They
say “ignorance is bliss”, but that's not really true. Everybody
knows it's not a good thing to bury your head in the sand. No matter
how uncomfortable or painful the truth is, IF you need to know THEN …
YOU NEED TO KNOW!
When I was very young, my mother had a habit of
answering my questions with “You'll understand it when you get
older.” Her answer was frustrating, but guess what? When I got
older, I did understand.
A lot of tech talk goes straight over my head. While I may not understand what's being said, I don't ignore it and forget it. I usually file it away in memory banks because … one day I might understand.
I apply my mom's wise words to nebulous (might be "nefarious") tech talk that I
hear or read. I sigh and think to myself: 'I'll understand later.'
π‘π‘π‘
While
casually perusing a newsletter I subscribe to, these statements
caught my eye.
“AI
data centers are draining the U.S. grid faster than it can be
rebuilt. Utilities are warning of shortages. Federal agencies are
fast-tracking permits. Big Tech is scrambling to lock up turbines,
pipelines, and baseload power just to keep their AI systems alive.”
(SOURCE: WealthDaily.com)
I
immediately thought: 'Nah! That can't be right. (O.o) Is it?'
So
I did a little research, and I have to say that … I'm a bit
alarmed.
- BOTTOM
LINE: The
core claims in that statement are broadly supported by current
reporting,
though with important nuance. Here’s what the evidence shows,
broken down by each part of the claim.
◆
1. “AI
data centers are draining the U.S. grid faster than it can be
rebuilt.”
(Supported, with nuance.)
Multiple authoritative sources confirm that
AI-driven data center
growth is outpacing the speed at which utilities can expand
generation and transmission:
The North
American Electric Reliability Corporation (NERC) issued a Level
3 alert—its highest warning—citing “significant risks” to grid reliability from large data centers and AI loads.
POWER
Magazine reports that power availability has become the limiting factor for where and how fast hyperscale AI data centers can be built,
because these facilities now require 300–600
MW each, comparable to a mid-sized city.
Forbes notes that regions like Northern Virginia and Texas are hitting capacity limits, with utilities struggling to keep up with demand.
Conclusion:
The grid is not literally being “drained,” but demand
from AI data centers is growing faster than grid infrastructure can
be expanded, creating real
constraints*.
[*So? They're NOT crying wolf?πΊ]
◆
2. “Utilities
are warning of shortages.”
(Supported.)
Utility CEOs in Texas*, the Midwest, and Virginia have publicly warned that
AI-related load growth is straining capacity.
Grid operators in Texas, the Great Plains, and the mid-Atlantic region project major electricity demand spikes driven largely by data centers.
NERC’s Level 3 alert itself is a formal warning of potential reliability shortfalls.
Conclusion:
Yes — utilities and grid monitors are issuing explicit warnings.
[*Really? I live in Texas, and the ads I see
on TV practically say data centers were sent from God! The Texas data centers offer
high-paying jobs and are used to run systems that keep our schools
and streets safe. Any idea how many mass shootings we've had at
schools in Texas?]
◆
3. “Federal
agencies are fast-tracking permits.”
(Partially supported.)
While the statement is directionally
correct, the evidence is more nuanced:
The Federal
Energy Regulatory Commission (FERC) issued Order
2023, which streamlines interconnection processes and accelerates timelines for connecting new generation and large loads.
This indirectly helps data centers by speeding up the clean energy projects they depend on.
Grid operators and regulators have formed special task forces to accelerate large load interconnections.
Conclusion:
Federal agencies are
taking steps to accelerate permitting and interconnection, though the
phrase “fast tracking permits” is a simplification.
[*Fast-tracking? Under the Trump
administration? Why am I not surprised?]
◆
4. “Big
Tech is scrambling to lock up turbines, pipelines, and baseload power
just to keep their AI systems alive.”
(Supported in spirit, though phrased
dramatically.)
Evidence shows that major tech companies are
aggressively securing dedicated power sources:
Some developers are contracting directly with private power producers or installing on-site natural gas generators because grid capacity is insufficient.
Utilities and developers are turning to natural gas and nuclear to provide
24/7 baseload power for AI data centers.
Exxon and Chevron have announced plans to enter the power supply market specifically to serve data centers, including natural gas-powered generation.
Conclusion:
Big Tech and energy companies are indeed racing to secure* reliable
baseload power, though the imagery of “scrambling” is rhetorical.
[*Seems like Big Tech and energy companies
are always racing to secure their “power”. What do they care?
They're moving to Mars. Right?]
◆◆
Overall Assessment
The WD statement is directionally
accurate and substantially
supported by current reporting,
though it uses dramatic language. Here is a summary of the VERIFIED
FACTS:
AI data center demand is growing extremely fast.
Grid capacity is struggling to keep up.
Utilities and regulators are issuing warnings and accelerating processes.
Tech companies are securing dedicated energy sources, including gas and potentially nuclear.
The only part that is somewhat overstated is
the implication* of a federal emergency–style “fast track,”
but federal and regional regulators are
accelerating interconnection and permitting reforms.
[*I used to work for the federal government.
Based on my real work experience, I'm inclined to believe the
“implication” is SPOT ON! When you're assessing anything being
done by the federal government and the US military, ALWAYS IMAGINE THE
WORST-CASE SCENARIO. If it turns out not to be how you imagined, you'll
breathe a sigh of relief.]
***
SOURCES/REFERENCES:
►► AI Data Centers
Straining the Grid
Baker, Mike.
“Data Centers Are Gobbling Up America’s Electricity.” The
New York Times, 12 Mar.
2024.
Nicas, Jack.
“The A.I. Boom Could Use a Lot of Power. How Green Will It Be?”
The New York Times,
9 Feb. 2024.
►► Utility Warnings
and Grid Reliability Alerts
North American Electric Reliability
Corporation (NERC). 2023–2024
Long-Term Reliability Assessment.
NERC, Dec. 2023.
North
American Electric Reliability Corporation (NERC).
NERC Issues Level 3 Alert on Large
Load Interconnection. NERC, 2024.
►► Federal Agencies
Accelerating Permitting / Interconnection
Federal Energy Regulatory Commission
(FERC). Order
No. 2023: Improvements to Generator Interconnection Procedures and
Agreements. FERC, 2023.
Eisen,
Joel B., and Emily Hammond.
“FERC’s New Interconnection Rules Aim to Speed Up Clean Energy
Deployment.” Harvard Environmental
Law Review, 2023.
► ► Big Tech Securing
Turbines, Pipelines, and Baseload Power
Meyer, Robinson.
“Big Tech’s Next Big Energy Play.” Heatmap
News, 2024.
Sorkin,
Andrew Ross. “Tech Giants Are
Buying Up Power to Feed A.I.” The
New York Times, 18 Apr.
2024.
DiChristopher, Tom.
“Exxon and Chevron Plan to Supply Power to Data Centers.” CNBC,
22 Jan. 2024.
► ► Regional Grid
Constraints (Virginia, Texas, Midwest)
Sverdlik, Yevgeniy.
“Northern Virginia’s Data Center Boom Is Running Out of Power.”
Data Center Knowledge,
2023.
Malik, Naureen.
“Texas Grid Strains Under Data Center and Crypto Mining Demand.”
Bloomberg,
2024.
It's not just Texas and Virginia. Georgia and Nevada are also having problems. There very likely will be more!
***
NOTE: As I stated earlier, a lot of tech talk goes over my head. So the research was done, and this content was prepared using a free Microsoft AI tool. The
notes marked [*...] are just ME (not AI), just babbling to myself. π