Skip to content
Creator Hubs›Applied Digital

Applied Digital

Applied Digital logo

News, updates, and expert insights from Applied Digital.

Discover the epicenter of AI innovation at Applied Digital, where our purpose-built data centers are engineered for peak performance. Revolutionize accelerated computing with our state-of-the-art design and power-centric facilities. Follow this channel for the latest from Applied Digital: product news, expert perspectives, and updates from the team.

137 episodesVisit website ↗
Channel Brief·Applied Digital · 137 episodes
Updated Apr 14, 2026

AI data centers demand a complete rethinking of infrastructure

Applied Digital's channel argues that scaling AI compute requires not just new facilities, but entirely new approaches to power, cooling, operations, and community integration. Evidence comes from on-the-ground updates and technical deep dives into 100+ MW facility design and commissioning.

Applied Digital argues that traditional data center expertise no longer applies to AI workloads, forcing a shift from retrofitting legacy facilities to building purpose-built "AI factories" from the ground up. The channel proves this through detailed construction updates from its Ellendale, North Dakota campus (Polaris Forge 1) and technical episodes explaining why conventional facilities cannot handle the density, thermal demands, and power volatility of accelerated computing.

Drawn from Applied Digital’s Data Center Design for a 100… and 1 more →

“Legacy data centers, built for traditional co-location, cannot handle the density, thermal demands, or power dynamics of accelerated computing.”

Episode 17: Applied Digital's Data Center Design for a 100 MW AI Factory Built from the Ground Up

By the numbers

600kW

NVIDIA's target rack density by end of 2027

$200B

Global AI infrastructure spending forecast by 2028 (IDC)

2.5B

Daily prompts to ChatGPT (2.5 years after launch)

100 MW

Critical IT load of Building 1, Polaris Forge 1

What the channel argues

DataIndustry shifted from 10kW racks to 120kW, with NVIDIA targeting 600kW by end of 2027.→
InsightAI data centers generate millions of data points per second, making real-time visibility essential for operations.→
DataApplied Digital built Polaris Forge 1 (Building 1) with 100 MW IT load in under a year using prefabrication.→
InsightAI workloads trigger rapid, synchronized power spikes and drops of tens of megawatts in seconds.→
DataIDC forecasts global AI infrastructure spending will exceed $200 billion by 2028.→
InsightCommissioning 100 MW+ facilities involves months-long systems validation, not just equipment checks.→

What you'll learn

•Why traditional data center design principles fail for AI: rack density has increased 60x in just a few years, forcing complete rethinking of electrical, cooling, and control systems.
•How remote locations like Ellendale, North Dakota enable both scalability and efficiency: natural climate reduces cooling load and land availability supports modular expansion to 400+ MW.
•What commissioning a 100 MW facility actually means: months of integrated systems testing of power, cooling, controls, and teams under real-world conditions, not just component validation.
•How data center operators are managing volatile power loads: real-time telemetry at scale (millions of data points per second) is now essential, not optional.
•Why rural communities are becoming strategic hubs: Applied Digital is integrating workforce development, housing, and civic projects (like pickleball courts) as core to facility success.

What to do about it

→Treat data center commissioning as a distinct phase requiring 3-6 months of integrated systems validation, vendor alignment, and on-site collaboration before going live on 100+ MW facilities.
→Design for 10x power density growth: plan electrical, cooling, and control infrastructure for 600kW+ racks now, even if deploying at lower densities today, to avoid stranded assets.
→Embed community integration into infrastructure strategy: workforce partnerships, local procurement, and civic amenities are no longer nice-to-have but operational necessities for rural deployment at scale.

Who and what shows up

Philbert Shih

Founder and host, Structure Research

Hosted the Architects of Acceleration series, conducting deep technical and strategic interviews on data center design, execution, and community integration.

Nick Phillips

EVP of External Affairs, Applied Digital

Spoke on community integration strategy, workforce, housing, and local partnerships as core to rural data center deployment success.

McGough Construction

Construction partner

Built Building 1 of Polaris Forge 1 to substantial completion with 100 MW IT load on schedule using off-site prefabrication.

NVIDIA

Technology partner and benchmark setter

2025 GTC roadmap targets 600kW racks by end of 2027, driving the industry shift in power density that forces complete data center redesign.

Applied Digital

AI infrastructure provider

Transitioned from Bitcoin mining to purpose-built AI factories; built Polaris Forge 1 in Ellendale as case study in greenfield AI infrastructure and rural community integration.

Questions this channel answers

Q

How do you keep a 100 MW AI data center online 24/7 when power demand swings tens of megawatts in seconds?

AI data centers require sophisticated power and cooling systems validated over months, real-time telemetry visibility across millions of data points per second, and infrastructure designed specifically for volatile, synchronized loads unlike traditional cloud workloads.

Power, Pressure, and Precision: What It Takes to Keep AI… →
Q

Why can't traditional data centers handle AI workloads?

Legacy facilities were built for sub-10kW racks; AI workloads now demand 120kW racks with NVIDIA targeting 600kW by 2027. The density, thermal demands, and power dynamics require ground-up redesign of electrical, cooling, and control systems.

Applied Digital’s Data Center Design for a 100 MW AI Fac… →
Q

What does it take to commission a 100 MW facility?

Months-long integrated systems validation of power, cooling, controls, and teams working together under real-world conditions, not just equipment checks. Success depends on early planning, vendor alignment, and in-person collaboration.

Power, Cooling, and Risk: What It Takes to Bring a 100MW… →
Q

Why build AI data centers in rural areas like North Dakota?

Rural locations offer scalable land, access to natural cooling (reducing energy and water consumption), available power generation capacity, and the opportunity to revitalize communities through workforce development and local partnerships.

Execution at Scale: How Applied Digital Is Powering AI I… →
Q

How do you operate at the scale of millions of data points per second?

Visibility infrastructure that transforms massive telemetry streams into clear operational decisions is now essential, not optional. Real-time data on power, cooling, and system health drives uptime, efficiency, and risk mitigation.

Visibility at Scale: How Data, Telemetry, and IT Archite… →
Topics:Liquid cooling systems100+ MW AI data centersPower density and managementRural infrastructure deploymentData center commissioning
Themes:Purpose-built beats retrofitted for AIOperations complexity shifted from physical to digitalInfrastructure and community strategy are inseparable

Industry context

AI workloads now represent over 40 percent of new colocation lease demand, driving rapid adoption of liquid cooling as a baseline expectation for operators competing for AI infrastructure contracts.

Want a show like this for your brand?

MarketScale produces and distributes branded shows like Applied Digital for B2B companies.

Book a Demo →