How do AI models generate videos?

Stay Ahead, Stay ONMINE

How do AI models generate videos?

Sure, the clips you see in demo reels are cherry-picked to showcase a company’s models at the top of their game. But with the technology in the hands of more users than ever before—Sora and Veo 3 are available in the ChatGPT and Gemini apps for paying subscribers—even the most casual filmmaker can now knock out something remarkable. The downside is that creators are competing with AI slop, and social media feeds are filling up with faked news footage. Video generation also uses up a huge amount of energy, many times more than text or image generation. With AI-generated videos everywhere, let’s take a moment to talk about the tech that makes them work. How do you generate a video? Let’s assume you’re a casual user. There are now a range of high-end tools that allow pro video makers to insert video generation models into their workflows. But most people will use this technology in an app or via a website. You know the drill: “Hey, Gemini, make me a video of a unicorn eating spaghetti. Now make its horn take off like a rocket.” What you get back will be hit or miss, and you’ll typically need to ask the model to take another pass or 10 before you get more or less what you wanted. [embedded content] So what’s going on under the hood? Why is it hit or miss—and why does it take so much energy? The latest wave of video generation models are what’s known as latent diffusion transformers. Yes, that’s quite a mouthful. Let’s unpack each part in turn, starting with diffusion. What’s a diffusion model? Imagine taking an image and adding a random spattering of pixels to it. Take that pixel-spattered image and spatter it again and then again. Do that enough times and you will have turned the initial image into a random mess of pixels, like static on an old TV set. A diffusion model is a neural network trained to reverse that process, turning random static into images. During training, it gets shown millions of images in various stages of pixelation. It learns how those images change each time new pixels are thrown at them and, thus, how to undo those changes. The upshot is that when you ask a diffusion model to generate an image, it will start off with a random mess of pixels and step by step turn that mess into an image that is more or less similar to images in its training set. [embedded content] But you don’t want any image—you want the image you specified, typically with a text prompt. And so the diffusion model is paired with a second model—such as a large language model (LLM) trained to match images with text descriptions—that guides each step of the cleanup process, pushing the diffusion model toward images that the large language model considers a good match to the prompt. An aside: This LLM isn’t pulling the links between text and images out of thin air. Most text-to-image and text-to-video models today are trained on large data sets that contain billions of pairings of text and images or text and video scraped from the internet (a practice many creators are very unhappy about). This means that what you get from such models is a distillation of the world as it’s represented online, distorted by prejudice (and pornography). It’s easiest to imagine diffusion models working with images. But the technique can be used with many kinds of data, including audio and video. To generate movie clips, a diffusion model must clean up sequences of images—the consecutive frames of a video—instead of just one image. What’s a latent diffusion model? All this takes a huge amount of compute (read: energy). That’s why most diffusion models used for video generation use a technique called latent diffusion. Instead of processing raw data—the millions of pixels in each video frame—the model works in what’s known as a latent space, in which the video frames (and text prompt) are compressed into a mathematical code that captures just the essential features of the data and throws out the rest. A similar thing happens whenever you stream a video over the internet: A video is sent from a server to your screen in a compressed format to make it get to you faster, and when it arrives, your computer or TV will convert it back into a watchable video. And so the final step is to decompress what the latent diffusion process has come up with. Once the compressed frames of random static have been turned into the compressed frames of a video that the LLM guide considers a good match for the user’s prompt, the compressed video gets converted into something you can watch. With latent diffusion, the diffusion process works more or less the way it would for an image. The difference is that the pixelated video frames are now mathematical encodings of those frames rather than the frames themselves. This makes latent diffusion far more efficient than a typical diffusion model. (Even so, video generation still uses more energy than image or text generation. There’s just an eye-popping amount of computation involved.) What’s a latent diffusion transformer? Still with me? There’s one more piece to the puzzle—and that’s how to make sure the diffusion process produces a sequence of frames that are consistent, maintaining objects and lighting and so on from one frame to the next. OpenAI did this with Sora by combining its diffusion model with another kind of model called a transformer. This has now become standard in generative video. Transformers are great at processing long sequences of data, like words. That has made them the special sauce inside large language models such as OpenAI’s GPT-5 and Google DeepMind’s Gemini, which can generate long sequences of words that make sense, maintaining consistency across many dozens of sentences. But videos are not made of words. Instead, videos get cut into chunks that can be treated as if they were. The approach that OpenAI came up with was to dice videos up across both space and time. “It’s like if you were to have a stack of all the video frames and you cut little cubes from it,” says Tim Brooks, a lead researcher on Sora. [embedded content] A selection of videos generated with Veo 3 and Midjourney. The clips have been enhanced in postproduction with Topaz, an AI video-editing tool. Credit: VaigueMan Using transformers alongside diffusion models brings several advantages. Because they are designed to process sequences of data, transformers also help the diffusion model maintain consistency across frames as it generates them. This makes it possible to produce videos in which objects don’t pop in and out of existence, for example. And because the videos are diced up, their size and orientation do not matter. This means that the latest wave of video generation models can be trained on a wide range of example videos, from short vertical clips shot with a phone to wide-screen cinematic films. The greater variety of training data has made video generation far better than it was just two years ago. It also means that video generation models can now be asked to produce videos in a variety of formats. What about the audio? A big advance with Veo 3 is that it generates video with audio, from lip-synched dialogue to sound effects to background noise. That’s a first for video generation models. As Google DeepMind CEO Demis Hassabis put it at this year’s Google I/O: “We’re emerging from the silent era of video generation.” [embedded content] The challenge was to find a way to line up video and audio data so that the diffusion process would work on both at the same time. Google DeepMind’s breakthrough was a new way to compress audio and video into a single piece of data inside the diffusion model. When Veo 3 generates a video, its diffusion model produces audio and video together in a lockstep process, ensuring that the sound and images are synched. You said that diffusion models can generate different kinds of data. Is this how LLMs work too? No—or at least not yet. Diffusion models are most often used to generate images, video, and audio. Large language models—which generate text (including computer code)—are built using transformers. But the lines are blurring. We’ve seen how transformers are now being combined with diffusion models to generate videos. And this summer Google DeepMind revealed that it was building an experimental large language model that used a diffusion model instead of a transformer to generate text. Here’s where things start to get confusing: Though video generation (which uses diffusion models) consumes a lot of energy, diffusion models themselves are in fact more efficient than transformers. Thus, by using a diffusion model instead of a transformer to generate text, Google DeepMind’s new LLM could be a lot more efficient than existing LLMs. Expect to see more from diffusion models in the near future!

The downside is that creators are competing with AI slop, and social media feeds are filling up with faked news footage. Video generation also uses up a huge amount of energy, many times more than text or image generation.

With AI-generated videos everywhere, let’s take a moment to talk about the tech that makes them work.

How do you generate a video?

Let’s assume you’re a casual user. There are now a range of high-end tools that allow pro video makers to insert video generation models into their workflows. But most people will use this technology in an app or via a website. You know the drill: “Hey, Gemini, make me a video of a unicorn eating spaghetti. Now make its horn take off like a rocket.” What you get back will be hit or miss, and you’ll typically need to ask the model to take another pass or 10 before you get more or less what you wanted.

So what’s going on under the hood? Why is it hit or miss—and why does it take so much energy? The latest wave of video generation models are what’s known as latent diffusion transformers. Yes, that’s quite a mouthful. Let’s unpack each part in turn, starting with diffusion.

What’s a diffusion model?

Imagine taking an image and adding a random spattering of pixels to it. Take that pixel-spattered image and spatter it again and then again. Do that enough times and you will have turned the initial image into a random mess of pixels, like static on an old TV set.

A diffusion model is a neural network trained to reverse that process, turning random static into images. During training, it gets shown millions of images in various stages of pixelation. It learns how those images change each time new pixels are thrown at them and, thus, how to undo those changes.

The upshot is that when you ask a diffusion model to generate an image, it will start off with a random mess of pixels and step by step turn that mess into an image that is more or less similar to images in its training set.

But you don’t want any image—you want the image you specified, typically with a text prompt. And so the diffusion model is paired with a second model—such as a large language model (LLM) trained to match images with text descriptions—that guides each step of the cleanup process, pushing the diffusion model toward images that the large language model considers a good match to the prompt.

An aside: This LLM isn’t pulling the links between text and images out of thin air. Most text-to-image and text-to-video models today are trained on large data sets that contain billions of pairings of text and images or text and video scraped from the internet (a practice many creators are very unhappy about). This means that what you get from such models is a distillation of the world as it’s represented online, distorted by prejudice (and pornography).

It’s easiest to imagine diffusion models working with images. But the technique can be used with many kinds of data, including audio and video. To generate movie clips, a diffusion model must clean up sequences of images—the consecutive frames of a video—instead of just one image.

What’s a latent diffusion model?

All this takes a huge amount of compute (read: energy). That’s why most diffusion models used for video generation use a technique called latent diffusion. Instead of processing raw data—the millions of pixels in each video frame—the model works in what’s known as a latent space, in which the video frames (and text prompt) are compressed into a mathematical code that captures just the essential features of the data and throws out the rest.

A similar thing happens whenever you stream a video over the internet: A video is sent from a server to your screen in a compressed format to make it get to you faster, and when it arrives, your computer or TV will convert it back into a watchable video.

And so the final step is to decompress what the latent diffusion process has come up with. Once the compressed frames of random static have been turned into the compressed frames of a video that the LLM guide considers a good match for the user’s prompt, the compressed video gets converted into something you can watch.

With latent diffusion, the diffusion process works more or less the way it would for an image. The difference is that the pixelated video frames are now mathematical encodings of those frames rather than the frames themselves. This makes latent diffusion far more efficient than a typical diffusion model. (Even so, video generation still uses more energy than image or text generation. There’s just an eye-popping amount of computation involved.)

What’s a latent diffusion transformer?

Still with me? There’s one more piece to the puzzle—and that’s how to make sure the diffusion process produces a sequence of frames that are consistent, maintaining objects and lighting and so on from one frame to the next. OpenAI did this with Sora by combining its diffusion model with another kind of model called a transformer. This has now become standard in generative video.

Transformers are great at processing long sequences of data, like words. That has made them the special sauce inside large language models such as OpenAI’s GPT-5 and Google DeepMind’s Gemini, which can generate long sequences of words that make sense, maintaining consistency across many dozens of sentences.

But videos are not made of words. Instead, videos get cut into chunks that can be treated as if they were. The approach that OpenAI came up with was to dice videos up across both space and time. “It’s like if you were to have a stack of all the video frames and you cut little cubes from it,” says Tim Brooks, a lead researcher on Sora.

A selection of videos generated with Veo 3 and Midjourney. The clips have been enhanced in postproduction with Topaz, an AI video-editing tool. Credit: VaigueMan

Using transformers alongside diffusion models brings several advantages. Because they are designed to process sequences of data, transformers also help the diffusion model maintain consistency across frames as it generates them. This makes it possible to produce videos in which objects don’t pop in and out of existence, for example.

And because the videos are diced up, their size and orientation do not matter. This means that the latest wave of video generation models can be trained on a wide range of example videos, from short vertical clips shot with a phone to wide-screen cinematic films. The greater variety of training data has made video generation far better than it was just two years ago. It also means that video generation models can now be asked to produce videos in a variety of formats.

What about the audio?

A big advance with Veo 3 is that it generates video with audio, from lip-synched dialogue to sound effects to background noise. That’s a first for video generation models. As Google DeepMind CEO Demis Hassabis put it at this year’s Google I/O: “We’re emerging from the silent era of video generation.”

The challenge was to find a way to line up video and audio data so that the diffusion process would work on both at the same time. Google DeepMind’s breakthrough was a new way to compress audio and video into a single piece of data inside the diffusion model. When Veo 3 generates a video, its diffusion model produces audio and video together in a lockstep process, ensuring that the sound and images are synched.

You said that diffusion models can generate different kinds of data. Is this how LLMs work too?

No—or at least not yet. Diffusion models are most often used to generate images, video, and audio. Large language models—which generate text (including computer code)—are built using transformers. But the lines are blurring. We’ve seen how transformers are now being combined with diffusion models to generate videos. And this summer Google DeepMind revealed that it was building an experimental large language model that used a diffusion model instead of a transformer to generate text.

Here’s where things start to get confusing: Though video generation (which uses diffusion models) consumes a lot of energy, diffusion models themselves are in fact more efficient than transformers. Thus, by using a diffusion model instead of a transformer to generate text, Google DeepMind’s new LLM could be a lot more efficient than existing LLMs. Expect to see more from diffusion models in the near future!

Stay Ahead

Explore More Insights

Stay ahead with more perspectives on cutting-edge power, infrastructure, energy, bitcoin and AI solutions. Explore these articles to uncover strategies and insights shaping the future of industries.

F5 to acquire CalypsoAI for advanced AI security capabilities

CalypsoAI’s platform creates what the company calls an Inference Perimeter that protects across models, vendors, and environments. The offers several products including Inference Red Team, Inference Defend, and Inference Observe, which deliver adversarial testing, threat detection and prevention, and enterprise oversight, respectively, among other capabilities. CalypsoAI says its platform proactively

HomeLM: A foundation model for ambient AI

Capabilities of a HomeLM What makes a foundation model like HomeLM powerful is its ability to learn generalizable representations of sensor streams, allowing them to be reused, recombined and adapted across diverse tasks. This fundamentally differs from traditional signal processing and machine learning pipelines in RF sensing, which are typically

Google enlists new ‘free’ data transfer service as the EU Data Act deadline approaches

Google beyond the basic requirement When asked about the rationale behind the introduction of the new service, a Google spokesperson said that the hyperscaler was going “above and beyond what the data act requires by eliminating the cost of in-parallel processing fees for customers in the EU and UK.” In

Cisco’s Splunk embeds agentic AI into security and observability products

AI-powered observability enhancements Cisco also announced it has updated Splunk Observability to use Cisco AgenticOps, which deploys AI agents to automate telemetry collection, detect issues, identify root causes, and apply fixes. The agentic AI updates help enterprise customers automate incident detection, root-cause analysis, and routine fixes. “We are making sure

USA EIA Reveals Latest Brent Oil Price Forecast

The U.S. Energy Information Administration (EIA) revealed its latest Brent spot price forecast for 2025 and 2026 in its September short term energy outlook (STEO), which was released on September 9. According to its latest STEO, the EIA expects the Brent spot price to average $67.80 per barrel in 2025 and $51.43 per barrel in 2026. In its previous STEO, which was released in August, the EIA projected that the Brent spot price would average $67.22 per barrel this year and $51.43 per barrel next year. A quarterly breakdown included in the EIA’s September STEO showed that the EIA sees the Brent spot price coming in at $68.35 per barrel in the third quarter of 2025, $59.41 per barrel in the fourth quarter, $49.97 per barrel in the first quarter of 2026, $49.67 per barrel in the second quarter, $52 per barrel in the third quarter, and $54 per barrel in the fourth quarter. In its previous STEO, the EIA forecast that the Brent spot price would average $67.40 per barrel in the third quarter of this year, $58.05 per barrel in the fourth quarter, $49.97 per barrel in the first quarter of next year, $49.67 per barrel in the second quarter, $52 per barrel in the third quarter, and $54 per barrel in the fourth quarter. “We expect the Brent crude oil price will decline significantly in the coming months, falling from $68 per barrel in August to $59 per barrel on average in the fourth quarter of 2025 and around $50 per barrel in early 2026,” the EIA said in its latest STEO. “The price forecast is driven by large oil inventory builds as OPEC+ members increase production. We expect global oil inventory builds will average more than two million barrels per day from 3Q25 through 1Q26,” it

ReconAfrica Signs PSC for Asset Offshore Gabon

Reconnaissance Energy Africa Ltd. said it has entered into a production sharing contract (PSC) and a joint venture agreement for the exploration, appraisal, development, and production of oil and gas on Gabon offshore Block C-7, renamed Ngulu. Under the terms of the joint venture, ReconAfrica will be the designated operator, with a 55 percent working interest, Record Resources Inc. will have a 20 percent working interest, the Gabon Oil Company (GOC) with 15 percent, and the Republic of Gabon with 10 percent. For the initial four-year term of the PSC, ReconAfrica plans to execute detailed geological and geophysical studies, advance 3D seismic reprocessing on an existing database, and drill one well on the block, the company said in a news release. The PSC covers 468.7 square miles (1,214 square kilometers) and is located in shallow water offshore central Gabon. The Ngulu block is near several sizeable producing fields that range in size from 38 million barrels to up to 250 million barrels, ReconAfrica said. The key aspects of the Ngulu Block include the Loba oil field discovery and over 28 seismically identified prospects in the Pre-salt Gamba / Dentale and Post-salt play, according to the release. According to ReconAfrica, the entry terms are attractive for a large concession secured at a low-entry cost with a minimal work program over the initial four-year period. Ngulu contains near-term oil production potential with the development of the Loba field and access to existing infrastructure within 6.2 miles (10 kilometers), which provides a low-cost production tie-back. The block also adds an extensive inventory of high-impact exploration projects for future drilling, in addition to exposure to a large pipeline of brownfield opportunities across Gabon, the company noted. The transaction advances ReconAfrica’s strategy of “becoming an offshore West Africa exploration and production company with an expanded

Oil and Gas Workforce Shrinks

The number of employees in the oil and gas extraction industry has dropped by thousands since the start of the year, data on the U.S. Bureau of Labor Statistics (BLS) website shows. According to the data, which was accessed by Rigzone, the number of employees in the sector has declined from 123,100 in January to 119,100 in August. Figures have recorded a month on month drop on four occasions so far this year, the data reveals. The number of employees in the oil and gas extraction industry stood at 122,400 in February, 122,300 in March, 122,600 in April and May, 121,600 in June, and 121,200 in July, the data highlighted. July and August figures are preliminary, the data pointed out. The BLS data, which spans from January 1972 to August 2025, showed that the highest oil and gas extraction industry employee figure was seen in March 1982, at 267,000. By December 2003, this figure had declined to 118,400, according to the data, which revealed that this figure climbed back up above 200,000, to 200,800, in October 2014. The lowest figure in the data set was seen in January 2022, at 110,800. The BLS data is taken from the national Current Employment Statistics survey, the data page outlines. The CES program produces detailed industry estimates of nonfarm employment, hours, and earnings of workers on payrolls, the BLS site states, adding that CES National Estimates produces data for the nation and CES State and Metro Area produces estimates for all 50 States, the District of Columbia, Puerto Rico, the Virgin Islands, and about 450 metropolitan areas and divisions. Each month, CES surveys approximately 121,000 businesses and government agencies, representing approximately 631,000 individual worksites, the BLS site notes. The oil and gas extraction subsector is part of the mining, quarrying, and oil and gas extraction sector, the BLS site

Norway Opens Block Nominations for Next Leasing

Norway’s upstream regulator is calling on companies with offshore assets in Norway to suggest blocks for the next licensing round. Nominations for the 26th leasing round for the Norwegian continental shelf (NCS) are due October 1, the Norwegian Offshore Directorate (NOD) said in a statement on its website. Areas that could be nominated are spread across Norway’s side of the Barents Sea, the Norwegian Sea and the Norwegian sector of the North Sea. “Available attractive acreage is the key to ensure continued high activity in the industry”, the NOD said. Numbered licensing rounds offer opened, frontier parts of the NCS with the greatest potential for large discoveries, according to the NOD. “The total number of nominated blocks should be limited to no more than 15”, the NOD says on an information page for the upcoming round. Companies cannot nominate areas included in Awards in Pre-Defined Areas (APA), the page says. The Energy Ministry launched the process for the 26th licensing round last month. “The companies’ input in the nomination process, together with the Offshore Directorate’s own subsurface assessments, will form the basis for the Directorate’s petroleum-related recommendation to the Ministry on which acreage should be announced in the 26th licensing round”, the ministry said in a press release August 12. Energy Minister Terje Aasland commented, “Norway will remain a long-term supplier of oil and gas to Europe, while the Norwegian continental shelf will continue to create values and jobs for our country. To deliver on this commitment, we must make more discoveries – and to make more discoveries, we must explore”. Last week the NOD said 20 companies had applied for the 2025 APA, signaling “significant interest in exploring new acreage in mature areas of the Norwegian continental shelf”. “We note that most of the companies that are active on the Norwegian

Major Aussie LNG Plant Gets 40 More Years

Australia’s first and largest liquefied natural gas plant cleared the final hurdle to operate until 2070, after an almost seven-year process during which operator Woodside Energy Group Ltd. faced opposition from environmentalists and traditional owners concerned about the site’s pollution. Environment Minister Murray Watt has allowed the North West Shelf project in Western Australia to keep operating for an additional 40 years, almost four months after he granted provisional approval. Woodside has agreed to 48 conditions to limit the site’s impact on local Aboriginal rock art. “This is the last procedural step in a lengthy process,” Watt said in a statement. “I have imposed conditions that will require a reduction in certain gas emissions below their current levels, in some cases by 60 percent by 2030 with ongoing reductions beyond that,” he said, without specifying the gases. Australia has had a rocky relationship with its LNG export industry since North West Shelf exported the nation’s first cargo of the superchilled gas in 1989. The nation’s 10 plants accounted for about a fifth of global exports last year – earning almost A$69 billion ($46 billion) from shipments focused on Japan, South Korea and China – but have been criticized for helping turn the nation into one of the biggest per-capita emitters while paying relatively little in royalties. North West Shelf is also one of the nation’s top polluters, and emissions through 2070 – including from the burning of its gas abroad – have been estimated to be about 10 times Australia’s current annual total. It has also faced criticism for its potential effect on a nearby ancient Aboriginal rock art site containing more than 1 million engravings – which Watt in July successfully lobbied to have designated as a United Nations Cultural World Heritage Site. The conditions require additional monitoring and management of air emissions to protect the

Petrofac Reaches Agreement in Principle with Thai Project Partners

Petrofac Ltd said Thursday it had reached an agreement in principle with Saipem SpA and Samsung E&A Co Ltd on their claims from a failed project in Thailand, which have derailed Petrofac’s financial restructuring plan. “The commercial terms are supported by the Ad Hoc Group of Bondholders, subject to the agreement of longform documents”, the energy engineering company, now based in London after moving from Jersey, said in a brief statement on its website. “This agreement will enable the restructuring to proceed with the consent of those parties. “The Group will now work to conclude discussions with key stakeholders on next steps towards implementation of the restructuring. It is expected that, subject to receipt of all requisite approvals and satisfaction of conditions, the restructuring will be completed by the end of November 2025 and the company will share further details as soon as appropriate to do so, in line with its disclosure obligations”. On July 1, in favor of Saipem and Samsung E&A, the Court of Appeal set aside the High Court of Justice’s sanction for Petrofac’s restructuring. Petrofac and the two companies had been part of a discontinued project, called Clean Fuels Project, to enable the production of cleaner fuels at a Thai refinery. Claims by Saipem and Samsung and those of other parties to the Thai project “will be compromised” under the restructuring plan, the Court of Appeal said in its decision, published on the UK judiciary’s website. Last month Petrofac said it would apply to the United Kingdom Supreme Court for leave to appeal the reversal of the High Court’s sanction. “The board has identified and is actively progressing a number of routes to deliver the restructuring, including through a plan that addresses the narrow grounds on which the Court of Appeal issued its judgment on 1 July

There are 121 AI processor companies. How many will succeed?

The US currently leads in AI hardware and software, but China’s DeepSeek and Huawei continue to push advanced chips, India has announced an indigenous GPU program targeting production by 2029, and policy shifts in Washington are reshaping the playing field. In Q2, the rollback of export restrictions allowed US companies like Nvidia and AMD to strike multibillion-dollar deals in Saudi Arabia. JPR categorizes vendors into five segments: IoT (ultra-low-power inference in microcontrollers or small SoCs); Edge (on-device or near-device inference in 1–100W range, used outside data centers); Automotive (distinct enough to break out from Edge); data center training; and data center inference. There is some overlap between segments as many vendors play in multiple segments. Of the five categories, inference has the most startups with 90. Peddie says the inference application list is “humongous,” with everything from wearable health monitors to smart vehicle sensor arrays, to personal items in the home, and every imaginable machine in every imaginable manufacturing and production line, plus robotic box movers and surgeons. Inference also offers the most versatility. “Smart devices” in the past, like washing machines or coffee makers, could do basically one thing and couldn’t adapt to any changes. “Inference-based systems will be able to duck and weave, adjust in real time, and find alternative solutions, quickly,” said Peddie. Peddie said despite his apparent cynicism, this is an exciting time. “There are really novel ideas being tried like analog neuron processors, and in-memory processors,” he said.

Data Center Jobs: Engineering, Construction, Commissioning, Sales, Field Service and Facility Tech Jobs Available in Major Data Center Hotspots

Each month Data Center Frontier, in partnership with Pkaza, posts some of the hottest data center career opportunities in the market. Here’s a look at some of the latest data center jobs posted on the Data Center Frontier jobs board, powered by Pkaza Critical Facilities Recruiting. Looking for Data Center Candidates? Check out Pkaza’s Active Candidate / Featured Candidate Hotlist (and coming soon free Data Center Intern listing). Data Center Critical Facility Manager Impact, TX There position is also available in: Cheyenne, WY; Ashburn, VA or Manassas, VA. This opportunity is working directly with a leading mission-critical data center developer / wholesaler / colo provider. This firm provides data center solutions custom-fit to the requirements of their client’s mission-critical operational facilities. They provide reliability of mission-critical facilities for many of the world’s largest organizations (enterprise and hyperscale customers). This career-growth minded opportunity offers exciting projects with leading-edge technology and innovation as well as competitive salaries and benefits. Electrical Commissioning Engineer New Albany, OH This traveling position is also available in: Richmond, VA; Ashburn, VA; Charlotte, NC; Atlanta, GA; Hampton, GA; Fayetteville, GA; Cedar Rapids, IA; Phoenix, AZ; Dallas, TX or Chicago, IL. *** ALSO looking for a LEAD EE and ME CxA Agents and CxA PMs. *** Our client is an engineering design and commissioning company that has a national footprint and specializes in MEP critical facilities design. They provide design, commissioning, consulting and management expertise in the critical facilities space. They have a mindset to provide reliability, energy efficiency, sustainable design and LEED expertise when providing these consulting services for enterprise, colocation and hyperscale companies. This career-growth minded opportunity offers exciting projects with leading-edge technology and innovation as well as competitive salaries and benefits. Data Center Engineering Design ManagerAshburn, VA This opportunity is working directly with a leading mission-critical data center developer /

Modernizing Legacy Data Centers for the AI Revolution with Schneider Electric’s Steven Carlini

As artificial intelligence workloads drive unprecedented compute density, the U.S. data center industry faces a formidable challenge: modernizing aging facilities that were never designed to support today’s high-density AI servers. In a recent Data Center Frontier podcast, Steven Carlini, Vice President of Innovation and Data Centers at Schneider Electric, shared his insights on how operators are confronting these transformative pressures. “Many of these data centers were built with the expectation they would go through three, four, five IT refresh cycles,” Carlini explains. “Back then, growth in rack density was moderate. Facilities were designed for 10, 12 kilowatts per rack. Now with systems like Nvidia’s Blackwell, we’re seeing 132 kilowatts per rack, and each rack can weigh 5,000 pounds.” The implications are seismic. Legacy racks, floor layouts, power distribution systems, and cooling infrastructure were simply not engineered for such extreme densities. “With densification, a lot of the power distribution, cooling systems, even the rack systems — the new servers don’t fit in those racks. You need more room behind the racks for power and cooling. Almost everything needs to be changed,” Carlini notes. For operators, the first questions are inevitably about power availability. At 132 kilowatts per rack, even a single cluster can challenge the limits of older infrastructure. Many facilities are conducting rigorous evaluations to decide whether retrofitting is feasible or whether building new sites is the more practical solution. Carlini adds, “You may have transformers spaced every hundred yards, twenty of them. Now, one larger transformer can replace that footprint, and power distribution units feed busways that supply each accelerated compute rack. The scale and complexity are unlike anything we’ve seen before.” Safety considerations also intensify with these densifications. “At 132 kilowatts, maintenance is still feasible,” Carlini says, “but as voltages rise, data centers are moving toward environments where

Google Backs Advanced Nuclear at TVA’s Clinch River as ORNL Pushes Quantum Frontiers

Inside the Hermes Reactor Design Kairos Power’s Hermes reactor is based on its KP-FHR architecture — short for fluoride salt–cooled, high-temperature reactor. Unlike conventional water-cooled reactors, Hermes uses a molten salt mixture called FLiBe (lithium fluoride and beryllium fluoride) as a coolant. Because FLiBe operates at atmospheric pressure, the design eliminates the risk of high-pressure ruptures and allows for inherently safer operation. Fuel for Hermes comes in the form of TRISO particles rather than traditional enriched uranium fuel rods. Each TRISO particle is encapsulated within ceramic layers that function like miniature containment vessels. These particles can withstand temperatures above 1,600 °C — far beyond the reactor’s normal operating range of about 700 °C. In combination with the salt coolant, Hermes achieves outlet temperatures between 650–750 °C, enabling efficient power generation and potential industrial applications such as hydrogen production. Because the salt coolant is chemically stable and requires no pressurization, the reactor can shut down and dissipate heat passively, without external power or operator intervention. This passive safety profile differentiates Hermes from traditional light-water reactors and reflects the Generation IV industry focus on safer, modular designs. From Hermes-1 to Hermes-2: Iterative Nuclear Development The first step in Kairos’ roadmap is Hermes-1, a 35 MW thermal demonstration reactor now under construction at TVA’s Clinch River site under a 2023 NRC license. Hermes-1 is not designed to generate electricity but will validate reactor physics, fuel handling, licensing strategies, and construction techniques. Building on that experience, Hermes-2 will be a 50 MW electric reactor connected to TVA’s grid, with operations targeted for 2030. Under the agreement, TVA will purchase electricity from Hermes-2 and supply it to Google’s data centers in Tennessee and Alabama. Kairos describes its development philosophy as “iterative,” scaling incrementally rather than attempting to deploy large fleets of units at once. By

NVIDIA Forecasts $3–$4 Trillion AI Market, Driving Next Wave of Infrastructure

Whenever behemoth chipmaker NVIDIA announces its quarterly earnings, those results can have a massive influence on the stock market and its position as a key indicator for the AI industry. After all, NVIDIA is the most valuable publicly traded company in the world, valued at $4.24 trillion—ahead of Microsoft ($3.74 trillion), Apple ($3.41 trillion), Alphabet, the parent company of Google ($2.57 trillion), and Amazon ($2.44 trillion). Due to its explosive growth in recent years, a single NVIDIA earnings report can move the entire market. So, when NVIDIA leaders announced during their August 27 earnings call that Q2 2026 sales surged 56% to $46.74 billion, it was a record-setting performance for the company—and investors took notice. Executive VP & CFO Colette M. Kress said the revenue exceeded leadership’s outlook as the company grew sequentially across all market platforms. She outlined a path toward substantial growth driven by AI infrastructure. Foreseeing significant long-term growth opportunities in agentic AI and considering the scale of opportunity, CEO Jensen Huang said, “Over the next 5 years, we’re going to scale into it with Blackwell [architecture for GenAI], with Rubin [successor to Blackwell], and follow-ons to scale into effectively a $3 trillion to $4 trillion AI infrastructure opportunity.” The chipmaker’s Q2 2026 earnings fell short of Wall Street’s lofty expectations, but they did demonstrate that its sales are still rising faster than those of most other tech companies. NVIDIA is expected to post revenue growth of at least 42% over the next four quarters, compared with an average of about 10% for firms in the technology-heavy Nasdaq 100 Index, according to data compiled by Bloomberg Intelligence. On August 29, two days after announcing their earnings, NVIDIA stocks slid 3% and other chip stocks also declined. This came amid a broader sell-off after server-maker Dell, a customer of those chipmakers,

Cologix and Lambda Debut NVIDIA HGX B200 AI Clusters in Columbus, Ohio

In our latest episode of the Data Center Frontier Show, we explore how powerhouse AI infrastructure is moving inland—anchored by the first NVIDIA HGX B200 cluster deployment in Columbus, Ohio. Cologix, Lambda, and Supermicro have partnered on the project, which combines Lambda’s 1-Click Clusters™, Supermicro’s energy-efficient hardware, and Cologix’s carrier-dense Scalelogix℠ COL4 facility. It’s a milestone that speaks to the rapid decentralization of AI workloads and the emergence of the Midwest as a serious player in the AI economy. Joining me for the conversation were Bill Bentley, VP Hyperscale and Cloud Sales at Cologix, and Ken Patchett, VP Data Center Infrastructure at Lambda. Why Columbus, Why Now? Asked about the significance of launching in Columbus, Patchett framed the move in terms of the coming era of “superintelligence.” “The shift to superintelligence is happening now—systems that can reason, adapt, and accelerate human progress,” Patchett said. “That requires an entirely new type of infrastructure, which means capital, vision, and the right partners. Columbus with Cologix made sense because beyond being centrally located, they’re highly connected, cost-efficient, and built to scale. We’re not chasing trends. We’re laying the groundwork for a future where intelligence infrastructure is as ubiquitous as electricity.” Bentley pointed to the city’s underlying strengths in connectivity, incentives, and utility economics. “Columbus is uniquely situated at the intersection of long-haul fiber,” Bentley said. “You’ve got state tax incentives, low-cost utilities, and a growing concentration of hyperscalers and local enterprises. The ecosystem is ripe for growth. It’s a natural geography for AI workloads that need geographic diversity without sacrificing performance.” Shifting—or Expanding—the Map for AI The guests agreed that deployments like this don’t represent a wholesale shift away from coastal hyperscale markets, but rather the expansion of AI’s footprint across multiple geographies. “I like to think of Lambda as an AI hyperscaler,”

Microsoft will invest $80B in AI data centers in fiscal 2025

And Microsoft isn’t the only one that is ramping up its investments into AI-enabled data centers. Rival cloud service providers are all investing in either upgrading or opening new data centers to capture a larger chunk of business from developers and users of large language models (LLMs). In a report published in October 2024, Bloomberg Intelligence estimated that demand for generative AI would push Microsoft, AWS, Google, Oracle, Meta, and Apple would between them devote $200 billion to capex in 2025, up from $110 billion in 2023. Microsoft is one of the biggest spenders, followed closely by Google and AWS, Bloomberg Intelligence said. Its estimate of Microsoft’s capital spending on AI, at $62.4 billion for calendar 2025, is lower than Smith’s claim that the company will invest $80 billion in the fiscal year to June 30, 2025. Both figures, though, are way higher than Microsoft’s 2020 capital expenditure of “just” $17.6 billion. The majority of the increased spending is tied to cloud services and the expansion of AI infrastructure needed to provide compute capacity for OpenAI workloads. Separately, last October Amazon CEO Andy Jassy said his company planned total capex spend of $75 billion in 2024 and even more in 2025, with much of it going to AWS, its cloud computing division.

John Deere unveils more autonomous farm machines to address skill labor shortage

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More Self-driving tractors might be the path to self-driving cars. John Deere has revealed a new line of autonomous machines and tech across agriculture, construction and commercial landscaping. The Moline, Illinois-based John Deere has been in business for 187 years, yet it’s been a regular as a non-tech company showing off technology at the big tech trade show in Las Vegas and is back at CES 2025 with more autonomous tractors and other vehicles. This is not something we usually cover, but John Deere has a lot of data that is interesting in the big picture of tech. The message from the company is that there aren’t enough skilled farm laborers to do the work that its customers need. It’s been a challenge for most of the last two decades, said Jahmy Hindman, CTO at John Deere, in a briefing. Much of the tech will come this fall and after that. He noted that the average farmer in the U.S. is over 58 and works 12 to 18 hours a day to grow food for us. And he said the American Farm Bureau Federation estimates there are roughly 2.4 million farm jobs that need to be filled annually; and the agricultural work force continues to shrink. (This is my hint to the anti-immigration crowd). John Deere’s autonomous 9RX Tractor. Farmers can oversee it using an app. While each of these industries experiences their own set of challenges, a commonality across all is skilled labor availability. In construction, about 80% percent of contractors struggle to find skilled labor. And in commercial landscaping, 86% of landscaping business owners can’t find labor to fill open positions, he said. “They have to figure out how to do

2025 playbook for enterprise AI success, from agents to evals

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More 2025 is poised to be a pivotal year for enterprise AI. The past year has seen rapid innovation, and this year will see the same. This has made it more critical than ever to revisit your AI strategy to stay competitive and create value for your customers. From scaling AI agents to optimizing costs, here are the five critical areas enterprises should prioritize for their AI strategy this year. 1. Agents: the next generation of automation AI agents are no longer theoretical. In 2025, they’re indispensable tools for enterprises looking to streamline operations and enhance customer interactions. Unlike traditional software, agents powered by large language models (LLMs) can make nuanced decisions, navigate complex multi-step tasks, and integrate seamlessly with tools and APIs. At the start of 2024, agents were not ready for prime time, making frustrating mistakes like hallucinating URLs. They started getting better as frontier large language models themselves improved. “Let me put it this way,” said Sam Witteveen, cofounder of Red Dragon, a company that develops agents for companies, and that recently reviewed the 48 agents it built last year. “Interestingly, the ones that we built at the start of the year, a lot of those worked way better at the end of the year just because the models got better.” Witteveen shared this in the video podcast we filmed to discuss these five big trends in detail. Models are getting better and hallucinating less, and they’re also being trained to do agentic tasks. Another feature that the model providers are researching is a way to use the LLM as a judge, and as models get cheaper (something we’ll cover below), companies can use three or more models to

OpenAI’s red teaming innovations define new essentials for security leaders in the AI era

Join our daily and weekly newsletters for the latest updates and exclusive content on industry-leading AI coverage. Learn More OpenAI has taken a more aggressive approach to red teaming than its AI competitors, demonstrating its security teams’ advanced capabilities in two areas: multi-step reinforcement and external red teaming. OpenAI recently released two papers that set a new competitive standard for improving the quality, reliability and safety of AI models in these two techniques and more. The first paper, “OpenAI’s Approach to External Red Teaming for AI Models and Systems,” reports that specialized teams outside the company have proven effective in uncovering vulnerabilities that might otherwise have made it into a released model because in-house testing techniques may have missed them. In the second paper, “Diverse and Effective Red Teaming with Auto-Generated Rewards and Multi-Step Reinforcement Learning,” OpenAI introduces an automated framework that relies on iterative reinforcement learning to generate a broad spectrum of novel, wide-ranging attacks. Going all-in on red teaming pays practical, competitive dividends It’s encouraging to see competitive intensity in red teaming growing among AI companies. When Anthropic released its AI red team guidelines in June of last year, it joined AI providers including Google, Microsoft, Nvidia, OpenAI, and even the U.S.’s National Institute of Standards and Technology (NIST), which all had released red teaming frameworks. Investing heavily in red teaming yields tangible benefits for security leaders in any organization. OpenAI’s paper on external red teaming provides a detailed analysis of how the company strives to create specialized external teams that include cybersecurity and subject matter experts. The goal is to see if knowledgeable external teams can defeat models’ security perimeters and find gaps in their security, biases and controls that prompt-based testing couldn’t find. What makes OpenAI’s recent papers noteworthy is how well they define using human-in-the-middle

Stay Ahead, Stay ONMINE