AI’s Biggest Bottleneck Isn’t GPUs — It’s Power | Sergii Gerasymovych

AI’s next bottleneck may not be GPUs. It may be the power they need for computing.

In this episode of An Hour of Innovation podcast, Vit Lyoshin speaks with Sergii Gerasymovych, founder and CEO of EZ Blockchain and an AI data center infrastructure operator, about the physical systems behind the AI boom. Sergii explains why his company is pivoting from Bitcoin mining infrastructure toward AI data centers, why power and land may be more constrained than chips, and why AI builders often underestimate how slowly the physical world moves compared with software.

This conversation looks beyond models and GPUs to the infrastructure layer that determines how fast AI products can actually scale.

They discussed:

* Why crypto mining infrastructure prepared Sergii for the AI data center boom

* Why power and land may be harder to scale than GPUs

* How utilities, transformers, substations, and power studies slow down deployment

* What an AI data center actually looks like in simple terms

* How modular data centers work and why startups may use them

* Why Sergii thinks future AI data centers may become smaller and more distributed

* What Sergii believes people misunderstand about data centers, power, water, and local communities

* Why demand response can make some large power users useful for the grid

* Why space data centers and blockchain-style distributed compute may be overhyped

* What AI founders should understand before assuming compute will be available quickly

Sergii Gerasymovych is an infrastructure operator whose company, EZ Blockchain, began by building and operating data center infrastructure for Bitcoin mining and is now applying those lessons to AI compute. His perspective is grounded in the physical side of technology: power contracts, land, modular deployment, grid constraints, cooling, and the real timelines required to bring compute online. For builders, founders, product leaders, and engineers, this episode is a practical reminder that AI may feel like software, but it depends on hardware, energy, and infrastructure that cannot always move at startup speed.

Timestamps

00:00 Introduction

01:41 The Shift from Crypto to AI Infrastructure

03:15 Lessons from Crypto Mining for AI Data Centers

06:40 Power and Land: The New Bottlenecks

10:01 The Role of Clean Energy in AI Data Centers

12:52 Chips: Supply and Demand Dynamics

15:25 Understanding AI Data Center Infrastructure

17:55 Public Perception and Controversies

23:30 Modular Data Centers: the Future of AI Infrastructure

27:23 The Evolution of Data Centers Over the Next Decade

31:14 The Feasibility of Space Data Centers

35:46 The Future of AI Compute Power

41:45 Misconceptions About AI Infrastructure

51:33 Advice from Sergii to Builders

54:09 Innovation Q&A

Connect with Sergii

* Website: https://ezblockchain.net/

* LinkedIn: https://www.linkedin.com/in/sergii-gerasymovych/

* X: https://x.com/SergiiGera

* Other: https://gerasymovych.com/

Support the Podcast

If you enjoy the podcast, you can support it by exploring the tools below.

These are affiliate links, meaning the show earns a small commission at no extra cost to you.

MeetGeek: Record, transcribe, summarize, and share insights from every meeting.

https://get.meetgeek.ai/yjteozr4m6ln

Amplemarket: Step into the future of sales: Human + AI. Empower reps, uncover opportunities, and grow revenue.

https://grow.amplemarket.com/04e068165684-affiliate

Databox: Turn business performance data into clear answers your team can understand, explain, and act on – instantly.

https://join.databox.com/vhw0wbr9qg9o

For inquiries about sponsoring An Hour of Innovation, email [email protected]

Connect with Vit

* Substack: https://anhourofinnovation.substack.com/

* LinkedIn: https://www.linkedin.com/in/vit-lyoshin/

* X: https://x.com/vitlyoshin

* Website: https://vitlyoshin.com/