@misc{9229887,
  abstract     = {{Mobile connectivity is set to change in the coming years due to AI generated data from agents and robotics. How this will develop remains uncertain since this type of data generation is not only a question of connectivity but also compute location and use cases. This creates a need for forward looking scenario analysis for ecosystem players so they can position themselves to win in a new paradigm.

The purpose of this research was to analyze how the introduction of artificial intelligence traffic would impact the information and communication technologies industry by 2035. It aimed to provide a prognosis of network traffic, scenarios that might play out and how the ecosystem might change in each scenario. The goal was not to predict a single correct future but to map the space of plausible ones and identify what conditions led to each scenario.

The research employed a two-round Delphi study as its primary method, gathering structured responses from 23 domain experts across academia, technology vendors, regulators, and industry associations. Industry reports were used as secondary data to contextualize and validate the expert findings.

Four scenarios looking toward 2035 were highlighted as possible futures, the Delphi panel was prompted on each scenario's plausibility. This resulted in one scenario with a strong positive consensus of 93 percent, two with 70 percent and one with no consensus splitting the panel at about 50 percent. Additionally, for each scenario: winners, losers, consequences and things that need to happen for the scenario to materialize were identified.

Three bifurcation points were found that have major implications for communication service providers (CSPs) and telecom equipment vendors. First, regulation is critical for CSPs. Market forces alone are not considered enough for them to gain traction. Second, the decoupling of infrastructure from economic value can be detrimental. If computing shifts to the edge but the economic value is still captured by the orchestrators, CSPs end up losing out, even though they are the ones supplying the compute. Third, geopolitical fragmentation makes data sovereignty a priority. This positions CSPs to displace global hyperscalers as primary compute providers.}},
  author       = {{Westring, Robert and Chamoun, Alex}},
  language     = {{eng}},
  note         = {{Student Paper}},
  title        = {{Machines Don’t Watch Netflix - A Delphi Study-based Scenario Analysis of the Network for AI Ecosystem in 2035}},
  year         = {{2026}},
}