Kamil Jankielewicz’s Post

Have you wondered how much power is required to keep your social media accounts running? Or what is really powering cloud- computing? This article from The Economist shares interesting steps taken by the tech giants particularly towards garnering clean energy.   Cloud computing and AI needs vast amount of electricity. JPMorgan Chase calculates that Microsoft, Alphabet, Meta consumed 90 terawatt-hours (TWh) of electricity in 2022, as much as Colombia. Yes, that’s a lot! So, the tech titans don’t just need a lot of power, they also want theirs to be clean. Microsoft and Brookfield recently announced a deal to build 10.5GW of renewable capacity in America and Europe by 2030. The arrangement is meant to enable Microsoft to meet its pledge to have 100% of its electricity use, 100% of the time, come from zero carbon sources by 2030. One problem is that data centers tend to consume power at a steady rate, including when the sun is not shining nor the wind blowing. So tech firms are also thinking of ways to make data-processing more flexible… It involves a combination of microgrids (which can run independently but also exchange energy with others nearby), batteries and advanced software in order to enable shifting less time- sensitive tasks, such as training AI models, to periods of fallow demand. Google, meanwhile, is dabbling in geothermal energy. It has signed the first- ever corporate deal to develop “enhanced” geothermal power with Fervo. Google, Microsoft and Nucor have announced that they will aggregate demand and jointly offer contracts to clean-energy project, both early-stage commercial ones and entirely novel “first-of-a-kind” ventures. Sam Altman has invested in Helion, a nuclear-fusion startup, and Exowatt, a startup developing solar modules that can act as both electricity generators and thermal-storage batteries. All these wagers may seem fanciful. Then again, 18 months ago so did the idea that ab AI would write essays or paint like a human. Copyright owned by Economist Newspaper Limited.   https://lnkd.in/dEVB2t8v

The insightful article by The Economist highlights the aggressive pursuit of #AI computing power by tech giants Alphabet Inc. Amazon and Microsoft They collectively invested $40 billion in AI-focused data centers between January and March. Meta also forecasts up to $40 billion in capital expenditure this year due to AI projects. With Microsoft expected to spend even more, AI's immense processing power requirements drive tech giants' significant investments in energy-demanding data centers. International Energy Agency (IEA) predicts that data centres (including those dedicated to ai and equally energy-hungry cryptocurrencies) will gobble up more than 800twh globally in 2026, more than double the amount in 2022 . Big tech companies are increasingly investing in green energy to power their data centers. With a focus on clean power, these deep-pocketed companies are shifting from green power-purchase agreements to more direct investments in renewable energy projects. Their efforts contribute to the projected rise in annual grid investment required for global electricity decarbonization, from $300 billion in 2022 to over $800 billion in 2050. Microsoft and Brookfield Asset Management announced a partnership to develop 10.5GW of renewable energy capacity by 2030, aiming to meet Microsoft's 100% zero-carbon electricity commitment. This deal could power 1.8 million homes and contribute significantly to grid decarbonization. Major tech companies like AWS, Google and Microsoft are actively seeking alternative energy sources for their power-hungry data centers. AWS acquired a nuclear-powered data center, while Google is exploring enhanced geothermal energy with Fervo. Microsoft and Google are collaborating with Nucor to secure promising technologies such as long-duration energy storage, clean hydrogen, and next-gen nuclear energy. Amid global energy uncertainties, tech giants strive to meet the soaring electricity demands of AI. One of the latest articles published by Lloyd's List suggests that the advancement of AI will not only improve market visibility and operational efficiency but also drive growth in the shipping industry. The growing demand for electricity to power AI data centers may lead to an increased need for natural gas, creating new opportunities in the #LNG #shipping sector and fostering overall industry expansion. As tech giants revolutionize the energy landscape with their AI power grab, the shipping industry must stay attuned to these developments and adapt accordingly. By embracing this shift towards A.I. and clean energy sources, the Maritime sector can maintain its competitiveness while contributing to a #greenerfuture. #betterworld #decarbonization Pls refer to the link for details : https://lnkd.in/eMPJhWdg from The Economist https://lnkd.in/eFZjeguC

⏭ 'Big tech’s great AI power grab' 'Alphabet, Amazon and Microsoft are on the hunt for new energy sources' ⏭ 'Amazon and Microsoft—the world’s cloud-computing giants—collectively invested $40bn between January and March, most of it in data centres equipped to deal with growing artificial-intelligence (ai) workloads' ⏭'Meta, which does not have a cloud business but does run a data-hungry social-media empire, said its capital expenditure could reach $40bn this year as a result of ai-related projects' ⏭ 'The comparison with the famously capex-happy energy industry is apt not just because of the sums involved. ai needs vast amounts of processing power.' "Dominion Energy, one of America’s biggest utilities, said that data-centre developers now regularly ask him for “several gigawatts” (gw). Dominion’s total installed capacity is 34gw." ⏭ 'the International Energy Agency (IEA), an official forecaster, to predict that data centres (including those dedicated to ai and equally energy-hungry cryptocurrencies) will gobble up more than 800 twh globally in 2026' ➡ "And not just any power will do. The technology titans want theirs to be clean." ⏭ 'Georgia Power, which had managed to fast-track the approval of 1.4gw of new fossil-fuelled generation by pointing to rising demand from data centres, that its members would build fewer of these in the southern American state if the utility spewed extra carbon' ➡ "BloombergNEF, an information firm, reckons that annual grid investment needed to fully decarbonise global electricity by 2050 will need to rise from about $300bn in 2022 to $600bn in 2030" ⏭ 'Deep-pocketed giants have already been the biggest force behind green “power-purchase agreements”, which helped kickstart America’s renewables boom by persuading utilities and other investors to build wind and solar farms.' ➡ "On May 1st Microsoft and Brookfield, one of the world’s biggest infrastructure investors, announced a deal to build 10.5gw of renewables capacity in America and Europe by 2030" An insightful and engaging piece from @theeconomist, read in full here: https://lnkd.in/gqJtPgzy

Good to see continued interest and reporting on the impacts from LLMs, in this case by Camilla Hodgson at the Financial Times. It's difficult to find sources willing to discuss the actual financial and environmental costs of the LLM arms race, which isn't surprising considering the scale of the impacts and amount of money at risk, though I think it's unwise of the industry not to be transparent. One of these days (soon, I hope), we'll get to the phase of discussing much smarter and more responsible alternatives to LLMs, like our KOS, which can reduce about 90% of the financial and environmental costs off the top due mostly to our system design, which is ultra-focused on high quality data and precision data management. Our next generation technology, now 15 years in R&D, is the synthetic genius machine (SGM), which is a combination of neurosymbolic AI with new types of compression and encryption. While both generations of technology require significant computing resources, it's nowhere near the vast waste in brute force LLMs, doesn't infringe on copyright, and both were designed with safety-critical system principles. The only reason the public is even aware of LLMs is because LLM firms did not follow safety-critical principles. Had they done so LLMs would not have been released to the public so prematurely if at all as LLMs are inherently unsafe as currently deployed -- interactive consumer bots for the general public. That they also require vast amounts of compute power, electricity, and water is directly related as the models are based on scale and inefficiency, which is one of the reasons they are also unsafe. LLMs are intended to use as many resources as possible to raise the bar of entry beyond anyone but incumbent Big Tech enablers. That's the hard, cold truth of the matter, and frankly why so few are willing to discuss it in public. Hundreds of billions of USD in industry spend is at stake, and over $5 trillion in market cap. That's a lot of incentive to do the wrong thing. It's also an extreme example of the power of competition -- we lack such perverse incentives and conflicts, by design.

The potential impact of adopting of #ai applications is exciting. Generating the huge amounts of #electricity required to do this will be daunting. Will enough #cleanpower be available to meet this need? #daretocare about building out #renewable #energy sources and #regulatoryreform

Can renewable energy keep up with Big Tech’s thirst for AI? The cost of annual investment to fully decarbonize the grid will hit $600bn a year by 2030 and $800bn by 2050. And companies like Meta and Alphabet are reaching into their deep pockets to foot the bill. In the meantime, #AI is rapidly increasing the power data centers are consuming and #renewableenergy supplies aren’t always consistent enough.

"Data is the new oil" A quote used extensively over the past 15 years across all industries during sales pitches, on stages presenting future trends and probably in board rooms during intense discussions how to pivot from physical to a digital business. In some cases the quote was accurate, in some cases not. In terms of energy consumption lets hope not. Or maybe this is what Clive Humby OBE meant all along with his famous quote? Most likely not. The current AI craze comes with a potential tremendous cost for society which we yet dont fully know the consequences of. It´s great to see local,regional and global media´s current attention and dedication towards the subject these days, because we need more awareness and understanding of the subject moving forward. Global tech companies spent 20+ years before getting serious about privacy during the internet boom, this time around we dont have the same amount of time to get control of our energy consumption with the exponential usage of artificial intelligence. Kicking the can down the road which was the common strategy for privacy during the internet boom cant and wont work this time around. The Guardian

This week I had the privilege of representing the electric sector at the White House to discuss how we can meet the energy demands of AI’s exponential growth. The discussion brought together key AI leaders like Jensen Huang (Nvidia) and Sam Altman (OpenAI); executives from hyperscalers including Microsoft, Google, Meta, and AWS; public sector leaders such as Secretary Granholm, Secretary Raimondo, Lael Brainard, John Podesta, and Jake Sullivan; and energy leaders like myself, Arshad Mansoor (EPRI), and Calvin Butler (Exelon). This meeting underscored the importance of energy infrastructure as a foundation for U.S. leadership in AI. Renewable energy, battery energy storage, and innovations like Grid Enhancing Technologies (GETs) are essential to meet the energy demands of AI-driven data centers—on time and sustainably. As Jensen Huang aptly remarked, "this industry is going to be producing intelligence, and what it takes is energy." https://lnkd.in/erK728Dx

Extremely insightful (and interactive) article on global data consumption within AI... Datacentres are dominating the landscape in previously rural areas 🗄🗄🗄🗄🏠 🌳 🗄🗄🗄🗄🗄🌳🗄🗄🗄🗄🗄 🏠 Some countries' AI power-drain will outstrip total green energy production as each country takes part in the 'AI race' (to where though? 👀 ) NVIDIA's status as the world's most valuable company 💵💵💵 sees no signs of abating as their H100 chips as seen as the de facto standard for AI datacentres Where will we get to with data sovereignty 👑 and regulation? It's been reported that Apple's AI (due to release in Q3 in US) is nowhere near an EU launch due to GDPR https://lnkd.in/emsTETGV