In this episode of The Bridgecast, host Scott Kinka welcomes back Brian Eichman, Vice President of Business Development and Solution Architecture at CoreSite, to explore the revolutionary concept of "AI inference zones." From CoreSite's billion-dollar milestone to the shift from AI training to monetization, Brian breaks down how proximity, ecosystem partnerships, and hybrid infrastructure are reshaping the future of enterprise AI deployment.

Brian introduces the critical concept of inference zones—strategic locations where AI models transition from training to real-world application and monetization. Using his signature analogy-driven approach, he explains how AI is moving from the "college phase" of learning to the "big city phase" of making money, requiring new infrastructure approaches that prioritize proximity to users and low-latency connectivity.

Want to learn how to build an infrastructure that can handle the next wave of AI? Tune in now.

Brian Eichman is the Vice President of Business Development and Solution Architecture at CoreSite, one of the leading colocation providers in the U.S. With over 13 years at CoreSite, Brian has transitioned from engineering to product development and now focuses on corporate strategy, ensuring CoreSite remains competitive in the rapidly evolving data center landscape. Under his leadership, CoreSite has achieved a major milestone, growing from $150 million in annual revenue when he started in 2012 to breaking the billion-dollar revenue mark. Brian is also a dedicated community volunteer with organizations like Junior Achievement and serves as a soccer coach, bringing the same collaborative and mentoring approach to his professional leadership.

Brian introduces the groundbreaking concept of inference zones—strategic geographic locations where AI models transition from training to real-world application and revenue generation. These zones represent key markets with high concentrations of enterprises, consumers, and robust internet infrastructure where cloud providers are establishing clusters of infrastructure. Unlike the training phase that happens in remote locations with expensive GPUs, inference zones are positioned in major Tier 1 markets like Northern California, Chicago, and Northern Virginia, where models can be served with high-speed, low-latency connectivity to end users and monetized at scale.

Using his signature analogy approach, Brian explains AI development as a journey from college to career. The training phase is like attending university—models are exposed to vast amounts of data, different perspectives, and formal education in isolated environments where organizations pay significant costs for GPUs, power, and space. The inference phase represents graduation and moving to big cities where that knowledge gets applied to make real-time decisions and generate revenue. This shift from learning to earning requires different infrastructure approaches, emphasizing proximity to users and integration with existing business systems rather than raw computational power.

Brian emphasizes that successful AI deployment requires abandoning traditional siloed decision-making in favor of comprehensive, interconnected infrastructure strategies. IT leaders can no longer make isolated decisions about data centers, cloud providers, or network connectivity without considering how these choices impact AI capabilities. The key principle is proximity—ensuring that every infrastructure decision considers low-latency access to AI models, cloud on-ramps, and ecosystem partnerships. This holistic approach enables organizations to remain flexible and choose the best fit for each application rather than being locked into suboptimal solutions based on price alone.

Brian offers a contrarian perspective on the current AI hype, suggesting that while the technology is transformative, the market may be experiencing overbuilding and bubble-like conditions similar to previous technology waves. Drawing parallels to the internet boom and cloud adoption cycles, he predicts some rationalization ahead where the initial hype will settle into practical, sustainable implementations. This perspective doesn't diminish AI's importance but suggests that organizations should prepare for a more measured adoption phase where proven use cases and ROI take precedence over experimental deployments.

Episode Resources:

The Bridgecast is handcrafted by our friends over at: fame.so