Photo by Maksym Kaharlytskyi
Data centers are notorious energy hogs. With each hyperscale data center using as much energy as an entire city, they’re polluting surrounding areas, straining grids, and raising electric rates nationwide.
This is creating a backlash, with 70 percent of Americans saying they oppose building a data center near them.
Companies have proposed a solution that sounds good on the surface: data centers bringing their own power. If they use their own “behind the meter” energy, tech companies can get data centers running faster without increasing transmission or generation costs for other consumers.
But these dedicated power sources almost always run on natural gas — a polluting fossil fuel. Behind the meter natural gas is projected to account for 40 percent of all data centers’ power by 2030.
Amazon recently announced a data center in Texas that will be powered by its own natural gas plant. The facility could emit up to 33 million tons of carbon dioxide a year, making it the most polluting power plant in the country.
Meta also built a massive data center in New Albany, Ohio powered by 400MW of dedicated natural gas turbines — enough energy to power over 300,000 homes.
And since there’s a shortage of gas turbines for power plants, tech companies are turning to any gas turbines they can get, many of which are far less efficient and more polluting — and often paired with backup diesel generators, which are even worse.
A study commissioned by the Piedmont Environmental Council in Virginia highlights the hazards communities face. It found that a single facility, the VA2 data center in Loudoun County, could inflict $53 million to $99 million per year in health-related damages in the surrounding community from increased deaths and hospitalizations.
Despite these alarming findings, most approved and proposed gas turbine plants to power data centers are even larger.
Just imagine what that means for pollution and health impacts nationwide — particularly in states like Texas and Ohio, which are fast-tracking these plants. (Texas recently issued a moratorium on data centers connected to the grid — but behind the meter facilities are exempt. That’s a massive loophole for superpolluters like Amazon!)
Picture the damage to our air, water, and land from increased natural gas fracking, pipelines, and methane leaks. Or the more extreme storms, wildfires, and droughts from climate change caused by burning fossil fuels.
On top of that, consumers will still pay more for energy. The increased demand for natural gas from these behind the meter facilities will likely drive up the price of natural gas for all households.
But there’s a better way to power AI data centers.
China, the leading U.S. competitor on AI, is increasingly powering its data centers with renewable energy instead of fossil fuels. While there are legitimate concerns around China’s AI policies overall, the country is planning to power these centers with at least 80 percent clean energy, even while doubling their energy use, by 2030.
Big Tech is shelling out hundreds of billions of dollars on data center infrastructure. They should spend some of those billions to drive the use of renewable energy, pairing wind and solar with battery storage to provide 24/7 energy.
They know this can be done because they’re already doing it. Google is building a data center in Minnesota powered by 1.6 GW of solar and wind with 300 MW of battery storage. And Amazon is building a 1.2 GW solar and battery storage facility in Oregon to power a data center.
In states that allow for virtual power plants — distributed networks of home batteries, rooftop solar panels, electric vehicle chargers that are coordinated via software to act like a power plant — companies could also purchase power from residentsand cover the cost of people installing their own solar and battery storage, drawing any excess generated power to meet peak demand.
For the sake of our communities, the United States must be a leader in powering AI with renewable energy.
The Money is Still Not Showing Up for the Big AI Companies
Still from Matrix Revolutions.
This week’s numbers…
When I saw this note from Torsten Slok, the chief economist for Apollo Capital, I knew I had my topic for the week. The point is that the big money in AI is far removed from the end product. The chipmakers are making money hand over fist, the energy providers are doing okay, the hyperscalers have less to show, and the AI companies are losing bucks bigtime.
This matters because at the end of the day if the AI companies are not making money, the whole thing breaks down. To use a common analogy, suppose that steel companies are making huge bucks producing steel for rails, and construction companies are making money laying the rail, but the companies that run the railroads are all going broke. That doesn’t look like a story of long-term prosperity. In the great-minds-think-alike category, Ed Zitron jumped on the same point in his excellent newsletter.
Anyhow, I take a somewhat different tack than Ed and focus on the Chinese competition. I realize that even if there was no competition from China, it is unlikely that AI would ever have the massive payoffs the hyperscalers are banking on — but the existence of that competition makes the story considerably less likely. And developments in the last couple of weeks seem to make the case for American AI even weaker.
Chinese AI Is Cheap and Getting Cheaper — US AI Less So
As I have frequently noted here in the past, Chinese AI costs far less per input or output token than US AI. For the cutting-edge models, the Chinese AI sells for one-fifth or even one-tenth the price of US AI. One response I have seen is that even though the Chinese AI costs less per token, it can still end up being more costly because the systems are less efficient and require more tokens per task.
I am not sure that the measure of cost per task is a sufficiently standardized metric to allow it to be compared in a meaningful way, but insofar as it can be, it looks like the US advantage has gone away. According to the Korean electronics industry publication, The Elec, the leading Chinese AI model is now cheaper on cost per task than the leading US model, and performance gaps continue to narrow.
In the same vein, both Google and DeepSeek released new flash models last week. The DeepSeek model scored better on several benchmarks — and it sells for less than one-tenth the price.
If that makes the picture look bleak for US AI producers, don’t worry — it will likely get worse. Alibaba reports having developed a modular design that will allow it to build data centers in 100 days, compared to 12-18 months in the United States. This should mean lower costs and greater capacity for Chinese AI producers. That means the flood of low-cost high-quality Chinese AI is likely to get even larger in the months ahead.
Chinese AI is Finding New Customers
Given its huge cost advantage, it’s not surprising that Chinese AI models are gaining ground rapidly at the expense of US models. I’ve noted before that Chinese AI seems to be winning out by large margins in most regions of the developing world; however, it also seems to be gaining ground in Europe. There are political considerations that could make European companies reluctant to rely on Chinese AI; however, given the erratic behavior of Donald Trump, it’s not clear going with the US provides greater security.
And it looks like Chinese AI is continuing to gain ground in the US market. It seemsthat Apple is looking to Chinese AI as a cheaper alternative to the Silicon Valley producers. Apple by itself is potentially a huge market, but perhaps more importantly it is a company that has been at the cutting-edge of innovative technology for more than a quarter century. Its decision to go with Chinese AI is sending a serious message.
And remember, the question for those expecting really big bucks for the AI makers is not just whether Anthropic, OpenAI, and the rest can hang onto a large share of the market. It’s whether they can do so while selling at prices that give them the huge profits the stock market is banking on.
Can Creative Financing Overcome the Problems?
As mortgage issuers sold ever more dubious mortgages to further inflate the housing bubble, the wizards of Wall Street assured us that their financial magic would make it all work. This attitude was best conveyed by former Treasury Secretary Larry Summers at an academic conference in 2005, where he dubbed a critic of the growing house of cards as a “financial luddite.” Somehow, Summers thought innovative finance would make the millions of underwater mortgages issued to people with weak employment prospects and no reserve assets all work out fine.
We might be getting the same story with the AI bubble. Getting back to Torsten Slok’s point about the chipmakers making big bucks — while AI producers are making big losses, it seems Nvidia is looking to address the problem. It has justarranged $500 billion in financing from major banks for the hyperscalers that buy its chips. Fans of markets everywhere are asking the obvious question: If there is so much money to be made in building the data centers, why does Nvidia have to arrange the financing?
The details are not clear at this point, like whether Nvidia will in any way be on the hook for the financing, but there is a suggestion that it could involve securitization with tranches carrying different levels of risk, sort of like mortgage-backed securities or collateralized debt obligations. It could be lots of fun!
This first appeared on Dean Baker’s AI Monitor blog.
No comments:
Post a Comment