robotics use cases. ³

D. Concrete 5G Advanced integration with Figure AI humanoid robots

• T-Mobile says it will connect its 5G Advanced network to Figure AI’s F03 humanoid robots, enabling connectivity from the moment robots power on. ¹
• Management says the partnership (among others) will let it explore how T-Mobile’s 5G Advanced capabilities—including network edge assets—can support the evolution of physical AI. ¹

E. “AI in the network” already deployed in operations (example: self-healing automation)

• T-Mobile also points to operational deployments: it says AI in the network was a “big reason” for real-world automated actions during “Winter Storm Fern,” such as antenna tilt being done automatically and network optimization, describing a “self-healing network” as reality rather than fiction. ³

---

2) 2026 capital expenditure implications tied to 5G network leadership and AI integration

The provided excerpts discuss AI integration and network capability direction, but they do not state a specific 2026 incremental capex line item dedicated to AI (e.g., “AI capex of $X in 2026”). Instead, they indicate where capex needs would likely concentrate based on management’s integration roadmap (AI embedded into core, AI RAN progression, edge compute).

What the excerpts do support about capex drivers:

• T-Mobile states it is introducing large amounts of AI into its network and expects that as it moves toward AI RAN, it will build compute into the network. This implies ongoing infrastructure investment to support compute placement (core/edge) and AI/RAN automation. ³
• It further describes that AI compute needs will be supported by creating “fallow compute” at the edge, combined with low latency requirements—again implying that network design/capacity planning and associated spending must support edge compute/latency performance for physical AI. ³
• For 5G leadership context, T-Mobile argues its 5G Advanced nationwide deployment provides the platform for physical AI and edge inferencing, supported by uplink and efficiency enhancements. Those kinds of enhancements typically require ongoing network modernization and engineering support, consistent with sustained capex aimed at preserving/expanding 5G Advanced performance leadership. ²

Capital allocation / investment posture in the same period (not explicitly “AI capex,” but relevant to 2026 spending envelope):

• In the earnings Q&A, a shareholder references an increase in capital allocation for the year of $3.6 billion (and contrasts it with $18.2 billion as part of the same capital allocation discussion), and asks about appetite for additional share repurchases. ⁴
• The excerpts also contain commentary that capital allocation philosophy prioritizes “investing in the core business,” “value-accretive M&A,” and then shareholder returns (dividend + buybacks). ⁵
• However, the excerpts do not quantify 2026 total capex, nor do they break out capex between AI/edge compute and other network or non-network investments.

5G network leadership continuity through integration

• T-Mobile discusses a timeline for UScellular customer migration (integration effort “over the spring, the summer and the fall,” with “substantially have it wrapped up by this year” as stated in the call) and characterizes the result as a “bolstered network advantage” in smaller markets and rural areas. While this is operational/integration-oriented rather than AI-specific, it reinforces that 2026 includes continued network and customer base integration work. ⁶

---

Direct answer: “plans for AI integration and capex related to 5G network leadership in 2026”

AI integration plans for 2026 (from the excerpts)

• Deploy network-native AI application: Roll out live translation on beta soon, with models embedded in the network core translating voice into 80 languages. ¹
• Push AI deeper into the network core and edge: Build AI capabilities in the network core, evolving toward a network that supports inferencing at the edge for physical AI. ¹
• Move toward AI RAN and network compute: Continue introducing AI into the network and progress toward AI RAN, including building compute into the network (especially to enable edge inferencing). ³
• Demonstrate physical AI connectivity: Connect T-Mobile’s 5G Advanced network to Figure AI’s F03 humanoid robots, using network edge capabilities to explore physical AI evolution. ¹
• Operationalize AI automation in the field: Use AI-driven automation to optimize and “self-heal” the network (example cited: automatic antenna tilt and optimization during “Winter Storm Fern”). ³

Capex implications in 2026 tied to 5G leadership (what can be concluded from the excerpts)

• The excerpts support that T-Mobile expects compute and infrastructure investment to follow its AI roadmap—specifically, moving from AI embedded in operations to compute-in-network as it approaches AI RAN and edge inferencing. ³
• They also support that continued 5G Advanced capacity/spectral efficiency improvements (including uplink-focused changes like uplink transmit switching, higher transmit power, and uplink MIMO) are part of T-Mobile’s physical AI readiness—implying ongoing network modernization effort consistent with leadership in 2026. ²
• But the excerpts do not provide a specific 2026 capex number or an explicit “AI capex” budget. ^{4 5}

In short: T-Mobile’s 2026 plan (as described here) is to integrate AI directly into its 5G Advanced network—starting with a network-native live translation beta, expanding toward AI in the core and inferencing at the edge, and partnering to enable physical AI connectivity—with capex/infra needs implied by compute-in-network and AI RAN progression, though not quantified in the provided excerpts. ^{1 3}