Latest news with #EnergyandAI
Yahoo
16-05-2025
- Business
- Yahoo
A New IEA Report and the Iberian Blackout End Dreams of an 'Energy Transition'
It's no secret that the Republican's 'Big Beautiful Bill' plans to axe large swaths of mandates and billions of dollars in subsidies directed at achieving a so-called 'energy transition.' If that budget axe falls, it will be the proverbial third strike that puts to rest the idea that the U.S., never mind the world, will abandon fossil fuels. The other two strikes already happened. Strike two came last month with the Great Iberian blackout. Preliminary forensics make clear that over-enthusiastic deployment of unreliable solar and wind power was the fulcrum that put 55 million people in the dark for days. Few politicians will want to risk allowing something like that to happen again, anywhere. And, as the North American Electric Reliability Corporation keeps warning, blackout risks are rising here, and for the same reason. Reliability used to be the core feature of electric grid designs, before the rush to push an energy transition in service of climate goals. And strike one came a few weeks prior to the Iberian calamity with the release of a new report from the International Energy Agency (IEA) titled Energy and AI. That report sought to answer the question about how to reliably meet the surprising jump in power demands expected in the coming decade's boom in artificial intelligence (AI) data centers. Answering that also answers, even if not intentionally, the same question about meeting society's future demands. As the IEA report noted, just one large AI data center uses as much electricity as two million households, and myriads are planned. Thus, digital infrastructures will soon create demands equivalent to—reliably—powering hundreds of millions of new households. Spoiler alert: the IEA forecast shows fossil fuels continue to play a central role. However, since the IEA is the chief cheerleader for an energy transition, the executive summary of this latest report leads by observing that half the expected data center demand will be 'met by renewables.' Not until deep into that report's 300 pages does one find the candid observation that natural gas supplies the other half in the U.S., and coal fills that role in China. The IEA's framing of the answer is a glass-half-full view of a failed vision, especially considering that trillions of dollars have been invested so far in pursuing the transition goal. Meanwhile, counting on far more renewables to supply half of new demands means ignoring the political and economic headwinds for U.S. solar and wind deployments. Long before the November 2024 election, or the Iberian grid collapse, the IEA itself flagged what many now know: China has unprecedented global dominance in wind and solar supply chains. Setting aside tariff impacts, the kind of spending required to build-out transition hardware would entail a massive wealth transfer to China. At the same time, it has become obvious that jamming wind and solar onto grids wreaks economic havoc on consumers. The economic fallout is starkly visible in Germany and the U.K., for example, where aggressive transition policies are further along, and have rendered those nations 'poster children' for de-industrialization and energy poverty. Cost of power, however, is not the central issue for the data center industry. After all, it has deep pockets. The Magnificent Seven, collectively, have about a trillion dollars of cash on their books. Even if ratepayers and most businesses are price sensitive, Big Tech is not. Why not just pay the premium for wind and solar? The answer: The prime drivers in digital domains are reliability and velocity. It's vital to ensure that power is ready when construction is done, i.e., the very near future. And it's vital to deliver that power continuously and reliably once operations start. Thus, we're seeing an almost covert reliance on massive quantities of natural gas turbines in nearly all the announced projects from Meta's Louisiana site, to Amazon's Virginia sites, to Microsoft's sites, and to Open AI's Stargate site in Texas. As Nvidia executive Josh Parker said at a recent energy conference, the tech community wants 'all options on the table' because at 'the end of the day, we need power. We just need power.' Likewise, households from Iberia to Indiana. Of course, nuclear energy is on everyone's wish list, but there's no prospect that it will make a significant contribution during the coming decade of furious data center buildout. This doesn't mean Big Tech or the IEA are backing off climate pledges. Nor does it mean the climate debate is settled. Nor will we see any diminution in transition fervor from the climate-industrial complex. Likely that fervor heats up as the Trump Administration attempts to deliver on its promise to defund the panoply of climate-energy programs marbled throughout federal agencies. What it does mean is that whatever one believes about the science of the climate, the fact is that mandates and subsidies can't change the physics of energy systems. Systems that can deliver reliable power at the scales necessary for robust growth remain anchored in precisely the fuels the transitionists want to abandon.<> Mark P. Mills is the executive director of the National Center for Energy Analytics, and the author of The Cloud Revolution.
Korea Herald
15-05-2025
- Business
- Korea Herald
Neuchips Champions Power-Efficient AI at COMPUTEX 2025
As AI's growth faces energy challenges, the company is focusing on energy efficiency — with the capability to run a 14-billion parameter model on a single AI card and chip at just 45W TAIPEI, May 15, 2025 /PRNewswire/ -- As energy demand from AI data centers is set to surge worldwide, Neuchips, a leading Artificial Intelligence (AI) Application-Specific Integrated Circuits (ASIC) provider, is stepping up to address the challenge with a focus on power-efficient AI solutions at COMPUTEX 2025. Known for security and privacy, the company will showcase how its AI solutions save power while running large AI models. Join Neuchips at Booth I0601a in Taipei Nangang Exhibition Center Hall 1 for COMPUTEX 2025 (May 20-23, 2025). The company's efforts come just after the International Energy Agency (IEA) released its Energy and AI report in April. The findings project that global electricity demand from data centers will more than double by 2030, to around 945 terawatt-hours (TWh) — just slightly higher than Japan's power consumption today. Specifically for data centers leveraging AI, electricity demand is expected to more than quadruple by 2030. "Caught between the energy challenges of today's world and the potential of AI, power-saving AI solutions are no longer optional — they are a standard," said Ken Lau, CEO of Neuchips. "At Neuchips, we are constantly aiming higher for energy efficiency, and currently our Viper series AI accelerator cards can run a full 14-billion parameter model at just 45W, comparable to a standard light bulb. Innovation must now be focused on performance and efficiency." Collaborating for energy-efficient innovation Neuchips leads for its energy-efficient AI hardware solutions that deliver maximum security and privacy by running LLMs offline, empowering enterprises across a wide range of industries to leverage AI while preserving sensitive data. The company's product offering includes its N3000 chip series and Viper PCIe accelerator cards, which both fully support both Intel® and AMD CPUs and are compatible with Ubuntu and Windows. They also support a variety of LLMs, including Mistral Small 3 (24B), Llama 3.3 (70B), DeepSeek distilled models, Gemma 3 (4B) and more. At COMPUTEX 2025, Neuchips will showcase several collaborations through on-site demos: About Neuchips Neuchips is dedicated to developing energy-efficient AI acceleration chips that deliver innovative inference solutions for both edge computing and data centers. Through strategic collaborations with ecosystem partners, Neuchips is driving the democratization and sustainable advancement of AI technologies across industries.
Yahoo
18-04-2025
- Business
- Yahoo
Soaring U.S. Power Demand Will Need All Energy Sources
AI data centers and the reshoring of some manufacturing will drive U.S. electricity demand growth over the coming decade. The world's biggest economy will need all energy sources to ensure power demand is met. Natural gas is the biggest near-term winner of AI advancements, but renewables will also play a key role in powering the data centers of next-generation computing, analysts say. The U.S. has seen a surge in proposed data centers over the past year. As of July 2024, Wood Mackenzie had identified about 50 gigawatts (GW) of proposed data centers in America. This figure had doubled in a few months, and the proposed data centers were nearly 100 GW by January 1, 2025, according to WoodMac's estimates. Policy and regulation 'will need to reflect the reality that there is a need for all generation technologies,' according to Wood Mackenzie's analysts, as the natural gas boom alone cannot meet the soaring electricity data centers are set to account for almost half of U.S. power demand growth by the end of the decade, the International Energy Agency (IEA) said in its Energy and AI report last week. Driven by AI use, America will consume more electricity for data centers than for the production of aluminum, steel, cement, chemicals, and all other energy-intensive goods combined, according to the agency. The U.S. and other key power markets will tap a diverse range of energy sources to meet rising electricity needs from data centers. But renewables and natural gas are set to take the lead due to their cost-competitiveness and availability in key markets, including the U.S., the IEA said. By 2035, data centers are set to account for 8.6% of all U.S. electricity demand, more than double their 3.5% share today, BloombergNEF said in a new report this week. U.S. data-center power demand will more than double over the next decade, rising to 78 GW in 2035 from nearly 35 GW in Sachs, for its part, sees U.S. electrical power demand rising by 2.4% each year through the end of the decade, with AI-related demand accounting for about two-thirds of the incremental power demand in the country. 'The US is reaching a power demand inflection point due to the rise of energy-intensive artificial intelligence (AI), the need for more AI-ready data center capacity, and the reshoring of manufacturing,' Goldman Sachs analysts wrote in a report earlier this year. The higher power demand will be met by a diverse mix of energy sources. While renewable energy is growing in the mix of power generation sources, natural gas is best positioned to capture the majority of incremental capacity growth, while nuclear energy's impact is unlikely to be felt this decade, according to Goldman gas is set to provide 60% of the incremental generation capacity related to data center demand, while renewables are expected to provide the remaining 40%, the investment bank's analysts said. Renewable investments could drop by 30% or more in the coming years if the Trump Administration axes key incentive provisions in the Inflation Reduction Act (IRA), WoodMac says. 'But despite these concerns, buyers are showing a willingness to pay a premium for renewable energy,' WoodMac's analysts noted. Natural gas is winning big from the Trump pro-fossil fuel policies and the U.S. power demand surge, and gas is leading the efforts to power a large part of the extra electricity demand. Gas capacity orders jumped by 146% in North America last year—the most in nearly a decade, and gas turbine orders increased by 36% year-on-year in 2024, said WoodMac. 'But gas alone cannot meet the rising electricity demand,' it added. Demand is rising, but there are uncertainties about how high the growth would be and how fast the data centers will be built. BloombergNEF (BNEF) has a relatively conservative forecast about U.S. data center buildout. This is not due to skepticism about AI's market potential 'but acknowledges real-world deployment challenges such as securing key elements (land, power, permits) and navigating the complex construction process,' said BNEF's Helen Kou, Head of US Power, and Nathalie Limandibhratha, Senior Associate US Power. The research firm has estimated that the development of a data center in the United States would typically take about 7 years—from initial planning to full operation. This timeframe includes 4.8 years in pre-construction and 2.4 years for construction. Still, the U.S. will be the world's most important market in power demand growth for data centers, BNEF said in a separate report this week. 'Over the next five years, demand from US data centers could outpace even electric vehicles' incremental demand, driven by the surge in AI training workloads that require significant compute capacity and highly energy-dense infrastructure,' according to BNEF. Amid surging power demand, natural gas will play a crucial role in meeting incremental U.S. electricity consumption. Renewables and potentially nuclear in the longer term will also provide the additional power needs. By Tsvetana Paraskova for More Top Reads From this article on
Forbes
14-04-2025
- Business
- Forbes
AI's Energy Dilemma: Can Tech Drive A Sustainable Energy Future?
An artificial Intelligence booth by Rittal Ltd. at the Hannover Messe 2025 trade fair in Hannover, ... More Germany, on Monday, March 31, 2025.. Photographer: Krisztian Bocsi/Bloomberg Artificial intelligence is no longer just the future—it's now the engine of the global digital economy. AI is infiltrating every sector, reshaping industries, and rewriting the rules of business. But behind this technological shift lies an uncertain reality: AI consumes enormous amounts of energy. And unless reined in, this demand could overwhelm global electricity systems. According to Energy and AI, the latest report from the International Energy Agency (IEA), the electricity consumed by data centers—the nerve centers of the AI revolution—could more than double by 2030, surpassing 1,000 terawatt-hours (TWh). That's conservative against a recent Goldman Sachs forecast that global power demand from data centers will increase 50% by 2027 and by as much as 165% by the end of the decade. At the heart of this surge are data centers—the physical infrastructure that powers AI's rapid computations and learning cycles. The rise of AI workloads, including large language models and generative platforms, is driving an exponential increase in power consumption that governments, utilities, and tech giants can no longer afford to ignore. 'There is no AI without electricity,' warned Dr. Fatih Birol, Executive Director of the IEA at the launch of the report. 'The energy industry is caught unprepared for the speed and demands of the AI industry.' This predicted surge in energy use poses a defining challenge. But there is also a clear opportunity: if harnessed correctly, AI could become a powerful tool for cutting energy consumption and carbon emissions - by up to around 5% of global emissions. The report doesn't just warn of growing demand—it highlights how AI itself could transform energy efficiency and reshape our response to climate and infrastructure pressures. AI is already playing a pivotal role in making energy use smarter and more sustainable. In power grids, it enhances forecasting and fault detection, allowing for better integration of renewable energy and between 30-50% fewer outages. In industry, it's optimizing operations to the point of saving as much energy as Mexico uses each year. AI-powered logistics could cut fuel use across transport systems, saving the equivalent of energy used by 120 million cars. And in the built environment—often slow to modernize—AI-driven controls for lighting, heating, and cooling could save 300 TWh annually, the same as Australia and New Zealand's total electricity use. In this way, AI embodies a paradox: it is both the driver of rising energy consumption and a key to reducing it. The future isn't only being shaped by demand—it's also being disrupted by innovation. In early 2025, Chinese startup DeepSeek introduced open-source AI models that match the performance of market leaders at a fraction of the energy cost. Its V3 model was trained for just $5.6 million, while reports put the training cost of GPT-4 at around $100 million. Breakthroughs like these are rewriting assumptions about the energy requirements of advanced AI and pointing to a more efficient path forward. Even Big Tech is reconsidering its approach. While companies like Microsoft, Meta, and Apple have announced massive investments in data centers, analysts are reporting a trend of scaling back. In the U.S. and Europe, Microsoft is reported to have paused several data center expansion projects. The move, spurred by efficiency gains and shifting business models, suggests the future may not be as energy-hungry as once feared. DeepSeek's breakthrough in cutting computational loads hints at a new era of leaner, greener AI. That shift may be essential. 'Almost half of the U.S. electricity demand growth between now and 2030 will be driven by data centers,' Birol said. 'To put it in context, the electricity used for AI data centers will exceed the consumption of the chemicals, steel, aluminum, and cement industries combined.' That's not just a shift—it's a wholesale reordering of industrial energy dynamics. It suggests that AI could become the single most dominant industrial force shaping the future of electricity grids, energy investments, and national energy security strategies. There's growing consensus that the solution lies not in curbing technological progress, but in building smarter systems. Jakob Jul Jensen, Head of Business Development for Data Centers at Danfoss, believes the future hinges on better design. 'Energy efficiency must be prioritized and integrated into every layer of data center design and operation,' he says. That means smarter cooling systems, AI-driven power management, and perhaps most overlooked, heat reuse. According to the IEA, the waste heat from data centers could meet 10% of Europe's space heating needs, provided it's redirected into homes, buildings, and local industries, impacting the energy system overall. As AI investment accelerates, access to secure, affordable, and clean energy is becoming a strategic differentiator for nations. Countries able to deliver on all three will lead the AI-driven economy and attract investment from the world's most valuable companies. 'The availability and cleanliness of electricity is becoming a strategic advantage,' says Birol. 'Countries that can provide secure, clean, and affordable electricity will be one step ahead in attracting AI investment. Those that fall short risk being left behind." That's why the IEA is taking this moment seriously. For the first time, it's integrating AI into its annual World Energy Outlook forecasts, recalibrating how governments and industries prepare for the road ahead. 'Everything around, above, and under energy is our business,' said Birol. 'We couldn't ignore this. Data always wins—and our job is to put that data on the table for governments and industry to act.' And act they must. Without coordinated policies and innovation-driven strategies, AI could lock in energy inefficiencies that will be far more expensive—and disruptive—to fix later. But by embracing AI-powered grid management, systemic energy integration, and innovations in efficiency and reuse, there is a real chance of turning this challenge into an opportunity driving the development of a smarter, cleaner, and more resilient global energy system. Artificial intelligence is not just the next great industrial consumer of energy—it's also poised to be its most transformative innovator. Whether the world leans into this dual identity will define not only the trajectory of technological progress, but the sustainability of the global energy economy itself. The question is no longer if AI will reshape the future of energy, but how sustainably we allow it to do so.
