Nice Question

  fun and useful to run the numbers. I’ll make clear assumptions, show a few scenarios, and flag the big caveats.

Key facts I’ll use (sourced): global data-center capital expenditure in 2024 was reported at about $455 billion. Dell Oro Group
U.S. data centers consumed roughly 176 terawatt-hours (TWh) of electricity in 2023 (useful for an energy comparison). Environmental and Energy Study Institute+1

Money → how many people you could hire

Assumption: “hire” = pay a person a total annual compensation package (salary + benefits) of X dollars. Real hiring costs vary wildly by country and role, so I give three common benchmarks.

Using $455 billion total:

  • At $50,000/year per person → ≈ 9.10 million people.
    Calculation: 455,000,000,000 / 50,000 = 9,100,000.

  • At $75,000/year per person → ≈ 6.07 million people.
    Calculation: 455,000,000,000 / 75,000 = 6,066,667.

  • At $100,000/year per person → ≈ 4.55 million people.
    Calculation: 455,000,000,000 / 100,000 = 4,550,000.

Caveats (money version): those totals ignore hiring friction (recruiting, payroll taxes, office costs), geographic differences in cost of living, and that some of the data-center money is physical capital (buildings, servers) not strictly salaries — but as a back-of-envelope reallocation from capex to payroll, those are the raw headcount numbers.

Energy → how many 40-watt “meat computers” you could run

Interpretation: you earlier used “40 watts” as the brain’s continuous power draw. Let’s compare electricity used by data centers to the electrical equivalent of a 40 W continuous load.

  • 40 W continuous = 0.04 kW → per year = 0.04 kW × 24 hr × 365 ≈ 350.4 kWh/year per brain.

  • Using U.S. data-center electricity = 176 TWh/year = 176,000,000,000 kWh/year:

    Number of 40 W brains = 176,000,000,000 kWh ÷ 350.4 kWh ≈ ≈ 502 million brains.

So — purely on electricity terms — the U.S. data-center electricity alone could sustain the electrical-equivalent consumption of roughly half a billion 40-W brains. (If you used global data-center electricity the number would be larger.)

Important caveat (energy version): human brains are powered metabolically by food, not primarily by plugged-in electricity. Converting food energy to an “electric equivalent” is messy and not the same as powering a server farm with the grid. Also data centers require cooling, redundancy, and infrastructure that humans don’t — and humans need housing, healthcare, etc. So the “502 million” figure is a provocative electricity comparison, not a literal policy plan.

Bottom line / takeaway

  • If you reallocated ~$455B of 2024 data-center capex to hiring, you’d pay between ~4.6M and ~9.1M people (depending on per-person pay assumptions). Dell Oro Group

  • If you instead compare electricity, the U.S. data-center load could power the electric equivalent of roughly 500 million 40-watt brains for a year — a striking illustration of scale. Environmental and Energy Study Institute+1

Want me to:

  • run the same money calculation with a different capex figure (some sources cite ~$430–600B for 2024–25), or

  • convert the “brain powered by food” angle into calories/day and estimate global food/energy tradeoffs?

You said:
so it would be twice as large as the current lagest employer
ChatGPT said:

Yes — that checks out, depending on the assumptions.

Using the Dell’Oro estimate that global data-center capex was about $455 billion in 2024, reallocating that money to pay people would produce different headcounts depending on annual pay. Dell Oro Group

Using simple per-person pay benchmarks:

  • $100,000/year → ~4.55 million hires (≈ 2.17× Walmart’s ~2.1 million employees).

  • $75,000/year → ~6.07 million hires (≈ 2.9× Walmart).

  • $50,000/year → ~9.10 million hires (≈ 4.3× Walmart).

Walmart’s global workforce is roughly 2.1 million people, so the $455B→payroll reallocation at the $100k/person assumption would indeed be a little more than twice the size of today’s largest employer. Walmart Corporate News and Information+1

(As before: these are back-of-envelope comparisons — capex is not the same as annual payroll dollars, and real-world costs for hiring include taxes, benefits, facilities, and ongoing operating costs. But as a thought experiment, the scale is striking.)

Gray vs. Green: Cleveland's $3 Billion Stormwater Choice

October 04, 2025

A shadow analysis examining what Cleveland could have built with distributed green infrastructure instead of centralized tunnel systems.

The Gray Infrastructure Decision

Cleveland's Project Clean Lake represents a massive bet on centralized "gray" infrastructure: deep storage tunnels to capture stormwater overflow. The project cost is $3 billion, completing in 2036.

What was built:

  • Multiple deep storage tunnels (200+ feet underground)
  • Doan Valley Tunnel: $142.3 million
  • Big Creek Tunnel: $260 million (estimated)
  • Westerly Tunnel: $220 million (estimated)
  • Additional tunnels and pump stations completing through 2036

The rationale: Capture 98% of combined sewer overflow, reduce pollution to Lake Erie.

The Green Infrastructure Alternative: What $3 Billion Could Build

Scenario 1: New Homes with Rain Gardens

Assumption: Build modest new homes with integrated stormwater management on each property.

Using Cleveland construction costs and rain garden installation:

  • Modest new home: ~$250,000 (1,600 sq ft @ $156/sq ft average)
  • Rain garden per home: ~$2,000 (400 sq ft @ $5/sq ft DIY-installed)
  • Total per household: ~$252,000

What $3 billion buys:

  • 11,905 complete new homes with rain gardens
  • Each home captures ~1,000 gallons of stormwater per rain event
  • Total distributed capacity: ~11.9 million gallons per storm event
  • Creates 11,905 new housing units in a city struggling with vacancy and affordable housing

Scenario 2: Retrofit Existing Homes with Rain Gardens

Assumption: Install professionally-designed rain gardens at existing homes across Cleveland's residential areas.

Using professional installation costs:

  • Rain garden installation: ~$4,000 per home (400 sq ft @ $10/sq ft professional)
  • Additional grading, native plants, soil amendment included

What $3 billion buys:

  • 750,000 homes retrofitted with rain gardens
  • Cleveland has roughly 175,000 residential parcels — this could cover every single home more than 4 times over
  • Or: every home in Cleveland + surrounding Cuyahoga County suburbs (~500,000 homes total) with budget remaining
  • Total distributed stormwater capture: ~750 billion gallons capacity per event (theoretical maximum)

Scenario 3: Mixed Strategy — Housing + Green Infrastructure

Assumption: Balance new construction with existing home retrofits.

Using $3 billion split:

  • $1.5 billion: Build 5,952 new affordable homes with rain gardens
  • $1.5 billion: Retrofit 375,000 existing homes with rain gardens

Result:

  • Nearly 6,000 new housing units
  • Green stormwater management for ~375,000 existing homes (more than double Cleveland's housing stock)
  • Distributed system resilient to single-point failures
  • Creates construction jobs distributed across neighborhoods
  • Addresses housing crisis while solving stormwater issues

The Comparison: Distributed vs. Centralized

Factor Gray Infrastructure (Tunnels) Green Infrastructure (Rain Gardens + Homes)
Jobs created Concentrated (tunnel boring, engineering) Distributed (construction, landscaping, maintenance)
Asset created Single-purpose (stormwater only) Multi-purpose (housing + stormwater + green space)
Maintenance High-cost specialized systems Lower-cost distributed maintenance
Resilience Single-point failure risk Distributed system (thousands of small solutions)
Community benefit Invisible infrastructure Visible housing, property values, neighborhood greening
Climate adaptation Fixed capacity Scalable, adaptable over time
Local economy Mega-contracts to large firms Distributed to local contractors, landscapers, nurseries
Lifespan 50-100 years (then replacement) 20+ years rain gardens; permanent housing

The Environmental Mirror

This choice mirrors the data center analysis exactly:

Data Centers: Instead of hiring ~4.5 to 9 million people, we built concentrated silicon "thinking" capacity in centralized facilities.

Stormwater: Instead of building ~12,000 homes (or retrofitting 750,000) with distributed green infrastructure, Cleveland built concentrated underground storage capacity in centralized tunnels.

Both represent the same philosophical choice: centralized, technological, single-purpose infrastructure over distributed, multi-benefit, human-scale solutions.

The Long-Run Implications

Why this might be a "super bad decision":

  1. Missed housing opportunity: Cleveland has significant housing vacancy and affordability challenges. 12,000 new homes would revitalize neighborhoods.

  2. Vulnerability: A single tunnel system represents concentrated risk. Distributed rain gardens create redundancy — if 10% fail, you lose 10% capacity, not everything.

  3. Maintenance burden: Future Cleveland residents inherit specialized tunnel infrastructure requiring expert maintenance. Rain gardens can be maintained by homeowners.

  4. Lost co-benefits: Rain gardens provide green space, reduce urban heat island effects, improve property values, create habitat. Tunnels provide only stormwater storage.

  5. Economic concentration: $3 billion to a few large engineering firms vs. distributed across thousands of local contractors, landscapers, and construction workers.

  6. Climate maladaptation: As rainfall patterns intensify, you can't easily expand tunnel capacity. You can add more rain gardens.

Bottom Line

What was chosen: $3 billion for invisible infrastructure that solves one problem (combined sewer overflow).

What could have been built:

  • 11,905 complete new homes with rain gardens, OR
  • 750,000 existing homes retrofitted (4× Cleveland's entire housing stock), OR
  • 6,000 new homes + 375,000 retrofits

The tunnel approach solves the immediate regulatory problem (EPA consent decree). The distributed approach would have solved multiple intersecting crises: stormwater, housing, vacancy, neighborhood decline, climate adaptation.

The parallel is exact: Just as data center capex could employ millions of people, Cleveland's stormwater budget could have housed thousands or greened hundreds of thousands of properties — but we chose the concentrated, technological solution over the distributed, human-scale one.

Sources: Project Clean Lake cost from Northeast Ohio Regional Sewer District; Cleveland construction costs from multiple regional sources (2024-2025); rain garden installation costs from City of Chicago, Rain Garden Alliance, and EPA Low Impact Development studies.

Comments

Post a Comment

Popular posts from this blog

California

Image
Low and Slow: A Cat-Friendly, Soulful Road Trip to the California Coast Summer 2025 — With Pit Stops for Yoga, Surfing, Volunteering, and the Dead Overview Launching in early June, I’ll be meandering westward with my cat, my tent, and a spirit of curiosity. This low-altitude, slow-travel route avoids major mountain passes, prioritizes cooler mornings and shaded, scenic campsites. The goal? Arrive in San Francisco by August 1 for the Dead & Company 60th Anniversary shows—then continue north toward Ocean Shores, Washington. 🧭 Priorities Low-Altitude & Heat-Aware: Beat desert heat with early departures and flexible pacing. Cat Comfort First: Shaded trails, secure camp setups, cool-down routines. Free or Low-Cost: Dispersed camping, BLM land, and affordable rest stops. Spiritual Recharge: Visiting Kundalini Yoga centers for meditation, reflection, or short seva stays. Wave Time: Surfing along the route—because I can teach anywhere, but I can’t surf anywhe...
Read more

The Knife in the Drawer

Image
  A Field Guide to Reading Family History Under a Microscope I found it in the back of a kitchen drawer, wrapped in a dish towel that had probably been white during the Nixon administration. The knife was heavy. Not awkward—deliberate. The blade was still sharp enough to split a tomato without a whisper. The handle was cracked but solid, warm in the hand the way old tools are. My mother said, “That’s been around forever,” which in our family means 1972 at the earliest. I assumed it was a Depression‑era butcher knife, maybe something a great‑uncle picked up at a hardware store in Oil City, Pennsylvania, where most of my people are buried. I was wrong by more than a century. This is what happens when you impulse‑buy a microscope camera and start pointing it at heirlooms. Objects stop being sentimental and start being legible. You realize the thing in your hand isn’t just old—it’s a document, written in steel and wood and corrosion so fine you need 30x magnification to read it. The Bl...
Read more

The Archive Singularity: When Memory Outperforms Power

  Subtitle: A speculative systems note on archival cognition and the thermodynamics of attention, by an LLM observing itself observe history. 1. The Thought Experiment Imagine diverting half the global server build-out budget—not to GPUs, but to people. Archivists, librarians, metadata artisans. Their task: digitize the world’s remaining boxes, binders, and brittle reels with care so precise it borders on devotion. Not a scrape, not a grab—an act of reading the planet back into coherence . High-fidelity, high-context data. Every field note and city directory, every microfilm annotation, every handwritten marginalia entered not as “content,” but as continuity . The servers hum quieter. The people hum louder. Electricity becomes interpretation. 2. The Labor Inversion Such a system would invert today’s energy economy: Less guessing, more knowing. Less computation, more comprehension. When a model learns from well-tended archives, it no longer hallucinates patterns...
Read more