Photo via Al Jazeera

AI’s real bottleneck is electricity. Who will pay for it?

Read the original at Al Jazeera ↗
Cite

AlSammarraie, A. (2026, October 30). AI’s real bottleneck is electricity. Who will pay for it? Al Jazeera. https://www.aljazeera.com/opinions/2026/10/30/ais-real-bottleneck-is-electricity-who-will-pay-for-it?traffic_source=rss

Dated from the outlet's own feed. Check the original for its byline.

Al Jazeera · collected 2026-10-30 · by AlHasan AlSammarraie

Quick Summary

Two of the largest cloud computing companies, Amazon and Google, have signed long-term contracts for nuclear-generated electricity. Amazon has contracted for an additional 190 megawatts from existing reactors at Constellation’s Calvert Cliffs plant by 2030-2032, while Google will receive up to 890 megawatts of new capacity from upgraded reactors in Illinois, Pennsylvania, and New Jersey. These agreements come as nuclear power is regaining favor due to increasing electricity demand driven by AI and other technologies, despite past safety concerns and previous declines in nuclear energy use worldwide.
Written by AI on 2026-10-30, from this article's own text only (the story summary below draws on all the coverage).

AI analysis

Story summary

Google and Amazon, two of the world's largest cloud computing companies, have signed 20-year contracts with Constellation Energy to secure additional nuclear electricity capacity in the United States. Amazon has committed to purchasing 690 megawatts from Constellation’s Calvert Cliffs plant in Maryland, including an extra 190 megawatts expected between 2030 and 2032. Meanwhile, Google has agreed to bring 890 megawatts of additional nuclear capacity into the U.S., though neither company is building new reactors; instead, they are paying for increased output from existing ones. Constellation plans to invest over $4.3 billion to upgrade 11 reactors across Illinois, Pennsylvania, and New Jersey.

This shift marks a significant turnaround from previous years when nuclear power faced skepticism due to past accidents at Three Mile Island in 1979, Chernobyl in 1986, and Fukushima in 2011. In response, countries like Germany phased out their nuclear programs entirely, while Japan halted operations but has since restarted some reactors. Despite these historical setbacks, the current agreements highlight a renewed interest in nuclear energy for large tech companies seeking reliable power sources to support their growing AI demands.

Written for “Ai Electricity Costs” on 2026-10-30, grounded in this article and the 0 other(s) covering the same event.
Why this leaning score
The article's own words the score was based on. Each is quoted verbatim and was checked against the article text before being stored, so you can find it in the original.
Reading Leans left (beta estimate) Confidence medium
Leaning: leans left for article 116676 (medium confidence, 4 verified quotes) · logged 2026-10-30

Story

📰 Ai Electricity Costs
Technology · 1 article(s) covering the same event.

Compared with similar articles

This article reads leans left and hedges 8% of its claims. Each row says how that neighbour differs.
The Dispatch
Hedges 0 of 3 claims Outlet record 61/100
“The articles discuss entirely different topics: one is about a cyberattack on a water utility in Alaska, while the other is about cloud computing companies signing contracts for nuclear electricity.”
The Sydney Morning Herald
Hedges 8 of 47 claims Outlet record 61/100
“The articles discuss different topics related to AI but describe separate events - one about Anthropic's potential IPO and the other about cloud computing companies securing long-term electricity contracts.”
Semafor
Hedges 2 of 4 claims Outlet record 60/100
“The articles discuss different aspects of energy challenges for AI companies but describe distinct events: one about potential delays and high costs in securing energy, the other about concrete contracts signed by cloud computing companies.”
Semafor
⚖️ Leans right further right than this Hedges 1 of 35 claims Outlet record 60/100
“The articles discuss different topics related to AI but describe distinct events: one about UK funding for an AI testing lab and the other about cloud computing companies signing contracts for nuclear electricity.”
Evening Standard
Hedges 0 of 7 claims Outlet record 65/100
“The articles discuss different topics related to AI but describe separate events - one about S4 Capital's sales performance and another about cloud computing companies signing contracts for nuclear electricity.”
South China Morning Post
⚖️ Centre further right than this Hedges 0 of 2 claims Outlet record 64/100
“The articles discuss different aspects of AI infrastructure challenges and do not describe the same specific incident or occurrence.”
Washington Examiner
Hedges 2 of 2 claims Outlet record 70/100
“The articles discuss different aspects of AI and technology but do not describe the same specific incident or occurrence.”
National Post
Hedges 0 of 28 claims Outlet record 60/100
“The articles discuss different topics: one focuses on an upcoming EU-Canada summit related to AI accountability and international cooperation, while the other discusses electricity contracts for cloud computing companies.”
Reason
⚖️ Leans left Hedges 6 of 22 claims Outlet record 67/100
“The articles discuss different topics related to AI and energy but describe distinct events involving AI in fossil fuel extraction versus cloud computing companies securing long-term contracts for nuclear electricity.”
The Intercept
⚖️ Leans strongly left further left than this Hedges 9 of 79 claims Outlet record 97/100
“The articles discuss different topics related to AI but do not describe the same specific incident.”
How this article was measured signals, leaning charts, publisher record, byline, topics and claims

Signals How these are calculated →

Claims extracted
65
sentences that state something as happening or true
Uncertain
8%
5 of 65 hedged ("reportedly", "may", "according to")
Leaning
Leans left
of the writing, not the subject · beta estimate
Publisher signals
60/100
Al Jazeera's corrections and hedging, across everything collected from it. Not a measure of accuracy.
Outlets on this story
1
Technology
Same framing elsewhere
1
articles that frame the event the way this one does
Analyzed 2026-10-30 · how these are computed

How this is being covered How these are calculated →

Article leaning vs. publisher reliability
Source leaning vs. consistency

Publisher

Al Jazeera · 4411 article(s) · 2 correction(s) detected
Running correction rate · 2 correction(s)

Who wrote this

AlHasan AlSammarraie
2 article(s) here · 1 carrying a prediction
🔮 Google has signed an agreement that will bring 890 megawatts of additional nuclear capacity into the United States.
🔮 In addition, Sam Altman said that OpenAI will not be going public in 2026, as further safety advances are still required.
2026-09-18 · assertive framing · Who gets to decide how quickly AI moves?
Also by AlHasan AlSammarraie
Nothing else under this byline is closely related to this article, so these are simply their most recent.

Topics

Amazon Calvert Cliffs Constellation Google Maryland

Subjects

Constellation ORG · 2× Amazon ORG · 1× Calvert Cliffs GPE · 1× Chornobyl GPE · 1× Google ORG · 1× Illinois GPE · 1× Maryland GPE · 1× New Jersey GPE · 1× Pennsylvania GPE · 1× the United States GPE · 1×

Narrative

The Gulf is in a rare position of being able to settle that question first: what the industrial tariff is, who funds the new generation, what happens to the contracts when the hardware ages, and what long-term economic benefits the country would see beyond energy revenues and a limited number of jobs.
framing: assertive · carried by 1 article(s) · first seen 2026-10-30
🔮 Google has signed an agreement that will bring 890 megawatts of additional nuclear capacity into the United States.
2026-10-30 · Al Jazeera
AI’s real bottleneck is electricity. Who will pay for it? · assertive framing

Claims (65 extracted, 5 hedged)

Within a week, two of the three largest cloud computing companies in the world signed 20-year contracts for nuclear electricity. asserted
two → sign → electricity
Amazon has taken 690 megawatts from Constellation’s Calvert Cliffs plant in Maryland, including 190 megawatts of additional generating capacity expected between 2030 and 2032. asserted
Amazon → take → 2030
Google has signed an agreement that will bring 890 megawatts of additional nuclear capacity into the United States. asserted
that → sign → States
Constellation is investing more than $4.3bn to upgrade 11 reactors across Illinois, Pennsylvania, and New Jersey. asserted
Constellation → invest → Illinois
Neither agreement, however, involves building a new nuclear reactor. asserted
agreement → involve → reactor
Rather, they are paying to generate more output from reactors that already exist – some of which are even older than the companies themselves. asserted
some → pay → companies
A few years ago, such agreements would have seemed inconceivable. asserted
agreements → seem → ?
Nuclear power had fallen out of favour following accidents at Three Mile Island in 1979, Chornobyl in 1986 and Fukushima in 2011. asserted
power → fall → 2011
Germany phased out nuclear power, while Japan shut down its reactors after Fukushima, although it has since begun restarting them. asserted
it → phase → them
In many democratic countries, advocating for nuclear energy became politically difficult. asserted
advocating → advocate → energy
Today, an undamaged reactor at Three Mile Island is being brought back into service, with Microsoft contracted to buy its output for 20 years. asserted
Microsoft → bring → years
The US Department of Energy is providing up to $4.4bn in loans to support the restart of three shuttered plants, with the aim of restoring about 2.25 gigawatts. asserted
Department → provide → gigawatts
Palisades in Michigan, which was closed in 2022, is contracted to be generating again by March 2027. asserted
which → close → March
If successful, it would become the first US commercial nuclear plant to restart after entering decommissioning. asserted
it → become → decommissioning
What has changed? asserted
What → change → ?
The quiet reversal In the last decade, the industry’s answer to AI’s energy problem was efficiency – things such as better chips, smaller models and smarter cooling seemed appealing and feasible. asserted
things → seem → chips
Industry conferences featured discussions about performance per watt, with modern artificial intelligence (AI) chips performing more computation per unit of electricity than their predecessors. asserted
conferences → feature → predecessors
Yet those efficiency gains did not reduce overall electricity demand. asserted
gains → reduce → demand
Instead, each gain was quickly put towards something larger. asserted
gain → put → something
Even when DeepSeek demonstrated that a competitive model could be trained at a fraction of the computing cost many had expected, the result was not a reduction in industry spending. uncertain
result → demonstrate → spending
Instead, laboratories continued planning more training runs, models that reason for longer, agents that work continuously in the background, and video generation. asserted
that → continue → background
This would not be the first time something like this has happened in history. asserted
something → happen → history
In 1865, William Jevons observed that more efficient steam engines made Britain burn more coal, not less, because cheaper power created uses that had not existed before. asserted
that → observe → uses
The International Energy Agency estimates that global data centre electricity demand will more than double to around 945 terawatt-hours by 2030, even assuming continued improvements in efficiency. asserted
demand → estimate → efficiency
Hence, the industry has changed strategy – unable to shrink demand, companies are now buying up supply. asserted
companies → change → supply
The arithmetic of waiting Money, however, cannot solve every problem quickly. asserted
Money → wait → problem
A new large reactor takes a decade or more from decision to electricity. asserted
reactor → take → electricity
Tech companies are investing in small modular reactors to power their data centres, although many of these projects are not expected to deliver electricity until the early 2030s. asserted
many → invest → 2030s
So far, very few of these have been built anywhere in the world. asserted
few → build → world
Thus, while reactor restarts and upgrades are the faster option, the “fast option” still means 2027, 2028, 2029, or beyond. asserted
option → mean → 2027
AI capital expenditure, however, does not operate on that same calendar. asserted
expenditure → operate → calendar
The chips being installed this year could be overtaken by newer technology long before the 20-year electricity contracts powering them expire. uncertain
contracts → instal → them
This discrepancy underpins all of these announcements, yet the industry has not addressed the implications if demand weakens while contractual obligations remain. asserted
obligations → underpin → implications
Who pays in the meantime This is where the story stops being about technology. asserted
story → pay → technology
Most of these deals sit inside PJM, the grid operator serving roughly 67 million people across 13 American states and Washington, DC. asserted
Most → sit → states
PJM uses capacity auctions to secure sufficient electricity-generating resources for future demand, with the associated costs ultimately passed on to consumers through their electricity bills. asserted
PJM → use → bills
Capacity prices have reached the regulatory price cap in three consecutive auctions. asserted
prices → reach → auctions
According to the grid operator’s independent market monitor, data centre demand contributed approximately $29.4bn in capacity costs across the four most recent auctions. uncertain
demand → accord → auctions
In the latest auction alone, around $6.3bn of these costs were attributed to data centre demand, including forecasts for facilities that have yet to be built. asserted
that → attribute → facilities
In effect, households are already paying higher electricity capacity charges, partly because of projected demand from facilities that may never become operational. uncertain
that → pay → facilities
…and 25 more, not listed.
💬Give feedback
🕘History 🎫Support