The 890 MW signature: 20 years, 11 units, one grid
Google and Constellation Energy committed to bring 890 MW of brand-new nuclear capacity onto the PJM grid under a 20-year purchase agreement announced on October 6, 2026. The plan rests on uprates at 11 existing reactors across Illinois, Pennsylvania and New Jersey, modernizing turbines, steam generators and digital controls, with the first added capacity targeted for 2028. According to ConstellationEnergy, the investment exceeds $4.3 billion , protecting 4,400 existing jobs and creating 7,200 construction roles. CNBC reports this slice is about a quarter of the total package on America's largest power grid.
The quieter half of the package is a 15-year supply agreement for another 2,700 MW, bringing the two legs to 3,590 MW combined. This second tranche is not tied to a single source; it gives ConstellationEnergy's operating plants long-term revenue certainty so they keep capacity in the PJM market. As the host notes on air, hyperscale operators are locked into a hunt for uninterrupted carbon-free electricity, and every extra MW squeezed from the existing fleet arrives faster than a greenfield plant. To me, the structure reads as a direct answer to the grid operator's bring-your-own-power pressure.
The Amazon front: 190 MW at Calvert Cliffs
A similar signature came from Amazon a week earlier: the 20-year agreement announced September 30, 2026 grows the Calvert Cliffs plant in Maryland with a 190 MW uprate and over $3 billion of investment into the 1,790 MW facility, with the new capacity due online between 2030 and 2032. Do not let the 690 MW figure from the broadcast confuse: the bulletin published on Nasdaq makes clear that 690 MW is the agreement's total power volume, while 190 MW is the genuinely new capacity joining the grid. Amazon's commitment also paves the revenue ground for a further 20-year license extension . Together, the two deals locked in 1,080 MW of nuclear growth in a single week.
Google's nuclear file does not end there: in a company blog post, its energy chief puts the new nuclear capacity unlocked for the US grid through uprates and restarts at over 1.5 GW . The eye-catching item is the plan with NextEraEnergy to restart Iowa's Duane Arnold plant; the 615 MW facility carries a 25-year purchase commitment, targets a return to production in the first quarter of 2029, and promises 400 direct jobs plus over $9 billion in economic benefit to Iowa. As CNBC recalls, Microsoft's Three Mile Island move belongs to the same wave. Upgrade agreements at Southern Co. units, mentioned by the host, extend the same frame.
The wall: turbine queues and transformer scarcity
Yet the wall in front of this appetite rises on the equipment side: data compiled by ModernPowerSystems shows buyers facing 2-3 year waits at gas-turbine makers, over five years for some frames, with GE Vernova's backlog nearing 100 GW in the first quarter of 2026. Reporting from TheNextWeb completes the picture: delivery of large power transformers has stretched to 128 weeks , and medium-voltage switchgear is spoken for into 2028 at some suppliers. Pulling pipeline, finding turbines, queuing years for a grid hookup, all while AI clusters demand ever more power: that is the picture the host sums up as waiting years. Next to Meta's single 5 GW campus plans, 890 MW looks modest, and that scale gap explains why.
Behind-the-meter setups sound like salvation, but MilkenInstitute researchers warn they are a temporary pause: the fine print of hyperscaler contracts targets grid connection in 5-10 years, so the load is deferred, not removed. Moreover, Alphabet's absorption of Intersect in March 2026 shows giants shifting from buying energy to producing it. The Talen-Amazon deal examined by PowerMag teaches the opposite lesson: an $18 billion, 17-year package reaching 1,920 MW from the Susquehanna plant was structured inside the grid , so the load joins the shared system from day one. In my view, the coming years' real fight is whose bill these costs land on.
The market scale: why are the stocks lagging?
To the broadcast's headline question: why do utility stocks lag despite all these deals? The numbers speak: the S&P utilities index lost 3.2% in a single session and 14.4% in 30 days per WalletInvestor data, the year-to-date picture cited on air is also negative, and the Constellation stock sits 24% down. Two clamps hold it: first, rate sensitivity , as rising bond yields erode the appeal of dividend stocks; second, tariff caps, meaning companies cannot pass rising costs into bills and profit margins squeeze. As regulators press down to keep prices flat, the market prices these stocks at a discount. My read: deal headlines pop stocks, but a lasting rally needs looser rates and tariff flexibility.
| Deal | Scale |
|---|---|
| Google-Constellation | 890 MW new, 20 years |
| Extra supply leg | 2,700 MW, 15 years |
| Amazon-Calvert Cliffs | 190 MW new, 20 years |
Key moments
AI commentary
"My stance as narrator is plain: the megawatts impress, but the story lives in the stocks. While tech giants lock in energy, who pays the bill and when shares recover is the front worth watching."
AI assessment
The other side argues most of this is already priced into stocks: producers like Constellation lock their balance sheets with 20 years of guaranteed revenue, so the risk sits in growth pace, not consumers. Nuclear uprates also carry technical risk: each unit upgrade needs regulator sign-off and long maintenance outages, so the 2028 timetable can slip. I do not find these objections unfair, yet 3,590 MW of contracted volume removes the luxury of wait-and-see.
What the broadcast leaves open is the final address of the cost: grid reinforcements and behind-the-meter investments may ultimately land on bills, which is exactly why the tariff brake stays engaged. There is also the host side: for energy voices on a Yahoo Finance show, such deal news is part of the daily flow, so I read the gloomy framing as pulse-taking rather than exaggeration.
The practical takeaway for readers is sharp: when picking utility stocks, look less at reactor counts and more at tariff regimes and the rate path; deal announcements create short-term jumps, while regulatory geography decides lasting returns. These stocks belong in portfolios as dividend and ballast items, not growth bets. Before adding weight, I want to see the index's monthly slide stabilize and the first 2028 uprates actually connect.
Sources
11 links; no other published story cites them. Stories sharing a link do not confirm each other; a source's origin is not inferred from how often it is cited.
- @youtube.com YouTube — Yahoo Finance
- @constellationenergy.com Constellation Energy press release
- @cnbc.com CNBC power deal report
- @nasdaq.com Nasdaq Constellation Amazon agreement
- @investor.nexteraenergy.com NextEra Energy Google collaboration
- @blog.google Google nuclear plants blog
- @modernpowersystems.com Modern Power Systems turbine backlog
- @thenextweb.com The Next Web grid equipment scramble
- @milkeninstitute.org Milken Institute behind the meter
- @walletinvestor.com WalletInvestor utilities index
- @powermag.com PowerMag Talen Amazon PPA
nuclear energy · google · constellation · pjm · data center · utilities · stock market