Expert Interview
Examining how Amazon’s AI-driven terrestrial fiber capex to support AWS expansion pressures retail margin recovery and capital allocation
Ticker(s): AMZN, EQIX, DLRInstitution - Swift Data Development
- seasoned data center and network infrastructure leader with deep, hands-on experience deploying large-scale terrestrial fiber networks that underpin hyperscale cloud and AI infrastructure across North and South America.
- Former Manager of Terrestrial Network Deployment at AWS, he led long-haul and metro fiber programs, combining network architecture, financial underwriting, and execution discipline to meet rapidly growing data center and machine learning capacity demands.
- Managing Partner at Swift Data Development, he applies this end-to-end expertise spanning site selection, grid and network interconnectivity, and economic development, to shaping the next generation of AI-ready data center campuses.
From your experience at AWS, how did AI-driven workloads change the required topology and density of terrestrial fiber compared with traditional cloud growth?
Added By: james_adminWhere does terrestrial fiber capex sit in the overall cost stack of an AI-oriented data center build, and how has that share evolved over the last few years?
Added By: james_adminHow should investors think about the return profile of owned vs. leased fiber when supporting hyperscale AI expansion?
Added By: james_adminWhat are the biggest technical or economic bottlenecks you see when scaling long-haul and metro fiber to support GPU-heavy clusters?
Added By: james_adminIn your view, does network infrastructure become a limiting factor for AI scaling before power, real estate, or compute, or after?
Added By: james_adminHow did capital allocation decisions around network infrastructure compete internally with other priorities, such as retail margin recovery or data center build-outs?
Added By: james_adminWhat metrics or KPIs mattered most when justifying terrestrial fiber investments inside AWS; latency, utilization, cost per bit, or something else?
Added By: james_adminAs AI models scale and training clusters grow more distributed, do you expect fiber requirements to grow linearly, exponentially, or plateau?
Added By: james_adminFrom your current role in site selection and development, how early does network availability now influence where AI data centers get built?
Added By: james_adminLooking forward, what would cause a hyperscaler like Amazon to slow or rethink terrestrial network expansion; technology shifts, utilization risk, or capital return pressure?
Added By: james_adminAre You Interested In These Questions?
Slingshot Insights Explained
Expert research benefits investors by giving them timely access to unbiased real world perspectives on highly specialized topics. Slingshot Insights' crowdfunded model makes this access available at a fraction of the cost of other expert networks.
Reason
*Slingshot Insights provides access to information, not investment advice. We work to support you and facilitate access to experts; however we are not responsible for monitoring calls for the disclosure of MNPI. You should obtain financial, legal and tax advice from your qualified and licensed advisers before deciding to invest in any security.
