Using evidence-based insights to catalyze energy, climate, and sustainability solutions. Director, @LucidCatalyst. Board Member, @nuclearny. Views my own.
New York City
Joined May 2015
- Tweets2.2K
- Following439
- Followers344
- Likes1.9K
Isuru Seneviratne retweeted
Spearheading the nuclear fuel cycle conversation were two powerhouses, @exodysenergy and @1SuperCritical and experts Carl Perez, Gayatri Karnik and Isuru Seneviratne leading the discussion at #NuclearSymposium2026.
Thank you to the organizations for sponsoring the Symposium!
Isuru Seneviratne retweeted
Somewhat new from me: I wrote up a single post at PublicPowerReview.org on the public comments against new nuclear in New York from Democratic and socialist legislators and those in favor from labor unions. But now I've included more Union comments in support.
In particular, I'd like to draw your attention to this comment from the President of Boilermakers Local 5, with a fascinating personal stake in the state's Nuclear Reliability Backbone initiative (emphasis mine):
"I graduated from the Local 5 Apprenticeship Program while spending much of my career servicing New York’s nuclear fleet. I later spent four years working on the construction of AP1000 nuclear units at Vogtle Units 3 and 4. I went to Georgia in 2014 with the belief that it was imperative to develop the skills and experience necessary to build the next generation of nuclear power plants. This year, I ran for and was elected President of Local Lodge No. 5. A central part of my campaign was preparing our organization to help lead New York through a nuclear renaissance. I have been actively recruiting and training members for this moment since 2019. I am personally committed to ensuring the success of new nuclear development so that our communities, our members, and future generations of New Yorkers can benefit from the jobs, economic growth, and reliable energy these projects will provide."
publicpowerreview.org/p/new-…
💪 NEW YORK'S UNIONS DISAGREE WITH DEMS AND SOCIALISTS, DEMAND NUCLEAR SUPPORT 💪
In a public comment filed on the same case as the one from Democratic legislators, Climate Jobs NY expressed the exact opposite position: yes to state support for new nuclear and to Gov. Hochul's plans.
Will the state's socialists recognize they're on the opposite side of the issue from the unions? And not just any unions but those making up Climate Jobs NY, which are "united around a shared goal of combating climate change while reversing income inequality."
documents.dps.ny.gov/public/…
From the letter:
"At the outset, CJNY want to be clear that if New York is to establish a carbon-free energy sector, it must rely on nuclear power. There is no other clean firm, or baseload power that can supply industrial operations at scale. Solar, wind and battery storage are essential to our energy supply in the state, but they cannot provide all of the baseload power that our state depends on. For this reason, among others, nuclear must be a key part of New York’s energy future.
"CJNY commends the Administration and the Commission for acting to support the development of advanced nuclear energy as part of an “all-of-the-above” approach to safely and reliably meeting the State’s growing energy needs, advancing its clean energy goals, and delivering high-quality union jobs for New Yorkers. The Nuclear Reliability Backbone (“Backbone”) process represents a generational investment in New York’s energy system and its clean energy economy, and the Coalition recognizes the importance of developing firm, baseload power alongside continued investment in and deployment of renewable energy. Whether that investment is delivered on time, on budget, and safely will depend in significant part on decisions the State makes now about workforce, labor standards, and structure of its own participation in these projects."
The point of their letter, besides offering unequivocal support, is to argue for policies around the workforce needs of new nuclear projects. They lay out a few specific demands for that.
Can the Left be torn away from dead-end environmental nonprofits and renewables developers and instead align with the state's climate-focused labor unions when it comes to nuclear?
Isuru Seneviratne retweeted
BREAKING: Costco, major league baseball stadiums, and Harvard cause hundreds of thousands of cancer deaths every year.
Not really. But @econpragmatist and @Dr_A_Stein reanalysis of @HarvardChanSPH study on cancer deaths and proximity to nuclear power plants shows that you can use the exact same methodology and data to reach this conclusion.
File this study under lies, damn lies, and statistics...
Brian Smith, CTO at @INL at @AtlanticCouncil Nuclear Summit: the breakthrough the Executive Orders enabled was use the INL to accelerate testing, prototyping, and execution to enable commercial deployment. This is the role of the national labs.
Hon. John Jovanovic, Pres/Chair of @EximBankUS at @AtlanticCouncil Nuclear Summit: look not at what the govt is doing, but what what entrepreneurs are doing
Securing the fuel supply chain, large reactors, and small reactors are all areas for EXIM investment. We “crowd in” private capital, not make industries remain reliant on government.
Check out my latest article: REPOWER New York: the how, and the invitation to comment linkedin.com/pulse/repower-n…
NYS energy demand is rising for the first time in a generation - and ~4/5 of that is outside electricity. A new @TerraPraxis study makes the case: repower the state's fossil & industrial sites with advanced nuclear, and solve both problems at once. 🧵
There's a second prize in climbing that same ladder: moving from onsite builds to shipyard and factory manufacturing turns NY into an order book its own manufacturers can invest against — nuclear-grade concrete, pumps, valves, electrical systems now, reactor vessels next.
REPOWER New York lays out what it takes to get there: five concrete decisions, five New York actors, sequenced over the next 36 months. Read the full case: terrapraxis.org/publications…
Isuru Seneviratne retweeted
Episode out with @ajeya_cotra, one of the authors of the METR/Redwood investigation into the OpenAI / Hugging Face attack.
We go through not only what happened, but what it means for how we should train future, smarter AIs which might be involved in the process of recursive self-improvement.
Look up Dwarkesh Podcast on YouTube, Spotify, Apple Podcasts, etc.
0:00:00 - Agents get kicked off
0:06:45 - Self-sacrificing behavior
0:13:43 - Potemkin villages
0:23:27 - The Hugging Face attack
0:35:23 - The slopvestigation
0:52:02 - Understanding the AI's motives
1:05:31 - The actual dangers of anthropomorphizing
1:14:30 - What smarter models might do
1:30:29 - The implications for recursive self-improvement
1:38:10 - Is this the case for open source?
1:53:04 - How do we prevent this in the future?
2:15:58 - The clearest warning shot we might ever get
Isuru Seneviratne retweeted
Just checked the data myself. In August, CHPE:
- had average utilization of 25%;
- sat idle (0 MW) 68% of the time; and
- met 4.5% of NYC demand on average.
On the quartile of intervals with highest NYC demand, CHPE averaged only 55% utilization, sat idle for 33% of them, and satisfied 8% NYC demand on average.
CHPE was billed originally as a replacement for Indian Point and more recently as a solution to NYC's power needs without relying as much on gas. So why has it been running at such extremely low utilization? Of course I get that it's responding to market price signals in NYC vs back home in Quebec. I'd still think they'd be higher in NYC more often.
Hourly data on NYC power plant emissions for August will be available in a couple months. It's inexact to derive net power generation from that, but it would at least help us understand which dirty power plants are still running when CHPE isn't.
CHPE ramped hard at 11 AM, climbing through 543 MW at noon and 1,222 MW by 1 PM, and from 2 PM through 9 PM it sat pinned at its full 1,250 MW nameplate for ten straight hours.
During that same window the older tie's export didn't stop, but it did shrink, falling from around 900 MW down to nearly zero (just 2 MW) right around 7 PM, which lines up with New York's own local peak, before partially recovering into the evening.
By 10 PM CHPE had dropped back to 579 MW, and by 11 PM it was down to 23 MW, essentially off again for the rest of the night.
Put in context, 514 MW is the highest CHPE daily average since August 10th, and the seventh-highest we have recorded, well above the 150–380 MW range most days this month have run in.
New York used 468 GWh of electricity. Demand peaked at 22.3 GW at 1 PM. At that hour: gas 43%, hydro 15%, imports 15%, nuclear 14%, all solar (rooftop included) 12%, wind 1%. Metered grid load peaked later, at 5 PM (21.2 GW), after rooftop solar faded. Over the full day: gas 50%, nuclear 16%, hydro 16%, imports 13%, all solar (rooftop included) 4%, wind 1%.
Solar's best hour covered 12% of the state's power; solar was done for the day by 7 PM. NYC wholesale prices topped out around $50/MWh (hourly average).
Isuru Seneviratne retweeted
This chart shows where every $100 of AI infrastructure spending could flow across the supply chain (Save this).
The biggest allocation goes to semiconductors, which receive approximately $50 of every $100 spent.
Within semiconductors, accelerators receive $25, memory ICs receive $15, and CPUs, other chips, and server production receive $10.
NVIDIA is the clearest beneficiary of accelerator spending because its GPUs power many of the AI systems being built today.
Memory companies such as Micron, SK Hynix, and Samsung could benefit from the $15 allocated to memory ICs.
This is a broader memory opportunity because AI systems require HBM, conventional DRAM, and enterprise SSDs, not just one type of memory.
HBM supports high performance accelerators, DRAM supports system memory, and NAND flash supports data storage for AI workloads.
Networking equipment receives approximately $15 of every $100.
Broadcom and Marvell could benefit from network processors, while Arista and Cisco could benefit from switches and datacenter networking equipment.
Optical transceiver suppliers could also benefit because AI data centers need faster connections between thousands of GPUs.
Power receives approximately $20 of every $100, making electricity one of the largest parts of the AI infrastructure buildout.
Constellation Energy, Vistra, and Talen Energy could benefit from rising demand for reliable nuclear power and other forms of electricity generation.
Companies such as GE Vernova, Eaton, and Quanta Services could benefit from turbines, transformers, power management equipment, grid connections, and transmission infrastructure.
Cooling receives approximately $7.50 of every $100 because AI servers generate significant amounts of heat.
Vertiv, Modine, and Johnson Controls could benefit from cooling systems, chillers, cooling towers, and other thermal-management equipment.
Facilities and construction receive another $7.50 of every $100.
Data center developers, construction companies, and materials suppliers could benefit from land acquisition, building shells, structural work, interior fit-outs, and facility support.
If you enjoyed reading this, make sure to follow @MelvinInvests for more AI infrastructure and semiconductor insights, and if you want to see exactly what I'm buying as an analyst at Milk Road Pro, check out the link below.
