THE APEX TIMES
Jeff McElfresh, AT&T’s COO, argued that the internet was optimized for video streaming, but that AI agents are now driving always-on network demand.
The internet was built for Netflix, AT&T’s COO said, using a familiar analogy for how modern broadband networks evolved around video streaming. Jeff McElfresh, speaking in an interview highlighted by Yahoo Finance, said the “actual direction” of bandwidth is now changing as software “agents” increasingly connect to networks throughout the day and night.
McElfresh framed the shift as a move away from primarily human-driven traffic that peaks around viewing times and toward machine-driven traffic that is more continuous. “We’re seeing agents tap the network in a 24/7,” he said, describing a world where programs that act on instructions from users or systems generate ongoing data demand.
The comment matters for companies that sit at the center of delivery. Netflix has long been associated with the mass-market streaming model, while telecom carriers like AT&T operate the underlying networks that must carry that traffic. In the Netflix-era, streaming workloads and content delivery strategies were engineered around predictable bursts and the user experience requirements of video.
McElfresh’s remarks suggest that AI use cases may stress the network in different ways. Rather than long-lived, interactive sessions anchored to screens, AI agents can be continuously querying, retrieving, and coordinating information across services. Even without specific technical details in the reported interview, the broad claim is that the pattern of demand is becoming more constant and more distributed.
For Netflix, the AI-agent narrative is less about whether streaming remains important and more about how the surrounding ecosystem changes. AI capabilities can increase the amount of data exchanged around entertainment and streaming, including recommendation logic, personalization, metadata retrieval, rights and billing workflows, and cross-service orchestration. Netflix did not provide new figures in the Yahoo Finance segment referenced in the report.
Telecom operators, meanwhile, have to manage capacity planning, routing, and reliability for traffic that may arrive in steadier streams. A key difference is that AI-agent traffic may reflect many simultaneous background connections and frequent small transfers rather than a smaller number of large video streams. That can affect how networks handle congestion and how quickly they can respond to changing load.
Still, the reported exchange leaves important questions unanswered. The interview does not quantify how much network demand AI agents will add, which applications are leading the shift, or how much of the change should be attributed to AI compared with other sources such as gaming, cloud applications, or additional video formats. It also does not specify whether carrier traffic is measured by bytes, sessions, latency sensitivity, or other performance indicators.
What to watch next is whether network operators and large content platforms begin to publish more explicit indicators of machine-driven traffic growth, along with clearer benchmarks for capacity and performance. If “agents tapping the network 24/7” becomes a common operational planning assumption, the next round of carrier guidance, network investment disclosures, and product roadmaps may increasingly reflect the economics of always-on demand rather than primarily peak-time video consumption.
Bonta says his office is weighing whether the law and facts justify action in an ongoing scandal involving Jensen Huang, as the regulator examines potential wrongdoing related to the artificial-intelligence chipmaker’s business.
A new market piece argues that Nvidia remains a dominant AI play, but that some investors may prefer other AI-linked stocks framed as having steadier risk profiles and more modest valuations.
A broad drop in US stock indexes reflected a cautious tone into the next NVIDIA earnings event, according to Yahoo Finance, as traders weighed near-term market direction against results from the chipmaker’s latest quarter.
At SC26 in Chicago, ASUS presented an AI Factory blueprint aimed at helping enterprises assemble AI infrastructure that can be tuned to specific workloads, performance goals, and deployment constraints, while pairing the concept with both Intel and AMD-based platforms.
Ahead of Nvidia’s earnings and guidance update, market participants are positioning around expectations for revenue guidance on the order of $130 billion, with options markets implying the stock could move nearly 7% in either direction.
The short seller Muddy Waters has disclosed a position against Nvidia, adding to investor scrutiny around the chipmaker’s leadership and governance narratives, according to a report published by Yahoo Finance.
Microsoft used its Ignite conference to push an approach to generative AI built around using smaller, more task-specific models, as businesses evaluate how AI consumption translates into real-world spending.
Microsoft’s Office installed base is already among the software industry’s largest, but an RBC analyst argues the next growth step is pricing power, as competitors push free, Office-like tools into the market.
The software giant says a new initiative will help organizations track what happens around AI models and agents, addressing a common enterprise problem: many systems cannot fully explain the path from a user request to a final outcome.
Bank of New York and Microsoft announced that software agents negotiated and executed a Treasury market transaction through electronic trading workflows, positioning the demo as an early milestone for machine-to-machine finance.
Comments
Community guidelines. Comments may be removed for violations.
Loading comments...