Fiber for Breakfast Week 5: Monopoly in the Machine? AI, Chips & the Battle for Network Innovation

Fiber for Breakfast Week 5: Monopoly in the Machine? AI, Chips & the Battle for Network Innovation 

The conversation around AI has shifted. What once centered on models and use cases is now increasingly about infrastructure – chips, data centers, fiber, and who controls access to them. That question framed this week’s Fiber for Breakfast, where Gary was joined by Asad Ramzanali, Director of Artificial Intelligence & Technology Policy at the Vanderbilt Policy Accelerator, to unpack what he calls the growing risk of a “monopoly in the machine” – and what that means for networks, competition, and innovation.  

Ramzanali comes to the issue from the inside. Until recently, he served as Chief of Staff and Deputy Director for Strategy at the White House Office of Science and Technology Policy (OSTP), where he helped coordinate science and technology priorities across the federal government. When Chat GPT burst into the public consciousness in late 2022, OSTP wasn’t started from scratch. The team had already been working on AI policy – from the National AI Research Resource to the Blueprint for an AI Bill of Rights – but the moment forced a shift from niche policy work to a whole-of-government priority. 

What concerns Ramzanali now isn’t just the speed of AI’s growth, but the concentration of power beneath it. A small number of companies dominate key layers of the AI stack – from advanced chips to cloud infrastructure to foundation models – and in some cases, control multiple layers at once. That kind of vertical integration, he warned, risks slowing competition where it matters most. 

“The innovation we care about is what sits on top,” he explained, whether that’s medical research, productivity tools, or weather forecasting that can save lives. When access to chips, compute, or models tightens, the consequences ripple far beyond Silicon Valley.  

For broadband providers, those ripples are already visible. AI-driven growth is fueling massive demand for fiber, especially around hyperscale data centers. But Ramzanali cautioned against overconcentration. In the first half of 2025, he noted, 90 percent of U.S. GDP growth was tied to AI and related technology investment – an economic imbalance that should give infrastructure providers pause. 

The conversation also touched on the growing backlash to data centers themselves. Once welcomed as economic development, many communities are now pushing back over energy use, water consumption, noise, and the reality that they bring few permanent jobs. In response, some hyperscalers are starting to negotiate directly with communities – offering broadband buildouts or infrastructure improvements – a shift that could give local networks more leverage than expected. 

On policy, Ramzanali pushed back on the idea that reasonable AI regulation will stifle innovation. The greater risk, he argued, is public distrust if abuses go unchecked – from deepfakes to market conflicts where companies compete directly with their own customers. The tools to address this don’t require sweeping new regimes, but familiar principles: transparency, non-discrimination, and targeted oversight that preserves openness.  

One theme kept resurfacing throughout the discussion: AI isn’t arriving alone. It’s colliding with advances in networking, edge computing, biology, and, eventually, quantum technologies. The disruption may feel sudden, but as Ramzanali reminded the audience, most overnight revolutions are decades in the making.  

For telecom and broadband leaders, the message is clear. AI is already reshaping demand, investment, and risk. Staying ahead won’t just mean chasing hyperscalers, it will mean understanding where concentration creates fragility, where infrastructure can unlock shared value, and where openness still matters. 

The genie is out of the bottle. What happens next depends on who controls and machine – and who keeps the ecosystem open.   

Click here to watch the full interview.