I met Andy Lee Robinson more than a few years ago, online that is. Last I heard he was in England. I do remember I paid him in British pounds sterling [£] for the lifetime privilege of using this image. This graph begins in 1979. Pertinent in comprehending this is realizing several points.

First, sea ice forms on the ocean surface, with contributions both from the salty ocean surface + fresh snowfall. Separate from the ice floes calving off tide-water glaciers.

Second, April—the light green line—is the maximum extent of sea ice, the end of Arctic winter of course.

Third, September—the black line—is the minimum extent of sea ice, in this case coincident with the end of Arctic summer.

The pictures you have often seen of the fluctuating surface area of Arctic sea ice represent just that, while this graph is a measure of total volume, more pertinent in understanding climate effects. When + where sea ice disappears, more dark blue ocean water absorbs far more solar energy, which is a main reason that the Arctic is warming up faster than the world as a whole.

It should be apparent to the naked eye where all of this is headed, with huge implications for geoscience, fishing, extinctions, navigation, territorial claims, seafloor mining, extraction of fossil fuels—and the survival of whales. Yes, whales, since with dwindling sea ice, pods of orcas have migrated north to predate on Arctic whales such as belugas + narwhales.

I post this every couple of years, + suggest somber contemplation. Especially if you work in a fossil fuel industry.

Deciding on an energy strategy—whether for a company or an institution—depends on simulating future prices for the energy sources under consideration.

In this context, it is crucial to evaluate the pricing structures and/or tariffs available for each source.

Equally important are the risk management tools that can be employed to ensure the final monthly cost of the chosen source(s) aligns with the established business plan.

One thing is certain: those who tackle this challenge head-on are more likely to succeed than those who simply take a "wait-and-see" approach.

Recently, I wanted to see a post I had written in 2015. (Some background: I started the Blogspot blog in January 2013. I transferred all my Blogspot posts - over 1200 then - to Substack last August and stopped posting any new posts on Blogspot last fall. In fact, as you’ll see if you go to the link at the beginning of this post, I will soon take down the Blogspot blog because it’s become too popular among AI models eagerly seeking information in posts I wrote years ago. In fact, my Blogspot blog had 672,000 pageviews last month alone, while I guess I receive no more than 10-15,000 pageviews per month in Substack and in Energy Central, where I also put up my posts. I believe the Blogspot pageviews are almost entirely from LLM training)

When I entered “2015” in the Substack search bar, I wasn’t terribly surprised that, while I immediately was shown every post where “2015” appeared in the text, the search doesn’t look at the date of the post itself, and therefore couldn’t show me which posts I wrote in 2015 (or any other date range). I asked the chatbot if it was possible to search on a date range. It cheerfully pointed out to me (it’s always cheerful, of course) that the feature I want isn’t supported in the search bar, but it would be happy to submit a new feature request for me. Even though I knew this would be useless, I let it do that anyway. Then it closed the ticket.

After that, I received the usual email asking me to clarify what I wanted, so I did that – again, not expecting anything more to come of it. However, this time a human emailed me to point out that there’s a completely different way – not involving the search bar – that I can see all of my posts within any date range. It works fine, so now I can find posts from 2015 or any other time period.

The important point about this is that I don’t think the person who solved my problem is some sort of brilliant individual that Substack is lucky to employ. They just did what a human who’s trying to be helpful does naturally (or should, anyway): go back to the problem I was trying to solve and figure out how it could be solved in the context of the Substack platform. The chatbot had categorized my problem as narrowly related to the search bar (even though it was formulated more generally), but my problem was really that I wanted to see posts from a range of dates – and I didn’t know the platform well enough to realize that the search bar isn’t the only possible way to do that. My guess is the human figured out the answer to my problem within a few seconds, even though the chatbot would probably never have done so without being given additional prompts that a human wouldn’t need.

The moral of this story is that generative AI, which is simply an elaborate statistical algorithm for predicting the next word in a sentence, will never reach the point where it can satisfactorily answer a question, if the questioner didn’t ask their question in a way that the answer will be straightforward – specifically, if the questioner didn’t ask the question in the context of the correct response. I asked my question in the context of the search bar, so that was all the chatbot focused on in its answer. But the human understood the question in the right context: “Given that I’m a Substack user, is there a way for me to find all of my posts within a particular date range?”

Of course, since this was Substack’s chatbot, it could have been trained always to consider every question that appears to be about a particular function within Substack to be instead a question about the platform as a whole. That might have allowed it to solve my problem without requiring human intervention, but what if I’d really been asking whether blogging platforms in general allow searches by posting date? That would have required an even bigger context, which the chatbot probably wouldn’t have been trained for.

Human beings don’t know all the answers, but they should at least be able to reformulate the question in its proper context. Generative AI can’t even do that and probably never will. 

Tom Alrich’s Blog, too is a reader-supported publication. You can view new posts for two months after they come out by becoming a free subscriber. You can also access all of my 1300 existing posts dating back to 2013, as well as support my work, by becoming a paid subscriber for $30 for one year (and if you feel so inclined, you can donate more than that or become a founding subscriber for $100). Whether free or paid, please subscribe. 

If you would like to comment on what you have read here, I would love to hear from you. Please comment in my chat or email me at tom@tomalrich.com.

.

Brazil has the largest electricity sector in Latin America, + in 2024, Brazil added a substantial 10.9 GW of new power generation capacity, with a total installed capacity of 209 GW, of which nearly 85% was renewable. [Recall that a gigawatt or GW is equivalent to the instantaneous or power rating of a typical nuclear power plant.]. "The installed capacity grew from 11,000 MW in 1970 with an average yearly growth of 5.8% per year."

Impressively, Brazil has the largest capacity for water storage in the world, based largely on hydroelectricity generation capacity, which meets over 60% of its electricity demand. "The national grid runs at 60 Hz, just like in the US, and is powered 83% from renewable sources." This dependence on hydropower makes Brazil vulnerable to power supply shortages in drought years, as was demonstrated by the 2001–2002 energy crisis.

"In 2023, the output of Brazil's electricity system, serving over 88 million consumers, exceeded that of all other South American nations combined." Anticipated investments surpassing $100 billion by 2029 aim to expand utility-scale + distributed generation, alongside transmission + distribution projects. "The National Interconnected System (SIN) comprises the electricity companies in the South, South-East, Center-West, North-East and part of the North region." Only 3.4% of the country's electricity production is located outside the SIN, in small isolated systems located mainly in the Amazonian region.

Fascinating country in so many ways. From my background in infectious disease, problems with zika, dengue fever, malaria. But the specter of drought hangs over the Amazon + Brazil's power sector, as in many areas of the world. Since Brazil is about 92% tropical, it seems to me they should diversify further into solar + storage.

From half-meter-long moths to fairy wasps smaller than sand grains, insects come in a stunning variety of shapes and sizes and constitute the most diverse animal group on Earth. "But the insect species discovered so far may represent just a fraction of the total crawling, flying, and burrowing around the planet, according to a new study published today in the Proceedings of the National Academy of Sciences." Using statistical methods borrowed from epidemiologists, a team of entomologists estimates there may be as many as 20 million insect species on our planet—more than three times the previous estimate.

"Over the past 3 centuries, biologists have described about 1 million insect species, but finding and describing them all would be a daunting—if not impossible—task." However, a subfamily of parasitoid wasps known as Microgastrinae, which infamously lay their eggs inside living caterpillars, are extremely well-studied. "Over the past several years, scientists conducting surveys of flying insects in the park have identified 388 species of Microgastrinae."

Independently, when scientists surveyed caterpillars that had been parasitized within the park, they identified only 889 wasp species. With almost no overlap, the mismatch between the 2 studies allowed a statistical estimate of a whopping 2394 Microgastrinae species. Applying this as a multiple to all 53,945 known insect species within Guanacaste suggests the park is actually home to 332,846 insect species, most of which have gone unobserved.

"The researchers then scaled this number globally using another diverse group of organisms: trees...[with] 1200 [to] 1500 tree species within Guanacaste and about 73,000 on Earth, meaning the park contains between 1.6% and 2.1% of global tree diversity." If the same percentage holds true for insects, then there’s anywhere between 13.3 million and 24.7 million insect species on Earth, with a safe middle-of-the-road estimate of 20.3 million species.

While you may think this epidemiological calculation represents biodiversity run riot, recall that many insect species are in decline from pesticides + habitat loss + climate change. Do not be sanguine.

Fuel cells, a proven way to make electricity without combustion, fission, or gravity, are making serious inroads into the red-hot market for producing power for artificial intelligence data centers.

Last Tuesday (June 30), fuel cell vendor Bloom Energy (NYSE:BE) announced what they described as a “$25 billion” deal with Canadian financial holding company Brookfield (NYSE,TSN:BN) that is a five-fold increase in Brookfield’s initial investment last October in the San Jose-based solid-oxide fuel cell developer. In a news release, Bloom said, “The expanded partnership reflects strong and sustained demand from hyperscalers and AI infrastructure developers for fast, reliable, and community-friendly power.” 

Aman Joshi, Bloom’s chief commercial officer, said, “When we formed this partnership, we said it was the first phase of a much larger vision. Today’s commitment reflects the momentum we are seeing in the market, as evidenced by recently announced large-scale deals. Bloom is uniquely positioned to address the urgent need for clean, reliable power to support the rapid growth of AI.”

Brookfield, with a trillion-dollar diversified portfolio that includes real estate, infrastructure, renewable energy, private equity, and insurance, is one of the world’s largest financial holding companies. In late 2022, it spun off private equity investor Brookfield Asset Management (NYSE,TSX:BAM), which relocated to New York City in 2024.

The Bloom-Brookfield compact was the second major AI coup for Bloom. In April, Texas-based technology giant Oracle Corp. announced a deal with the company to develop a behind-the-meter 2.45-GW artificial intelligence  data center in the New Mexico desert, dubbed Project Juniper, powered by Bloom’s fuel cells.

 At the time, Oracle explained its choice of fuel cell technology: “Fuel cells generate electricity without combustion, meaning the Bloom microgrid is highly efficient with low emissions and water use. Compared to its previously planned gas turbines, Project Jupiter with the Bloom microgrid will reduce NOₓ emissions by approximately 92% and will use a negligible amount of water.” 

A week before the Bloom-Brookfield deal (June 24), Connecticut-based FuelCell Energy (Nasdaq: FCEL) and Fit Energy USA LP of Boca Raton, Fla., a “Foreign Limited Partnership” formed in February under Florida law, announced a “strategic agreement for up to 380 megawatts (MW) of clean, baseload on-site power for data centers using FuelCell Energy’s utility-scale fuel cell technology. The agreement includes an immediate deposit for an initial 30 MW of power scheduled to begin delivery later this year.”

Jason Few, FuelCell Energy president and CEO, said, “This agreement further validates our decision to scale our operations to 500 MW, preserving our ability to serve a broad and growing pipeline of customers.” Fit Energy CEO Joel Leonoff, said, “Today’s announcement marks a critical step in building the power foundation required for the next generation of AI infrastructure. FuelCell Energy’s technology aligns with our growth objectives and our goal of delivering behind-the-meter power solutions to data centers at gigawatt scale.”

According to a joint news release, the two companies said that “under the arrangement, Fit Energy will be eligible to receive warrants tied to future deployment milestones of up to 380 MW. The warrant structure is designed to align long-term value creation with successful project execution and customer deployment.”

Fuel cells generate electricity through electrochemical reactions between a fuel (often pure hydrogen) and an oxidizing agent (typically oxygen) without combustion. As the Department of Energy describes it, “A fuel cell consists of two electrodes—a negative electrode (or anode) and a positive electrode (or cathode)—sandwiched around an electrolyte. A fuel, such as hydrogen, is fed to the anode, and air is fed to the cathode….The electrons go through an external circuit, creating a flow of electricity.”

In the middle of the new fuel cell announcements, Norwegian energy consulting firm Rystad Energy issued a very bullish report on the fuel cell-data center connection, predicting “a tenfold increase in fuel cell market revenues by 2030, rising from around $2.8 billion in 2025 to roughly $30 billion, as AI computing demand drives unprecedented growth in data center construction.”

Backing up that prediction, says Rystad, is a “contracted order book of approximately 9 gigawatts (GW), including framework agreements with Oracle, AEP, Equinix, and Brookfield, points to growing confidence among major operators in fuel cells as a viable long-term power source.”

Fuel cells are particularly well suited for what the consultants see as a decided move to “on-site power generation rather than grid connection. Unlike conventional grid connections or large gas plants, fuel cells can be deployed quickly and run on natural gas today, transitioning to biogas, renewable natural gas or hydrogen as supply matures, while producing lower on-site emissions than combustion alternatives.”

Rystad’s Lein Mann Bergsmark said, “Power availability has become one of the defining constraints on data center growth, and operators are increasingly looking beyond the grid for solutions. Fuel cells have moved from a niche application to a measurable part of the firm power mix. The question now is whether the supply chain can scale at the same pace as demand.”

The Quad Report, covering energy policy and politics

$/kWh refers to the cost of all consumed energy sources normalised to a single unit.

kWh/activity represents energy consumption relative to a specific operational measure, such as tons delivered, units produced, or—in the case of buildings—the number of occupants.

By tracking these two metrics month-by-month, one can gauge the performance of energy managers on two fronts: the cost of energy procurement and energy efficiency relative to output!

Best of all, monitoring a chart is easy, simple, and quick for everyone involved with energy management—an area that typically ranks among the top 10 (and often the top 5) cost categories.

A new framework from the International Institute for Sustainable Development finds that solarising Indian villages can slash electricity costs well below state benchmarks, but only if utilities plan storage, surplus power and grid integration together rather than chasing rooftop installation targets village by village.

👉 Read the full story: https://indoen.com/news/rural-indias-solar-reckoning-why-village-grids-need-more-than-panels-to-work

But a quiet infrastructure success story unfolding in rural Haiti turns this capital-intensive, top-down philosophy entirely on its head. It suggests that in highly volatile or capital-constrained environments, the ultimate grid architecture might not be a massive upgrade at all—but rather a modular network of "nickel-and-dime" mesh grids.

Bypassing the Centralized Bottleneck

In my recent coverage for Forbes--picked up by Yahoo Finance--I looked closely at how decentralized solar mesh grids are successfully delivering clean, affordable electricity to thousands of rural Haitian households. These are not mini-utilities or traditional microgrids that require centralized generation and a mini-distribution network. Instead, they connect small clusters of homes via localized, modular nodes. If one node or home drops off, the rest of the mesh network self-heals and keeps running.

For utility professionals, the structural takeaways from this model are profound:

• Radical De-Risking: Traditional grid extensions require massive upfront capital expenditure before the first customer flips a switch. Mesh grids scale organically. You build out a few homes at a time, matching capital deployment precisely with localized demand.

• Operational Resilience Under Strain: Conventional wisdom says infrastructure requires political stability. Mesh networks prove that localized, modular assets can thrive in environments of absolute instability precisely because they have no single point of failure. If it can maintain reliability in rural Haiti, the underlying architecture has cleared the ultimate stress test.

• The Capital Sequencing Blueprint: Perhaps the most scalable lesson is how these projects are funded. Developers utilized early-stage philanthropy to absorb the initial execution risk. Once the operational data proved the model's viability, it unlocked major follow-on funding from multilateral institutions like the World Bank and the IDB Lab.

The Macro Value of Micro-Networks

As global utilities face mounting challenges from extreme weather, physical security threats, and skyrocketing interconnection queues, the "all-or-nothing" approach to grid expansion is looking increasingly fragile.

Haiti’s mesh grid success isn't just a humanitarian milestone; it’s a technical proof of concept. It proves that a highly decentralized, modular power architecture can deliver 40% lower costs and unmatched resilience. For an industry staring down the barrel of a complex energy transition, it’s time to realize that sometimes, thinking small is the most strategic way to scale.

https://www.forbes.com/sites/kensilverstein/2026/07/01/despite-crippling-poverty-haiti-is-quietly-switching-the-lights-on/?ss=energy