Group that blocked New Mexico utility merger eyes data centers in Blackstone deal
NEW YORK (Reuters) -Blackstone Infrastructure's plans for data centers in New Mexico will be a deciding factor in whether stakeholders challenge the private equity group's $11.5-billion proposed acquisition of electric company TXNM Energy, the group that blocked TXNM's previous merger plan told Reuters this week.
TXNM, which is a holding company for regulated utilities, including PNM in New Mexico, announced its sale agreement with Blackstone on Monday in the latest of several recent U.S. power industry deals propelled by rising electricity demand from Big Tech's AI data centers.
The agreement will require the approval of state regulators, with input from PNM stakeholders, including the New Mexico Department of Justice, consumer advocates and clean power groups such as New Energy Economy.
New Energy led the effort to ultimately thwart TXNM's last agreement to sell to power company Avangrid, the U.S. unit of Spanish electric company Iberdrola. After the fight over the proposed acquisition escalated to the New Mexico Supreme Court, Avangrid abandoned its $8.3-billion bid for TXNM in late 2023.
Since the foiled deal, data centers have emerged as the biggest driving force behind U.S. electricity demand, which is on track to reach record highs this year and in 2026.
How Blackstone plans to capitalize on that demand in New Mexico will be a key issue in New Energy Economy's scrutiny of the TXNM purchase, the nonprofit's director, Mariel Nanasi, said.
Among the considerations are whether Blackstone intends to own data centers in New Mexico, either directly or through affiliates, and how it handles the costs of upgrading electrical systems to connect the large energy loads, Nanasi said.
"We are going to want to have real guardrails around that," she said.
TXNM and Blackstone representatives, on a call with investors shortly after the acquisition announcement, said they planned to meet with stakeholders over the next 90 days before filing their plan with the state.
As artificial intelligence data centers proliferate and grow to use record amounts of electricity, regulatory fights have emerged over who pays for additional infrastructure and upgrades needed for the giant energy consumers.
The regulated utilities under TXNM can power data centers, but they are barred by state regulations from developing, owning, or operating the centers for third parties, said one person familiar with the Blackstone-TXNM arrangement, who spoke on condition of anonymity.
Any transmission upgrades or power generation built to serve data centers would be paid for by the data center companies, the person said.
Hashtags
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
The Verge
20 minutes ago
- The Verge
Here's everything you need to know about preordering the new Google Pixel phones
Google announced its entire Pixel 10 generation of phones at a jam-packed Made for Google event this week. The familiar four-phone lineup includes the Pixel 10, Pixel 10 Pro, Pixel 10 Pro XL, and Pixel 10 Pro Fold. All of them have received processor bumps and other technical improvements, along with a commitment by Google to provide seven years of operating system updates. Naturally, there are some new AI features, too. The Pixel 10, 10 Pro, and 10 Pro XL start at $799, $999, and $1,199, respectively, and begin shipping on August 28th. The Pixel 10 Pro Fold, meanwhile, starts at $1,799 and launches on October 9th. If you're confused about what makes each phone tick or which to choose, we've broken everything down, along with preorder information if you decide to pick one up. The new Pixel 10 is Google's latest entry-level flagship smartphone. It has 6.3-inch 120Hz OLED display, Google's new Tensor G5 processor, a battery that can supposedly last more than 30 hours, and a three-lens rear camera system that consists of a 48-megapixel wide, a 13-megapixel ultrawide, and a 10.8-megapixel telephoto lens. The phone comes in two storage tiers, 128GB or 256GB, and both have 12GB of RAM. Google says its G5 processor is 60 percent more powerful than the G4 for on-device AI tasks, and is 34 percent faster overall. The phone also supports Pixelsnap, a new Qi2-certified wireless charging feature. If you prefer wired charging, Google says you can charge its battery from 0 to 55 percent in about 30 minutes using a 30W USB-C power adapter. The Pixel 10 begins shipping on August 28th, and you can preorder the 128GB model at Amazon, Best Buy, or directly from the Google Store for $799. The 256GB model costs $899. Google's Pixel 10 has a 6.3-inch OLED display, Google Tensor G5 processor, 12GB of RAM, support for Pixelsnap wireless charging, and a triple-lens camera system including a new 10.8 megapixel telephoto. The Pixel 10 Pro has the same sized screen and processor as the Pixel 10, but it's an improvement in almost every other way. Its OLED display has a 1Hz to 120Hz refresh rate, higher maximum brightness, and slightly higher resolution. It supports Wi-Fi 7 (the standard Pixel only supports Wi-Fi 6E) for faster networking, and an Ultra-Wideband chip. Its three-lens camera system includes a 50-megapixel wide, a 48-megapixel ultrawide, and a 48-megapixel telephoto, plus the ability to record 8K video. You also get 16GB of RAM and four storage tiers: 128GB, 256GB, 512GB, and 1TB. The Pixel 10 Pro has a slightly smaller battery than the Pixel 10, but Google says its runtime is the same. The Pixel 10 Pro also comes with a 12-month subscription to Google AI Plus (typically $19.99 a month). The Pixel 10 Pro begins shipping on August 28th, and you can preorder the 128GB model at Amazon, Best Buy, or directly from Google starting at $999. The 256GB model costs $1,099, the 512GB model costs $1,219, and the 1TB model costs $1,499. The Pixel 10 Pro has a 6.3-inch OLED display, Google's Tensor G5 processor, 16GB of RAM, Pixelsnap wireless charging at 15W, support for Wi-Fi 7, and an Ultra-Wideband chip. The Google Pixel 10 Pro XL has most of the same specs as Google's standard Pro phone, including an identical processor and the same RAM, camera system, and support for Wi-Fi 7. That being said, it sports a 6.8-inch OLED screen, as opposed to a 6.3-inch display. The phone's bigger display and larger battery cancel one another out, so Google says you get the same battery life. It has the same amount of memory and storage as the standard Google Pixel Pro, too; however, its Pixelsnap charging feature is Qi2.2-certified at speeds up to 25W. If you prefer wired charging, Google says you can charge the 10 Pro XL's battery from 0 to 70 percent in roughly half an hour using a 45W USB-C power adapter. It's only available in three storage sizes — 256GB, 512GB, and 1TB — and includes a 12-month subscription to Google's AI Plus service. The Google Pixel 10 Pro XL is slated to arrive on August 28th, and you can preorder the 256GB model at Amazon, Best Buy, or the Google Store starting at $1,199. The 512GB model costs $1,319 and the 1TB model costs $1,549. The Pixel 10 Pro XL is largest non-foldable smartphone. It has a 6.8-inch OLED screen, Google's Tensor G5 chip, and support for Pixelsnap wireless charging at up to 25W. Google's third-gen foldable has the same processor, RAM, maximum brightness, and battery life as its other Pro phones, but there are some key differences. When folded, the Google Pixel 10 Pro Fold leverages a 6.4-inch 60Hz to 120Hz OLED display. When you open it up, it showcases an 8-inch OLED screen with a refresh rate of 1Hz to 120Hz. Its three-lens rear camera system comprises a 48-megapixel wide, a 10.5-megapixel ultrawide, and a 10.8-megapixel telephoto, with the ability to record 4K video. The Pixel 10 Pro Fold also carries an IP68 rating for dust and water resistance — which is impressive for a foldable phone — and comes with a 12-month subscription to Google's AI Plus service, just like the aforementioned 10 Pro XL. The Pixel 10 Pro Fold begins shipping on October 9th, and you can preorder the entry-level 256GB model at Amazon, Best Buy, or directly from Google starting at $1,799. The 512GB model costs $1,919, and the 1TB model costs $2,149. The Google Pixel 10 Pro Fold has a 6.4-inch outer screen and eight-inch inner screen. The new foldable has Google's Tensor G5 processor, 16GB of RAM, Wi-Fi 7 support, Pixelsnap wireless charging at 15W, and an IP68 rating. If you don't want to pay for your entire phone up front, several US carriers are currently offering deals if you purchase or upgrade through them. Be mindful that taking advantage of these deals will lock your phone to that carrier, and likely mean you'll end up breaking up your purchase into monthly payments that'll increase your monthly bill. These deals are available online, but you may also be able to take advantage of them in-store. Sign up for Verge Deals to get deals on products we've tested sent to your inbox weekly. Posts from this author will be added to your daily email digest and your homepage feed. See All by Brandt Ranj Posts from this topic will be added to your daily email digest and your homepage feed. See All Deals Posts from this topic will be added to your daily email digest and your homepage feed. See All Foldable Phones Posts from this topic will be added to your daily email digest and your homepage feed. See All Gadgets Posts from this topic will be added to your daily email digest and your homepage feed. See All Google Posts from this topic will be added to your daily email digest and your homepage feed. See All Google Pixel Posts from this topic will be added to your daily email digest and your homepage feed. See All Mobile Posts from this topic will be added to your daily email digest and your homepage feed. See All Phones Posts from this topic will be added to your daily email digest and your homepage feed. See All Tech
.jpg&w=3840&q=100)
WIRED
20 minutes ago
- WIRED
Do Large Language Models Dream of AI Agents?
Aug 20, 2025 2:00 PM For AI models, knowing what to remember might be as important as knowing what to forget. Welcome to the era of 'sleeptime compute.' Photo-Illustration:During sleep, the human brain sorts through different memories, consolidating important ones while discarding those that don't matter. What if AI could do the same? Bilt, a company that offers local shopping and restaurant deals to renters, recently deployed several million agents with the hopes of doing just that. Bilt uses technology from a startup called Letta that allows agents to learn from previous conversations and share memories with one another. Using a process called 'sleeptime compute,' the agents decide what information to store in its long-term memory vault and what might be needed for faster recall. 'We can make a single update to a [memory] block and have the behavior of hundreds of thousands of agents change," says Andrew Fitz, an AI engineer at Bilt. "This is useful in any scenario where you want fine-grained control over agents' context,' he adds, referring to the text prompt fed to the model at inference time. Large language models can typically only 'recall' things if information is included in the context window. If you want a chatbot to remember your most recent conversation, you need to paste it into the chat. Most AI systems can only handle a limited amount of information in the context window before their ability to use the data falters and they hallucinate or become confused. The human brain, by contrast, is able to file away useful information and recall it later. 'Your brain is continuously improving, adding more information like a sponge,' says Charles Packer, Letta's CEO. 'With language models, it's like the exact opposite. You run these language models in a loop for long enough and the context becomes poisoned; they get derailed and you just want to reset.' Packer and his cofounder Sarah Wooders previously developed MemGPT, an open-source project that aimed to help LLMs decide what information should be stored in short-term vs. long-term memory. With Letta, the duo has expanded their approach to let agents learn in the background. Bilt's collaboration with Letta is part of a broader push to give AI the ability to store and recall useful information, which could make chatbots smarter and agents less error-prone. Memory remains underdeveloped in modern AI, which undermines the intelligence and reliability of AI tools, according to experts I spoke to. Harrison Chase, cofounder and CEO of LangChain, another company that has developed a method for improving memory in AI agents, says he sees memory as a vital part of context engineering—wherein a user or engineer decides what information to feed into the context window. LangChain offers companies several different kinds of memory storage for agents, from long-term facts about users to memories of recent experiences. 'Memory, I would argue, is a form of context,' Chase says. 'A big portion of an AI engineer's job is basically getting the model the right context [information].' Consumer AI tools are gradually becoming less forgetful, too. This February, OpenAI announced that ChatGPT will store relevant information in order to provide a more personalized experience for users—although the company did not disclose how this works. Letta and LangChain make the process of recall more transparent to engineers building AI systems. 'I think it's super important not only for the models to be open but also for the memory systems to be open,' says Clem Delangue, CEO of the AI hosting platform Hugging Face and an investor in Letta. Intriguingly, Letta's CEO Packer hints that it might also be important for AI models to learn what to forget. 'If a user says, 'that one project we were working on, wipe it out from your memory' then the agent should be able to go back and retroactively rewrite every single memory.' The notion of artificial memories and dreams makes me think of Do Androids Dream of Electric Sheep? by Philip K. Dick, a mind-bending novel that inspired the stylishly dystopian movie Blade Runner . Large language models aren't yet as impressive as the rebellious replicants of the story, but their memories, it seems, can be just as fragile. This is an edition of Will Knight's AI Lab newsletter. Read previous newsletters here.
Forbes
20 minutes ago
- Forbes
The Radical Changes AI Is Bringing To Higher Education
From the printing press to typewriters to personal computers to the internet, technology has been instrumental in reshaping higher education over the past 500 years. At the same time, much of that reshaping has been closer to the edges rather than a complete upheaval. Artificial intelligence has the potential to transform and redefine higher education at its core, and will likely do so in the next year. Using a modified quote from ChatGPT, AI is coming to higher education like a squirrel on espresso! As with other industries, AI is set to redefine higher education, for better and worse. Here is a select set of areas that will be impacted: In all, AI-integrated higher education has the potential to hold much promise. Of course, the potential has already been met with a backlash. Similar to the church's reaction to the printing press, higher education institutions are spending a disproportionate amount of time attempting to limit the use of AI, assuming the current educational model should go unchallenged and unchanged. Rather than covering one's ears, the key will be to limit the potentially ineffective aspects of AI by opening up the higher education enterprise (e.g., open-source the educational approach to ensure adequate critique). If done well, AI can revolutionize higher education for the better.
