Synthetic intelligence is essentially reshaping the panorama of world connectivity, performing as a catalyst for a brand new, subtle “connectivity material.” Cisco is constructing the safe, American-led infrastructure for the AI period. On Thursday July 30, Cisco’s Bob Everson, Chief Architect for Cisco’s Service Supplier Mobility workforce, testified earlier than the Senate Commerce Telecom Subcommittee, highlighting in his written testimony and opening assertion how AI is redefining networks, how the U.S. ought to lead by means of balanced spectrum and American-led tech stack, and the way we will outline American financial energy and nationwide safety for many years to return:
- A Shift in Community Visitors Patterns: AI is essentially altering the conduct of information on our networks. Cisco tasks that AI-driven campus and department site visitors will surge by 209% over the following three years. In contrast to conventional purposes, that are largely optimized for downstream content material consumption, AI workloads are closely uplink-intensive and require extra complicated, two-way communication.
- The Unprecedented Calls for of AI Brokers: The rise of autonomous AI brokers is inserting vital pressure on legacy infrastructure. Cisco’s inside testing revealed that AI brokers generate 450% extra site visitors per process than human customers. To handle this load effectively, Everson emphasised that edge computing will play an more and more important function in lowering infrastructure prices and securing delicate knowledge.
- AI-Native Operations for Machine-Velocity Networks: Managing and securing trendy, distributed networks requires leveraging agentic operations—utilizing AI to observe and handle infrastructure in actual time. By figuring out faults, predicting threats, and automating responses, this strategy permits networks to self-heal and function at machine velocity.
Opening Assertion: Bob Everson
Chairman Fischer, Rating Member Luján, Chairman Cruz, Rating Member Cantwell, and distinguished Senators, thanks for the invitation to testify.
My identify is Bob Everson. For 30 years at Cisco, I’ve labored on the programs powering our nation’s essential networks, and I’ve by no means seen a shift this consequential.
Synthetic intelligence isn’t just one other software. It’s driving a brand new connectivity material that can assist outline U.S. financial energy and nationwide safety for many years. In the present day, I’ll concentrate on how AI is remodeling our networks and the way Cisco is utilizing AI to enhance them.
First, AI is altering not solely the amount of community site visitors, however the conduct. Cisco measured a fourfold enhance in AI inference site visitors over eight months. Networks have historically been optimized for content material flowing downstream. AI is much extra two-way and uplink-intensive: prompts, context, sensor knowledge, and agent exercise all journey again towards AI fashions, and the ensuing connections are energetic longer than typical internet transactions.
AI brokers amplify these results by working at software program velocity. In our testing, an agent generated 450 % extra site visitors than an individual performing the identical process, and roughly 70 % of that extra site visitors was inference.
Assembly these calls for requires extra balanced community designs—better upstream capability, resilient paths with end-to-end visibility, stronger safety, and high quality of service that may tailor the AI site visitors. As fashions turn out to be quicker, the areas the place computing happens—from centralized clouds to the sting—will more and more have an effect on efficiency, price, and knowledge governance.
A community this quick and distributed can’t be operated solely at human tempo. By way of AgenticOps, Cisco makes use of AI to help self-healing operations—figuring out faults, rerouting site visitors, and recovering at machine velocity. Our deep community fashions can distinguish respectable AI site visitors from anomalies, assist predict and isolate threats, establish vulnerabilities, and information operators in deploying patches.
AI-native cell networks can use spectrum and infrastructure extra effectively, adapt in actual time, and allow new providers. For service suppliers, this may additionally flip the community right into a platform for development, opening new income streams past connectivity by means of providers like inference optimization, edge AI, trusted AI providers, and industry-specific options. One noteworthy instance is built-in sensing and communications, which permits AI to mix radio indicators with imaginative and prescient and different inputs so wi-fi networks can join gadgets whereas additionally detecting objects and motion—supporting purposes akin to safer robotics, extra environment friendly buildings, enhanced public security, and healthcare.
Collectively, these capabilities outline a brand new technology of clever, safe, and distributed networks that carry AI, use AI to enhance operations, and allow AI-native purposes and providers—strengthening financial competitiveness and nationwide safety.
I commend the NTIA for this week’s Name to Motion for 6G Management and Safety, which advances the worldwide cooperation wanted to make sure trusted companions form the 6G future.
To maintain U.S. management, I supply three coverage priorities.
First, speed up the U.S. AI-native stack. Cisco is investing throughout a number of dimensions of AI-native networking—bringing new capabilities to 5G-Superior in the present day whereas constructing the muse for 6G. One instance of this dedication is AI-WIN—a collaboration amongst Cisco, NVIDIA, MITRE, ORAN Improvement Firm, Booz Allen, and T-Cell—which brings AI, compute, and wi-fi collectively to create a safe, American-led path from 5G-Superior to AI-native 6G. I’d encourage Congress to lean on American strengths in these areas.
Second, modernize allowing and infrastructure. As computing turns into extra distributed, allowing should allow speedy and accountable deployment. As this Committee considers the way forward for the Common Service Fund, it ought to account for the evolving prices of AI-ready networks so rural and concrete communities can share in the advantages.
Third, preserve a balanced spectrum coverage. The 800 megahertz of licensed spectrum not too long ago made out there by Congress is important to high-capacity, high-uplink connectivity. The FCC’s authorization of the 6 gigahertz band for unlicensed use is equally vital to assembly enterprise demand. A reliable pipeline of each is foundational to American management, and I thanks in your efforts to rebuild it.
Cisco stands able to work with this Committee to make sure the USA stays the world’s premier vacation spot for AI and wi-fi innovation.
Thanks for the chance to testify. I look ahead to your questions.
