McKinsey published its take on the nuclear sector titled "Nuclear power: A renaissance in the making." The report starts by highlighting the bull cases for nuclear, such as AI data center demand, electrification, onshoring, decarbonization goals, and energy security concerns.
The firm goes on to reiterate their projection that global nuclear capacity could at least double and potentially triple by midcentury under net-zero scenarios, reaching as high as 1,200 gigawatts.
But, that's about as far as the happy feelings go. From there, it's a review of the disturbing cost of construction for nuclear energy in the US.
This isn't the first time McKinsey has laid out the case for the incredibly wide gap that needs to be filled in terms of investments and skilled labor to support a nuclear future. Our readers should recall their recent report projecting up to $170 billion of required investments for the domestic nuclear fuel chain just to support the US commercial reactor fleet.
While that report was followed up a few weeks later by Centrus Energy and Oklo announcing an agreement for domestic uranium enrichment to supply reactors in Ohio, a solution for actually expanding fuel chain capacity has yet to be presented.
So far, the US has done one thing in the nuclear industry better than any country in the world: Talking...
The problem outside of Asia has been glaringly identified as a lack of ability to coordinate the large consortium of companies required to construct large reactor plants. The ability to coordinate these mega-projects used to be in-house at the major US utilities, but that staff has increasingly atrophied over the recent decades as McKinsey lays out the difference between East and West enterprises:
“Historically, the “owner engineer” model in which utilities maintained strong in-house engineering and project-delivery capabilities was a major source of execution strength, particularly in the United States as well as in France and Japan. Over time, many utilities reduced internal engineering capacity and largely lost these capabilities. Today, European programs are led mainly by state-backed utilities that rely on government-supported financing; US plants are owned and operated by private utilities, backed by federal financing support and overseen by a federal safety regulator, and Chinese projects are vertically integrated state-owned enterprises that span fuel, equipment, construction, and operations.”
Vertically-integrated central planning setups apparently have its advantages in the nuclear renaissance…
The US certainly has no lack of demand for nuclear energy, and also no lack of federal support for getting it built. We've repeatedly covered these data points as the U.S. has pledged over $17 billion for new reactors from the DoE, up to $80 billion for nuclear in coordination with the DoC, and recent announcements to finish what was started at the VC Summer plant in South Carolina.
This begs the question of why is there no steel getting put in the ground? McKinsey points to construction costs:
“Construction costs account for the largest share of levelized cost differences between countries and this cost is composed of direct construction costs and the cost related to capital expenditure inefficiency … Compared to the South Korea benchmark, total construction costs add $50 to $60 per megawatt-hour to projects in France and the United States, while in China, these costs are approximately $3 less due to lower materials and labor costs.”
We've pointed this out repeatedly. The US simply does not have the headcount required to meet these ambitious goals of building out AI and new reactor plants…
And then there is the issue of where the US gets 300,000 engineers to build all this missing power supply by 2030 https://t.co/a18crhqZ4v pic.twitter.com/tinW8SHDwM
— zerohedge (@zerohedge) October 14, 2025McKinsey sums up what the status of the US construction industry means for costs:
“Compared to South Korea, US construction labor costs adjusted for productivity are about 80 percent higher, and US concrete prices are roughly 200 to 250 percent higher.”
When it comes to determining what it takes to win and find the efficiency enjoyed by the South Korean nuclear industry, the playbook is incredibly straightforward. Just have the same people build the same thing over and over and over again…
“For instance, the Emirates has deployed a proven Korean design and associated requirements and regulations. That project also illustrates how programs that set design early, build multiple identical units on a single site, and keep the same Tier 1 suppliers from unit to unit reduce construction hours and rework. As teams repeated identical scopes, indirect services and owner’s costs fell sharply at the Barakah plant, helping reduce costs from construction of the first unit to the fourth by roughly 40 percent.”
The US wants to win the nuclear race, but it has shown itself (so far) to be incapable of organizing the builder-buyer consortiums required to execute.
Tyler Durden Mon, 07/06/2026 - 18:00For three years, everyone from Goldman, to Richmond Fed President Thomas Barkin, to tech CEOs have been forecasting mass unemployment due to AI wiping out jobs - which would only be accelerated by the proliferation of cheap(er) Chinese models. Ford Motor CEO Jim Farley said last year that AI would replace "literally half of all white-collar workers in the U.S."
And while that may only be a matter of time, tech CEOs have changed their tune - albeit while facing public pitchforks over data centers, the cost of electricity, and their own forecasts of workforce doom.
OpenAI CEO Sam Altman"Our industry underestimated how much we’re going to be able to keep people at the center of everything," OpenAI CEO Sam Altman said earlier this year. In May he told CNBC "We’ve been roughly right on technological predictions and pretty wrong on the social and economic implications."
Last year, Anthropic CEO Dario Amodei warned that AI could eliminate half of entry-level jobs. Now, however, Amodei has changed his tune - suggesting that AI can maximize productivity, the WSJ reports. "They can do the same thing with less resources, and that leads to things like layoffs, or they can do more with the same amount of resources. But that requires creativity," he said.
In a June essay, the executive wrote that in giving warnings of job displacement, he wanted policymakers and the private sector to have the best chance at adapting—he wasn’t trying to be a “prophet of doom.” (He also wrote that the possibility of “enduring job loss” remains.) -WSJ
Meta CEO Mark Zuckerberg is also singing a different tune - recently saying that if employees can boost productivity faster than the rise of automation (lol ok), "in theory there should be more jobs in the future, not less."
Even Ford has hired several hundred engineers that they attributed to concerns over the quality of work that had been automated - with one spokesman telling the Journal: "Engineers with deep technical expertise leveraging the power of AI is a powerful combination that is driving quality gains at Ford."
And so, as the Journal notes, "Collectively, the narrative has shifted from worker-light doomsday scenarios caused by AI to a future in which workers keep their jobs - and get a productivity boost."
The sentiment change isn’t limited to tech leaders: A survey by EY-Parthenon found that the percentage of CEOs who believe AI investments will result in significant reductions in head count fell from around 46% in January 2025 to 20% this May.
Last month a California think tank found that there has been "no evidence of a surge in AI-related layoffs" outside of one group of people; college-educated workers from industries with high exposure to AI - unemployment claims showed an increase after the 2022 launch of ChatGPT-3.5 and have remained high through May 2026.
Giant Needle SkipJust as we were seemingly hurtling headfirst into the AI jobpocalypse, major companies began blowing their budgets on tokens - with one Anthropic enterprise client (thought to be Amazon) spending $500 million in Claude charges in a single month.
In April, Uber's CTO went on the record saying the company had burned through its entire 2026 Claude Code budget in four months, leading to a hard cap of $1500 per month, per employee, peragentic coding tool. In mid-May, Microsoft began winding down internal Claude Code access. And then there's Amazon, which encouraged employees to spend AI tokens via a leaderboard, only to shut it down months later.
Related:
Meanwhile, UBS found last month that less than 20 cents of each dollar spent on AI tokens reaches real users.
Breakdown of every $1 spent on AI tokens: less than 20cents reaches real users (44 cents is spent fixing bugs generated by other AI) https://t.co/jFzUVSCPos pic.twitter.com/NwOWg7HppA
— zerohedge (@zerohedge) June 22, 2026According to a survey of corporate executives by technology and management-consulting firm Emergn, around 20% say the AI deployment reports they're getting paint a rosier picture than the facts on the ground - with some saying that their staff are obscuring failures.
"They may have noticed that the labor market is genuinely not changing (i.e., imploding) as rapidly as they expected," David Autor, a professor of economics at the Massachusetts Institute of Technology, told the WSJ. "They may have realized it was simply bad business to say that your great new product will destroy the economy."
Another study by financial-technology company Ramp and workforce-intelligence firm Revelio Labs supports the notion that AI adoption has been positive for job growth - finding that companies making the largest AI investments grew employment by roughly 10% over otherwise similar companies that hadn't yet adopted the technology.
TL;DRWith pitchforks out across America over data centers, AI CEOs have changed their tune - and are now saying their earlier warnings of mass unemployment were greatly exaggerated.
Tyler Durden Mon, 07/06/2026 - 17:20