AI data centers consume up to 4% of U.S. electricity annually, projected to rise to 12% by 2030. By 2027, AI's water demand for cooling could exceed Denmark's annual usage, causing logistical and public image issues.
Companies like NVIDIA are developing more energy-efficient processor chips, with NVIDIA's latest model being 25 times more efficient than its predecessor. Major AI sponsors like Amazon, Google, Meta, and Microsoft are also designing their own energy-efficient chips.
AI companies are exploring the use of non-potable water sources, such as wastewater from local septic systems or industrial discharge water, to cool their data centers instead of municipal water supplies.
Trump's administration could eliminate the $7,500 federal tax credit for EVs, which is crucial for startups with higher production costs. Additionally, grants and loans from the Department of Energy may be cut, and proposed tariffs could increase input costs for these already cash-strapped companies.
EV startups faced falling consumer demand, rising costs, and higher interest rates, leading to a series of bankruptcies. Many companies had set unrealistic growth expectations during the low-interest-rate COVID era, which they failed to meet, causing investor sell-offs.
Chinese EV companies like BYD and CATL are seen as leaders in the industry, offering superior products at lower costs. If the U.S. deprioritizes the EV sector, China's lead could grow, potentially leaving the U.S. behind in the global EV market.
You want a straightforward path to your goals, but at Merrill, we know things may get in the way.
Or new opportunities can put you at a crossroads. With the bull at your back, you get a personalized plan and a clear path forward. Go to ml.com slash bullish to learn more. Merrill, a Bank of America company. What would you like the power to do? Investing involves risk. Merrill Lynch Pierce Fenner & Smith Incorporated. Registered broker dealer. Registered investment advisor. Member SIPC. A wholly owned subsidiary of Bank of America Corp.
Welcome to Tech News Briefing. It's Friday, December 6th. I'm Danny Lewis for The Wall Street Journal. Artificial intelligence is hot, literally. The data centers AI companies rely on to train and run models like ChatGPT consume huge amounts of electricity and water to power their chips and cool them down. We'll hear what AI developers are doing to cut back on their energy use.
And then, it hasn't been a good time lately to be an electric vehicle startup thanks to rising costs, supply chain obstacles, and cooling customer demand. And now, with President-elect Donald Trump, an outspoken EV critic, returning to the White House, the situation could get worse for many young EV companies. WSJ reporter Amrith Ramkamar breaks down the state of the EV industry. But first, the AI industry makes a lot of promises for how the tech could transform everyday life.
but developing that tech takes a lot of resources. According to projections from consulting firm McKinsey, the data centers AI companies rely on use as much as 4% of total electricity in the U.S. annually, and that could rise to as high as 12% in 2030.
Meanwhile, researchers at the University of California, Riverside and the University of Texas at Arlington say by 2027, AI's demand for water globally to cool the chips in data centers could account for more than the annual demand of Denmark. That's all causing logistical and public image problems for the growing industry. WSJ contributor Bart Ziegler looked into what AI companies are doing to reduce their demand for resources. He spoke with my colleague Julie Chang.
But what is one way companies are looking to cut AI's demand for power and water? So they're looking at processor chips that will be more efficient and not need quite as much energy and not emit quite as much heat. NVIDIA, which is the biggest maker of these AI processor chips, says that its newest version will be about 25 times as energy efficient as the previous high-end model.
And meanwhile, some of the biggest producers of AI or sponsors of AI computing, including Amazon, Google, Meta, Microsoft, they're designing their own AI processing chips that they also believe will be more energy efficient. These systems consume a lot of water, specifically in the equipment used to cool data centers. How are some companies looking to address water use?
All the IAI companies are looking at how can they not use so much water and how can they, when they do use water, use water that's not drinking water. Amazon, for example, their data centers in Santa Clara, California, they're using water that comes from the local septic system. It's highly processed, but it's basically wastewater they use to help cool their computer centers rather than taking water from the municipal water supply plant.
Other uses companies are looking at are possibly taking water from rivers or water from industrial processes that discharge water. It's all an attempt not to waste potable water on AI computing.
One thing you wrote about that I found interesting was researchers looking at capping the amount of electricity used by AI computers. Tell us more about that. Yeah, that's a fascinating group of studies I came upon. They're actually discovering, this one group at MIT and Northeastern University, that they could cut the power transmitted to AI centers by 22 to 24%.
And all that did was expand the time to come up with a response from the AI by about 5% to 8%, which they think looks like a valuable way to handle some of the power that these centers use. They believe it could lead to significant reduction in energy consumption, it
Meanwhile, there's a group at several other universities that is looking at when they train the AI systems initially to vary the power instead of not having the power be on constantly. And they also think that could cut down power use fairly substantially. Is there anything they can do about data specifically since these models consume such vast amounts of data?
Yes, that's another area of research. They're looking at whether instead of these incredible large language models, they could cut them down and get the same quality of response. There are a number of studies going on in that area, and they're actually finding that if the data is more specialized or it removes redundancies or removes what they consider junk data, they can get even better responses to AI queries. And also these smaller databases use less energy.
So that's a really promising area. That was WSJ contributor Bart Ziegler speaking with our producer Julie Chang. Coming up, electric vehicle startups were struggling before President-elect Donald Trump won the election last month. What does his return to the White House mean for the industry? That's after the break. Say this is your financial life. Over time, things can get more complex. With a personalized plan...
Merrill can help you navigate it all. Learn more at ml.com slash bullish. Merrill, a Bank of America company. What would you like the power to do? Investing involves risk. Merrill Lynch, Pierce, Fenner & Smith, Inc., registered broker-dealer, registered investment advisor, member SIPC, a wholly owned subsidiary of Bank of America Corp. EV startups have not been having a great year.
High production costs and cooling demand have made the market tough for young companies. And several high-profile companies, including electric SUV maker Fisker and bus manufacturer Arrival, have filed for bankruptcy. And now that President-elect Donald Trump is a few weeks away from returning to the White House, things could get even harder for companies that once had high hopes for making it big. WSJ reporter Amrit Ramkumar joins us now with more.
Amrith, what did the EV startup environment look like heading into the presidential election? EV startups were already struggling heading into November. It's been a brutal few years. These companies have seen falling consumer demand, turbulence there. Then they have rising costs, higher interest rates. And so that sort of double whammy is really hitting a lot of them hard. We've seen a slew of bankruptcies already this year. And then, yeah, Trump's election could be a death blow for, frankly, many of these companies.
Right. I mean, President-elect Trump has been critical of the EV industry. How might that play out in his second administration?
Executives in the industry are hopeful that Trump's alliance with Elon Musk will sort of limit some of the fallout, but both Trump and Musk have said they want to get rid of a $7,500 federal tax credit. So that right away is a huge deal for many of these startups because many of them are just starting to ramp up production, which means their cars tend to cost more. They haven't been able to bring the cost down. So that $7,500 federal tax credit is a huge deal that's expected to go away.
And then we're also expected to see a bunch of grants and loans from the Department of Energy and other agencies. Those are expected to be on the chopping block as well. And so a company like Rivian, which announced a $6.6 billion federal loan agreement recently,
Investors aren't sure whether they can get that done by inauguration day in late January and whether that one will actually happen. So there's a great deal of uncertainty and it's the same old saying on Wall Street, investors hate uncertainty. And again, the context here is so important. It's not like these businesses were doing great and then the election happened.
These businesses were really struggling. Just the fundamentals are very challenged right now. We're even seeing established automakers like Ford and GM dial back some of their EV expansion plans. So if you're a startup and all your eggs are in the EV basket, it's just a really difficult time in the US right now. Another thing to note is that this company called Northvolt in Europe that had raised something like $15 billion from big banks and investment firms, they recently filed for bankruptcy.
Right. Northvolt is a Swedish battery startup. Amrith, why were these startups struggling to begin with? You have to go back a few years to like 2021 and 2022 when a lot of these companies went public. At that time, interest rates were essentially zero. We were coming out of COVID. And anyone that said they could try to be the next Tesla became worth billions of dollars seemingly overnight. Right.
So a lot of these companies just had expectations that were set way too high. And then inevitably, they haven't been able to really meet those. A lot of them promised to grow sales and double them every couple of years and hit a billion dollars in sales in historically fast fashion. So yeah, when you make those lofty promises and then don't deliver, investors sell off the stock. So we've seen that happen with a lot of these companies.
And a lot of them, like Faraday Future, Canoo, names like that, they've been distressed for over a year, almost two, even three in some cases. So a lot of these names are sort of hanging on for dear life. And what will be really interesting is if government loans and government support can bail any of them out. So in addition to the government subsidies and loans and other supports, could these companies also be impacted by Trump's proposed tariffs?
Definitely. Proposed tariffs are another huge headwind for a lot of these companies, especially again, their costs are already very high. They're already burning billions of dollars or hundreds of millions of dollars per quarter per year in many cases. So the idea that their input costs are going to go up even further, which is what many economists predict with tariffs.
These tariffs on goods from Mexico and Canada and China, who knows where else and who knows at what level, but many executives are already bracing for higher costs. And that again, when you're already unprofitable and burning through cash, that has a huge impact. And there's also competition in the EV market from Chinese companies. What could this mean going forward?
That's the million, billion, trillion dollar question, depending on how much you value the EV industry of the future at. I mean, a lot of the Chinese EVs investors and executives say are superior in many ways, and they cost a fraction of what EVs cost in the US and Europe, and depending on what you're looking at. So you can factor in tariffs, but even with the tariffs,
the Chinese car still might be roughly cost competitive, which is pretty mind boggling. So companies like BYD, CATL, they're the global leaders in the space. And what many people are worried about is that if the U.S. takes a backseat in the EV industry and doesn't prioritize this area in the next four years, that China's lead will get even bigger. That was our reporter Amrith Ramkumar.
And that's it for Tech News Briefing. Today's show was produced by Julie Chang. I'm your host, Danny Lewis. Jessica Fenton and Michael LaValle wrote our theme music. Our supervising producer is Catherine Millsap. Our development producer is Aisha Al-Muslim. Scott Salloway and Chris Zinsley are the deputy editors, and Falana Patterson is the Wall Street Journal's head of news audio. We'll be back this afternoon with TNB Tech Minute. Thanks for listening.
You want a straightforward path to your goals, but at Merrill, we know things may get in the way.
Or if new opportunities can put you at a crossroads, with the bull at your back, you get a personalized plan and a clear path forward. Go to ml.com slash bullish to learn more. Merrill, a Bank of America company. What would you like the power to do? Investing involves risk. Merrill Lynch Pierce Fenner & Smith Incorporated. Registered broker-dealer. Registered investment advisor. Member SIPC. A wholly owned subsidiary of Bank of America Corp.