June 22, 2026—AT&T (NYSE:T) closed at $22.10 per share, with a market capitalization of approximately $153.6 billion and a trailing twelve-month P/E ratio of 7.17. For a leading U.S. telecom giant with annual revenue exceeding $120 billion, this valuation level reflects persistent market skepticism about its long-term growth prospects: the ceiling for traditional wireless and broadband businesses now seems within reach.
However, a rapidly emerging new paradigm may be reshaping this valuation framework. The explosive growth of AI applications is fundamentally transforming both the structure and scale of network traffic. As the physical backbone for data transmission, communications infrastructure now stands at the heart of this structural shift. By 2026, when AI inference traffic is projected to account for more than two-thirds of total AI traffic for the first time—and as AI traffic makes up roughly 30% of backbone network utilization—the network is no longer just a "pipeline." Instead, it becomes an indispensable strategic asset in the AI era.
Can AT&T leverage its fiber network, 5G coverage, and edge computing nodes to secure a strong position in this wave of AI-driven infrastructure upgrades? This article will analyze that question by examining the underlying logic of AI data traffic growth, AT&T’s latest financials, infrastructure progress, and strategic moves.
The AI Inference Era: A Fundamental Shift in Traffic Patterns
The first step to understanding AT&T’s potential growth story is to clarify how AI is impacting network traffic. This influence goes far beyond a simple "increase in volume"; it involves a threefold transformation in traffic scale, structure, and network performance requirements.
In terms of scale, AI-driven network traffic is entering a rapid acceleration phase. According to industry research, the number of global AI agents will reach between 50 and 100 billion by 2026 and could soar to 2–5 trillion by 2036. As a result, global bandwidth usage is set to skyrocket from about 100 exabytes per day in 2026 to roughly 8,100 exabytes per day in 2036—a compound annual growth rate of 51%. Mobile network data traffic grew 22% between Q1 2025 and Q1 2026, outpacing previous forecasts.
More significant than the sheer growth in volume is the profound change in traffic composition. AI traffic is shifting from "training-dominated" to "inference-dominated." In 2023, training accounted for about two-thirds of all AI traffic. By 2025, inference and training will be roughly equal, and by 2026, inference will surpass two-thirds for the first time. This marks a transition from centralized, periodic model training to distributed, continuous inference services—raising the bar for network coverage, low latency, and edge node density.
Ericsson’s June 2026 Mobile Market Report highlights another key trend: AI applications are driving much faster growth in uplink traffic than downlink. Of 55 global operators analyzed, 43 saw uplink traffic grow faster than downlink, and 17 experienced uplink growth at 1.5 times or more the rate of downlink. Traditional cellular networks are optimized for downlink, so this surge in uplink demand calls for a systemic overhaul of network infrastructure.
At the Fiber Connect 2026 conference, Cisco Senior Business Development Manager Robin Olds noted that AI traffic now accounts for about 30% of backbone network utilization—up from less than 1% just two years ago. Olds likened this shift to the "birth of the internet," arguing that AI is fundamentally changing traffic patterns and forcing service providers, data center operators, and cloud providers to rethink infrastructure design.
For communications infrastructure providers like AT&T, the takeaway is clear: the more pervasive AI becomes, the more essential high-quality network connectivity becomes. The question is, can AT&T translate this surging demand into revenue and profit growth?
AT&T’s Fundamentals: Stability and Limits of Traditional Business
Before evaluating the AI narrative, it’s important to review AT&T’s traditional business fundamentals.
In Q1 2026, AT&T delivered a stronger-than-expected performance. Total revenue reached $31.5 billion, up 2.9% year-over-year and beating the consensus estimate of $31.25 billion. Adjusted EPS came in at $0.57, above the expected $0.55. The Advanced Connectivity segment (covering domestic 5G and fiber services) generated $28.5 billion in revenue, up 4.7% year-over-year, with wireless service revenue at $16.9 billion, a 1.7% increase.
On the user side, AT&T added 294,000 postpaid phone subscribers in Q1, surpassing analyst expectations of 272,000. Net internet subscriber additions hit 584,000—split evenly between fiber and fixed wireless access—marking the company’s best-ever first quarter and the sixth consecutive quarter with net adds over 500,000. Fiber now reaches more than 37 million locations, with a target of 60 million by 2030.
AT&T’s convergence strategy is its most prominent growth engine. Forty-two percent of home broadband customers also use AT&T wireless services. Excluding the impact of the Lumen acquisition, the organic convergence rate is close to 45%, up more than 3 percentage points year-over-year. This approach effectively reduces churn and increases average revenue per account.
Yet the limitations of the traditional business model are clear. Wireless service revenue growth (1.7%) lags behind the Advanced Connectivity segment’s overall growth (4.7%) and is driven mainly by device sales rather than service revenue. For the full year, the company expects adjusted EPS of $2.25–$2.35, free cash flow of $18 billion or more, and capital expenditures between $23 billion and $24 billion. Even at the high end of guidance, the current share price implies a forward P/E of less than 10—indicating the market is not pricing in significant growth.
Competitive pressures are also mounting. In early June 2026, Oppenheimer downgraded AT&T from "Outperform" to "Perform," citing concerns about competition from SpaceX’s Starlink broadband. Analysts believe Starlink will continue to erode AT&T’s share in low-density broadband, edge cable networks, FWA alternatives, and enterprise backup links. While Starlink cannot fully replace traditional fiber and cellular networks in the short term, its presence increases uncertainty for AT&T’s broadband expansion in low-density areas.
Strategic Positioning in the AI Era: What Is AT&T Building?
If traditional business determines AT&T’s downside risk, its AI strategy defines its upside potential. Recent disclosures show AT&T is building AI-ready network infrastructure on multiple fronts.
Network Capacity Upgrades. AT&T has committed to building an "AI-ready network," expanding performance through ongoing investment. This includes boosting capacity to 1.6 Tbps on key metro and long-haul routes. The company has extended its 400G wavelength service to 40 U.S. metro areas and 130 interconnection nodes, enabling AI data to move at 400 Gbps between data centers, cloud facilities, AI clusters, and enterprise sites.
Deep Integration with Cloud Giants. In March 2026, AT&T announced a collaboration with AWS to launch the preview of "AWS Interconnect – last mile" (expected to be available in Q2 2026), extending AT&T’s 5G and fiber connectivity from enterprise locations directly into AWS environments. This partnership aims to create a secure, resilient "on-premises-to-cloud" architecture for AI workloads, supporting use cases like real-time analytics, machine learning, and agent-based AI. AT&T had previously partnered with Microsoft Azure for 5G SA core network operations; the AWS deal further deepens its ties with hyperscale cloud providers.
Differentiated Edge Node Advantage. Shawn Hakl, AT&T’s SVP of Product and Business, emphasized, "AI needs not just more compute, but a flatter network and faster connectivity." AT&T’s roughly 5,000 central offices and 65,000 cell sites give it a physical footprint unmatched by cloud providers. As AI inference shifts from centralized data centers to the edge, these nationwide edge nodes could become AT&T’s most valuable strategic asset.
AI-Driven Network Operations Optimization. AT&T has also launched Geo Modeler, an AI-powered simulation tool that can predict wireless network performance in near-real time, helping engineers identify potential weak spots in advance. This shows AT&T views AI not only as a demand driver but also as a tool to improve its own network operations.
Strategically, AT&T’s AI narrative is not just hype. The company operates one of the largest fiber networks in the U.S. (covering 37.5 million fiber locations), has broad 5G coverage, and a vast edge node network. As AI inference traffic continues to grow and low-latency connections become critical, the strategic value of these assets is being reassessed.
However, it’s important to note that a revaluation of infrastructure doesn’t automatically translate into high revenue growth for AT&T. As some Seeking Alpha analyses point out, AT&T is building networks for AI cloud companies, but the primary beneficiaries may not be AT&T’s own revenues. While increased network traffic boosts utilization, whether telecom operators can turn that into revenue growth depends on pricing power, value-added services, and business model innovation—all of which remain uncertain.
Risks and Constraints: Real-World Hurdles for the AI Narrative
AI-driven network demand growth gives AT&T a new growth story, but turning that narrative into sustainable shareholder value still requires overcoming several real-world hurdles.
Tension Between Capital Expenditures and Free Cash Flow. AT&T plans to invest $23–24 billion annually in capex from 2026 to 2028. The company expects free cash flow of $18 billion or more in 2026. With only about a $5 billion gap each year, there’s little room for error. If AI-driven network upgrade demand accelerates, capex pressure could rise further.
The Challenge of Monetizing the "Pipeline." Growth in network traffic doesn’t automatically mean revenue growth. The telecom industry has long faced the "scissors effect," where traffic grows faster than revenue. AT&T must demonstrate it can capture a premium through differentiated connectivity services—such as dedicated network slices for AI workloads or low-latency guarantees—rather than just providing commoditized "pipeline" services.
Multi-Faceted Competitive Pressures. Beyond Starlink’s threat in low-density broadband, AT&T faces competition from Verizon in 5G enterprise services and from Lumen, Zayo, and others in fiber interconnect markets. While collaborations like AWS Interconnect strengthen AT&T’s ecosystem, they also mean ceding some value to cloud providers.
Debt and Financial Flexibility. Market analysts warn that if AT&T doubles down on fiber and wireless infrastructure investment in response to AI demand, it may have to take on more debt. In today’s interest rate environment, higher debt costs would further constrain financial flexibility.
Conclusion
Should AT&T be reevaluated? The data suggests AI’s impact on network traffic is no longer theoretical—AI traffic’s share of backbone network utilization jumped from under 1% to about 30% in just two years; AI inference traffic will exceed two-thirds for the first time in 2026; and global mobile data traffic grew 22% in a single year. These figures all point to a clear trend: AI is shifting pressure from the compute layer to the network layer, placing communications infrastructure at the center of this structural transformation.
AT&T holds a significant strategic position in this trend—with one of the largest fiber networks in the U.S., extensive 5G coverage, and an edge node network of about 5,000 central offices and 65,000 cell sites. Deep partnerships with cloud giants like AWS and Microsoft Azure further embed its connectivity into the AI workload delivery chain. The 4.7% year-over-year revenue growth in the Advanced Connectivity segment in Q1 2026 also validates, to some extent, the effectiveness of its convergence strategy.
But challenges remain. The tight balance between capex and free cash flow, the industry-wide challenge of monetizing the "pipeline," and threats from new competitors like Starlink are all key variables determining whether the AI narrative can translate into sustainable growth.
AT&T’s AI story is not about a "short-term boom," but rather a "structural revaluation." It doesn’t hinge on a single product breakthrough or an exceptional quarter, but on whether AI-driven network demand can continue to grow—and whether AT&T can evolve from a "communications pipeline" to an "AI infrastructure service provider." Proving this thesis may take three to five years, but for investors focused on long-term infrastructure value, this could be the underappreciated factor in today’s market pricing.
