Stuff You Should Know: Data Centers – Can't Live With 'Em, Can't Live Without 'Em
Podcast: Stuff You Should Know
Hosts: Josh and Chuck
Date: January 15, 2026
Episode Theme:
This episode explores the past, present, and future of data centers—those massive, often mysterious buildings packed with servers that make modern digital life possible. Josh and Chuck unpack what data centers are, how they evolved, their critical (but often invisible) role in the global economy, and the mushrooming impact—particularly as artificial intelligence (AI) sends data center demands into overdrive. They cover technical history, cultural moments, and the environmental, financial, and social risks of this rapid expansion.
1. What Are Data Centers?
- Everyday Definition: A data center is any place where you can store and access data, from a server closet to a massive hyperscale facility.
- Memorable Quote:
- Chuck: “Technically the PC is a data center. Anywhere you can store and access data, that’s technically a data center.” (03:01)
- Key Point:
- Data centers have evolved in step with computing needs—from closets in office buildings (and podcast studios) to hyperscale buildings powering the global Internet.
2. Timeline: The Evolution of Data Centers
- Early Beginnings:
- Mainframes & Military:
- The first programmable electronic digital computer was Colossus at Bletchley Park, WWII-era UK, used to break enemy codes. The ENIAC in the US handled missile trajectories.
- Quote: Josh: “You walked up to them, asked what’s the trajectory of this missile, and ENIAC would go beep, bop, boop, boop, and then tell you.” (08:06)
- Mainframes & Military:
- Mainframes for Business:
- 1950s: Companies like British tea chain Lyons use mainframes (the LEO) for payroll and inventory, and even lease processing time to the military.
- 1960s: IBM becomes the main player; leasing a mainframe could cost $16,000/month (1952 money).
- Quote: Chuck: “A $10 billion...data center. 400 jobs. Because these things are so efficient, everything is just so advanced...” (44:27)
- PC Revolution and the Internet:
- Having a PC = your own mini data center.
- The Internet’s boom in the ’90s and 2000s catalyzed huge leaps: more companies needed bigger data centers to keep up.
- Cloud computing in the 2000s changes the game: no need for everyone to own their own server.
- Quote: Chuck: “People thought it was just floating up in the ether somewhere. It’s still being stored on stuff. It’s just not being stored locally.” (17:49)
- The Hyperscale Era:
- Examples: Google's data center in Oregon (>1.3 million sq ft), China Telecom in Inner Mongolia (10.7 million sq ft, or 250 acres).
- Covid and remote work gave another big push; AI is now “blowing up” demand to new levels.
- Quote: Chuck: “Covid actually gave it another bump...Remote working finally, finally established itself as like, no, we're doing this...” (22:04)
3. Data Explosion: How Much Do We Use?
- In 2010, humanity consumed 2 zettabytes of data; by 2024, that number had grown to 150 zettabytes—a 75x increase in 14 years.
- Notable Definition: 1 zettabyte = 1 trillion gigabytes.
- Quote: Josh: “In 2024...we used 150 zettabytes of data. That’s anything from uploading a TikTok to buying a song on iTunes.” (03:23; 04:18)
4. The Era of AI and Hyperscale
-
AI’s Insatiable Hunger:
- The release of ChatGPT in 2022 unleashed a race—AI computation requires massive data centers.
- Standard CPUs aren’t fast enough; data centers now string together tens or hundreds of thousands of GPUs (graphics processing units, originally for video games) for “parallel processing.”
- Example: ChatGPT trained on 20,000 Nvidia GPUs.
- Quote:
- Josh: “You need hundreds of thousands of these things strung together. Instead of a CPU running a couple of servers...all of them are strung together to form one giant supercomputer that the AI operates on.” (24:56)
-
Corporate Arms Race:
- Microsoft: $88B on data centers in 2025 alone
- Amazon: $150B pledged over 15 years
- Google & Meta: $750B (equipment) over the next two years
- Morgan Stanley projects $3 trillion will be spent (2025–2030).
- Quote: Josh (joking): “Stanley Morgan says...what did I say? Stanley Morgan? Yeah, I think we should leave that in there.” (27:54)
5. Speculative Frenzy and Warnings
- Are We in an AI Bubble?
- Critics note that only 5% of business AI programs currently secure a return on investment.
- The financing is built on “private credit”—unregulated, shadowy lending with potential to tip the global economy, reminiscent of 2008’s debt meltdown.
- Quote: Josh: “The IMF...are flashing the warning signs saying like, this could...make a trillion dollars by 2028 or it could like wreck the global economy.” (34:09)
6. Environmental and Social Impact
- Energy and Water Suckers:
- Data centers use 1–1.5% of the world’s electricity (as much as some small countries).
- In Ireland, they use 20% of national power; in Northern Virginia, data centers use as much power as 60% of VA households.
- Barclays: By 2030, US data centers could use 13% of national electricity.
- Evaporative cooling means massive water use: Meta and Microsoft’s Phoenix data centers, 7M gallons/day.
- Quote:
- Josh: “Some of these AI data plants use the same amount of electricity as a town of 50,000.” (37:14)
- Chuck: “Meta and Microsoft use 7 million gallons of water every single day for their data centers.” (42:29)
- Local Disruption:
- Spikes in electricity and water bills for locals (e.g., +267% electricity in Data Center Alley, VA, since 2020).
- “Local towns are running out of water. They go to turn on their water and water doesn’t come out because of this.” (42:48)
- Few Local Benefits:
- Huge investments often generate only a few hundred jobs, and money flows back to the tech giants, not the hosting communities.
- Quote: Chuck: “A $10 billion or 10 billion-pound data center. 400 jobs. Because these things are so efficient...” (44:27)
7. Government, Policy & ‘Gee Whiz’ Enthusiasm
- Governments from local to national levels keep rolling out the red carpet for data center projects, lacking effective plans for oversight or community benefit—lured by the technological promise and lobbying of tech giants.
- Quote:
- Josh: “There’s this factor of gee-whiz, like these titans of the AI industry are good at razzle dazzling elected officials into doing whatever they want.” (46:06)
- Chuck: “Maybe it'll all work out great.” (46:31)
- Josh: “Sure, it probably will. It usually does. Astoundingly, it usually does work out well.” (46:35)
8. UK/EU Context
- The UK is the third-largest market for data centers (after the US and Germany).
- Current investments include Microsoft’s $30B in UK centers and plans for ~100 new UK AI data centers; repurposing of old factories is happening, but controversy exists around jobs, water, and economic benefit.
- UK data centers consume 10 billion liters of drinking water per year.
9. Memorable Moments & Quotes
On how old data centers were managed:
- Josh: “When the PC came along and then the Macintosh came along, they took that thing and just made it very small so you could put it on all of your employees desks.” (09:53)
On the nostalgia for less connected times:
- Chuck: “Imagine working from home at that time (2008). It was just—you couldn’t. You went home and you homed.” (20:06)
On the anxiety about AI:
- Josh: “We're gonna reach some point potentially where we just put that extra last GPU in there and all of a sudden the balance is tipped and the thing becomes super intelligent.” (26:48)
10. Final Thoughts
- Data centers are foundational to everything from banking to remote work, AI, and the next digital revolution, but their impacts—environmental, economic, social—are outsized, mounting, and not well managed.
- Everyone using digital services is intimately tied to this infrastructure, for better or worse.
Key Timestamps
- What are data centers? (03:01)
- Colossus/ENIAC, mainframe history (07:30–10:49)
- Defining the data boom (zettabytes consumed) (03:23–05:00)
- Rise of the Internet and Cloud Computing (15:29–19:52)
- Hyperscale & global arms race (20:12–23:41)
- AI and GPUs explained (23:41–25:55)
- Financial frenzy and warnings (27:54–35:58)
- Environmental/energy costs (37:08–43:09)
- Local/community impact, lack of jobs (43:09–44:27)
- Policy and future risks (45:02–46:35)
Recommended Further Listening:
- “The End of the World with Josh Clark” (Josh’s limited series, with an episode on AI risks) (27:02)
Overall Tone:
Conversational and humorous, but with undercurrents of skepticism about unchecked technological expansion and a call for awareness on the broad implications of our data-driven world.
