The Environmental Impact of AI: Balancing Innovation and Energy Consumption

Energy-guzzling AI derails tech firms' climate goals

As major technology companies race to take advantage of the artificial intelligence (AI) boom, the energy-guzzling technology is derailing their efforts to rein in emissions and become carbon neutral or negative. Google revealed that its greenhouse gas emissions rose 13 percent in 2023 and 48 percent since 2019, clashing with its goal of becoming net-zero by the end of the decade. The primary culprit? AI. "As we further integrate AI into our products, reducing emissions may be challenging due to increasing energy demands from the greater intensity of AI compute," Google wrote in its annual environmental report. Microsoft has seen its emissions jump 29 percent since 2020, according to its annual sustainability report released in May. The tech giant, which aimed to be carbon negative by 2030, similarly cited AI as the cause of its growing emissions. "In 2020, we unveiled what we called our carbon moonshot. That was before the explosion in artificial intelligence," Brad Smith, Microsoft's vice chair and president, told Bloomberg at the time. "So, in many ways the moon is five times as far away as it was in 2020, if you just think of our own forecast for the expansion of AI and its electrical needs," he added. AI requires significantly more energy than other processes. A single ChatGPT request uses 2.9 watt-hours of electricity, while a typical Google search uses just 0.3 watt-hours, according to the International Energy Agency (IEA). Generating images with AI requires even more energy than generating text. On average, image generation uses over 60 times the energy of text generation, according to a study by researchers at Carnegie Mellon and the AI startup Hugging Face. This translates into more emissions. Using the popular text-to-image generator Stable Diffusion XL to create 1,000 images produces the same amount of emissions as driving an average gas-powered car 4.1 miles, the study found. Despite Microsoft's rising emissions, Smith has argued that "the answer is not to slow down." "We fundamentally believe that the answer is not to slow down the expansion of AI but to speed up the work needed to make it more environmentally friendly," he said. "I guarantee there's one way to fail: It's to give up." Environmental advocates aren't convinced. Michael Khoo, the climate disinformation program director at Friends of the Earth, argued that tech firms are very focused on "AI for good" and not "AI for bad." "So much of the AI discussion coming from Silicon Valley is centered on how AI is going to save humanity, whether you're looking at words from [OpenAI CEO] Sam Altman or from DeepMind or from [Elon] Musk, all of them use this as a talking point," Khoo told The Hill. "I can't predict the future ... but currently, today and tomorrow, it's doing really, really bad things for the planet," he added. With the rise of AI, as well as slowing efficiency gains, data center energy demand is rising rapidly. The IEA estimates that global electricity consumption by data centers could double between 2022 and 2026, adding the equivalent of "at least one Sweden or at most one Germany." By 2030, Goldman Sachs Research estimates that data center power demand will have grown 160 percent. The accompanying uptick in carbon dioxide emissions could have a social cost between $125 billion and $140 billion. With this uptick in demand, the question becomes whether that additional energy comes from carbon-free sources, like wind and solar, or "brown" sources, like natural gas and coal, said Benjamin Lee, a professor of electrical and systems engineering and computer and information science at the University of Pennsylvania. This is complicated by the fact that data center construction is currently outpacing renewable energy installation, Lee noted. "When we're looking at how much renewable energy is being put on the grid in any given year, the data centers could consume all of it and then some," he told The Hill. "So, there's a mismatch between the growth rates that we've seen." Khoo emphasized that the U.S. is currently "in the middle of a very tough build out of our renewable energy infrastructure." "There is no excess supply of renewable energy in America right now," he said. As a result, utility companies are increasingly looking to fossil fuels to keep up with growing energy demand. "Utilities don't see how they can meet that, in the short term at least, without more fossil fuel, more thermal plant operation," former Energy Secretary Ernest Moniz said at an energy conference in March. The use of fossil fuels -- coal, oil and gas -- is the main driver of climate change. This warming has led to an increase in the frequency and intensity of extreme weather, including flooding and droughts, in addition to heat waves. These extremes are expected to worsen if the planet continues getting hotter, especially if key "tipping points" that change the system and are difficult to reverse -- such as the melting of the Greenland ice sheet and Arctic permafrost -- are reached. Overall, the world's average surface temperature has warmed by more than 1.1 degrees Celsius (about 2 degrees Fahrenheit), according to a United Nations report last year that warned of a "rapidly closing window" to combat warming. "This couldn't come at a worse time," Khoo said. "We're at a point where we can just all look outside in America today and most days for the greatly expanding effects of climate change."

These 2 Innovative Tech May Help Reduce A.I.'s Carbon Footprint -- If Data Centers Are Ready

A crop of startups are developing affordable technologies to reduce the growing carbon footprint of A.I. data centers. Earlier this month, Google (GOOGL) revealed that the company's greenhouse gas emissions have climbed nearly 50percent percent since 2019 thanks, in part, to A.I. The search giant's annual environmental report is the latest addition to a growing body of research that shows how energy and resource-intensive it is to power A.I. workloads in data centers, which could hinder companies from reaching their net-zero emission goals. Sign Up For Our Daily Newsletter Sign Up Thank you for signing up! By clicking submit, you agree to our <a href="http://observermedia.com/terms">terms of service</a> and acknowledge we may use your information to send you emails, product samples, and promotions on this website and other properties. You can opt out anytime. See all of our newsletters While some tech giants are turning to renewable energy -- including potentially nuclear -- to meet A.I.'s energy demand, other companies claim they have devised new, more affordable technologies to reduce the carbon footprints of data centers. They come as Microsoft, Amazon, and other Big Tech companies spend billions building data centers across the U.S. -- an initiative that could double the country's electricity consumption by 2030, according to the Electric Power Institute. "As of now, there is so much more conversation around data center energy consumption than ever before," Abhijit Sunil, a senior analyst at Forrester who researches green IT and data centers, told Observer. One possible solution companies are exploring is immersion cooling. Traditionally, data centers are cooled using air conditioners and fans, which require tons of energy. An immersion cooling system, however, submerges data center equipment into a special liquid that absorbs the heat. That could reduce as much as 90 percent of the energy needed for cooling, which makes up to 40 percent of a data center's total energy consumption, according to Peter Boulin, the CEO of Green Revolution Cooling, an immersion cooling provider. As data centers implement chips like Nvidia (NVDA)'s GPUs -- which require massive amounts of electricity -- to power large A.I. workloads, immersion cooling could be a way to make A.I. usage less energy-intensive. The "exponential increase" in power at the chip level is "fundamentally changing the way data centers are being designed and constructed," Joe Capes, the CEO of LiquidStack, another company that creates immersion cooling systems, told Observer. "A majority of existing data centers are not well equipped to support these workloads." In addition to new cooling systems, some startups are making software that enterprises can use to optimize their current workloads to be less energy-intensive. Incredibuild, an Israeli company, claims its platform can tap into "idle computing," or power not being used by a program, and convert it into a usable source of energy for other tasks in the software development process. As a result, data centers can extract more electricity from their existing power source to support those workloads instead of using an additional source, which could reduce overall emissions, Dori Exterman, Incredibuild's chief technology evangelist, told Observer. Exterman explains that this technology, which he said is used by Amazon Web Services and Intel, helps reduce carbon emissions now that companies are beginning to adopt A.I. assistants like Microsoft Copilot that software engineers and employees on other teams can use to boost their productivity. "The future of dealing with all this demand A.I. generates is with more efficient software," Exterman said. Making data centers greener comes with challenges Despite their potential to reduce A.I.-driven emissions, these technologies are still relatively new products that have yet to be fully explored by existing data centers. In terms of immersion cooling, a "lack of standards" and "best practices" around deploying the technology across data centers may hold companies back from implementation, Forrester analyst Sunil explained. "It's an emerging technology, and everything evolving with it, including standards and technical know-how, is a big part of where we are seeing challenges now." Another challenge is the cost of adopting a new technology. While some immersion cooling companies say the technology is more cost-effective than traditional cooling measures, one says that major data warehouse providers may be skeptical about whether it's worth the effort to retrofit their existing infrastructure solely to address the energy problem. "It has to make economic sense for them," Boulin, CEO of Revolution Green Cooling, told Observer. "If you're a Dell or an HP, given the scale of your companies, you need to see a massive market before you allocate engineering resources to designing a platform specifically for immersion." Incredibuild's Exterman said getting data centers to adopt such new technology is a "matter of education" about what the product can deliver. But as A.I. giants like OpenAI seek more computing power to improve their tools, the chief technology evangelist believes that energy optimization software and other carbon-reduction tools must be deployed at a massive scale to meet the growing energy demand -- all while reducing emissions. "The industry is trying to catch up," Sunil said.

The evolution of AI: A bane for the energy industry, economy, and climate?

It looks like the constant rise of AI can be a double-edged sword for today's digitalised world With the larger adoption of artificial intelligence (AI), businesses seem to have automated most of their functions. This eventually is expected to threaten the stability of the energy industry, economy, and climate due to its high energy consumption. Experts suggest that the tech sector's decarbonization goals are challenged by AI's power demands. As reported by Google, there has been a 48 percent increase in carbon emissions over the last five years. As AI picks up speed, the rapid increase in power consumption can pose a direct threat to the tech sector's ability to follow its decarbonization plans. The threatening market of AI AI's energy demand is growing at an alarming rate, between 26% and 36% annually, as reported by the World Economic Forum. By 2028, this could push AI's energy consumption beyond countries like Iceland. In addition to this, each AI-powered internet query can consume about ten times more energy than traditional internet searches. It looks like the constant rise of AI can be a double-edged sword for today's digitalised world. While it unlocks incredible potential of automating tasks and other advancements, its energy footprint can be a cause for concern. A recent Goldman Sachs report predicts a 160% increase in data centre power demand by 2030, with AI being a key driver. This means that data centres are potentially consuming 8% of US power by 2030, compared to just 3% in 2022. The dark side of AI The variability and unpredictability of renewable energy pose significant challenges for AI applications. For example, solar and wind power generation are inherently intermittent and dependent on weather conditions. AI algorithms need to sync with the dynamic nature of these energy sources to optimize production and distribution effectively. However, developing accurate predictive models accountable for fluctuations in renewable energy generation remains a formidable challenge. This drawback eventually impacts the reliability and stability of energy grids. The energy industry already seems to face several challenges when it comes to the endorsement of rapid digitalization. It is believed that the energy sector's software architecture is much older than that of other sectors such as finance. The energy industry also needs to ensure that any form of change is compatible with 'on-the-ground infrastructure' located across several places. This makes the implementation of modern technology more costly and difficult, especially for smaller companies. Moreover, combined with the complexity of training models and limitations in accessing adequate computational power, the cost-benefit analysis of energy AI requires further investigation. Looking from an economical perspective, automation with AI and increased connectivity has the potential to make the energy sector vulnerable to cyber threats. For example, aging and unprotected points in the electric grid can be exploited to gain access to the entire ecosystem. The energy sector can be vulnerable to cyberattacks, with each average 2020 attack costing about $6.4 million in damages, according to a study by MIT. According to a report by the BBC, AI-powered services can involve considerably more computer power and electricity, in comparison to standard online activity. Another recent study by scientists at Cornell University suggested that generative AI modules such as ChatGPT can use up to 33 times more energy than computers running task-specific software. Industry reacts Industry experts suggest that the widespread adoption of AI and its utilization in the field of energy and other domains may result in significant job displacement. As AI systems become more advanced and widespread, their energy consumption skyrockets. This rise in energy consumption can eventually strain power grids and increase reliance on fossil fuels. Furthermore, "This surge in demand exacerbates greenhouse gas emissions, undermining global climate goals. Economically, the escalating costs of energy required to support AI infrastructure could drive up operational expenses for businesses, potentially slowing innovation and growth," Jaspreet Bindra, founder, Tech Whisperer, explained. Additionally, "A significant amount of heat is generated due to data centers and cooling equipment needed for the running of AI. This could be potentially contributing to climate change. AI could create socio economic imbalances, as fossil fuels have been the primary source of power for centuries," Sirajuddin Ali , founder and CEO, Malitra India, said. Despite their ambitious sustainability commitments, big tech companies such as Microsoft, have seen their carbon emissions continue to rise partly due to AI's energy requirements. Moreover, "Big tech companies' expansion of data centers to support AI has contributed to increased infrastructure emissions. This highlights the complex nature of achieving net-zero goals while scaling AI capabilities," Agam Chaudhary, CEO, Two99, highlighted. Moreover, " The key to extract the full potential of AI is the responsible implementation and collaboration between industry leaders and policymakers.This will help to align AI development with sustainability goals and economic stability" Prashant Singh, co-founder and CEO, Blue Planet, concluded.

For The Love Of AI

Artificial Intelligence (AI) has undeniably taken the world by storm. Slap the AI label on anything, and it's bound to get noticed. Like any new technological advancement, the use and misuse of this technology are rampant. We've all received an email or seen a work product that screamed "AI-generated." I've personally sent work back because it was just a series of well-structured words in English, devoid of any real depth. We've all seen those AI-generated sales and marketing emails that are meaningless and clog our already full mailboxes. Here's an easy pro-tip for those guilty as charged: put your real name instead of [Your Name], and delete the prompt response from the AI output before hitting send! AI has the potential to revolutionize how we work, communicate, and create, but its misuse can lead to superficial and ineffective communication. Consider the case of OpenAI's GPT-3 being used to generate automated news articles. While some articles were coherent, others were lacking depth and context, highlighting the importance of human oversight. While AI can assist in organizing thoughts and generating content, it cannot replace the authenticity and emotional intelligence of human input. Properly harnessed, AI can be a powerful tool, but it must be used thoughtfully and ethically to add real value. The current AI tools, let's call them AI 1.0, are often marketed with a powerful freemium version. For a small monthly fee, you get a more powerful paid version that can read attachments, generate graphics, and more. Regardless of the small amounts being collected for the premium version, it's hard to see how this business model will become profitable at $20/month. Currently, AI requires computing power from thousands of servers housed in data centers, which need massive amounts of electricity. Something will have to change: either processors will need to consume less power, we will need to find vast amounts of clean power, or demand will need to be curtailed. The root cause of AI misuse and abuse is that people aren't starting with their own original thoughts. As Henry Ford famously said, "Thinking is the hardest work there is, which is probably the reason why so few engage in it." AI chat tools excel at taking original ideas and structuring them into cohesive paragraphs using vast databases to predict how sentences and paragraphs should be structured. However, AI cannot express original thoughts, genuine emotion, or emotional intelligence. AI results can be significantly enhanced when directed to a data set relevant to your organization or contextual to your queries. This is where the next generation of AI tools, let's call them AI 2.0, will start to have a massive business impact. With advancements in processing technology over the past decade, data center workloads nearly tripled between 2015 and 2019, yet power demand remained flat at 200 terawatt-hours per year due to more efficient servers and processors. However, this changed with the advent of AI. According to Goldman Sachs, a ChatGPT query needs nearly ten times as much electricity as a Google search query. For perspective, Nvidia literally sells half a million of its H100 AI GPUs per quarter, each consuming up to 700W of power when operating, which is more than an average US household. The same Goldman Sachs research estimates that AI could increase data center power consumption by as much as our total consumption in 2019, around another 200 terawatt-hours per year! Where will that power come from? There are several ways to quench this thirst for energy: Both Microsoft and AWS are building mini nuclear power plants to power their AI initiatives. Nuclear power is a potential power source for AI because it can meet the large amounts of electricity that AI data centers require. As the modern economy becomes more integrated with AI technologies, so will their demand for energy and scale. Some say that nuclear power could help avoid exacerbating climate change. As of July 2024, tech companies like Amazon Web Services are in talks with nuclear power plant owners to provide electricity for new data centers. However, there are some uncertainties to consider, including: Pundits are talking about the ethical implications of using AI, such as plagiarism, privacy concerns, job obsoletion, and universal income. While AI is powerful when used correctly, it can sometimes increase workloads instead of decreasing them. A pressing issue is the creation of AI-generated deep fakes and phishing emails, which generate worthless and dangerous workloads for society. Another problem is the proliferation of overly wordy emails and long documents. Generative AI should improve communication, not think for you. One of the most crucial skills right now is learning to write your original thoughts, prompt effectively, and genuinely edit the final AI version to reflect your original intent and voice. With non-renewable energy sources being depleted because AI is used to create "novels" that could be summarized in 140-character statements, we're hastily building nuclear power plants to satisfy our addiction to instantaneous gratification. While we should support safe nuclear energy, we must also use our brains and faculties wisely. AI's ethical implications extend beyond convenience and efficiency; they impact our energy consumption, security, and the authenticity of human expression. It's essential to approach AI use thoughtfully and responsibly to harness its true potential.

Council Post: Powering The Future Of AI: Addressing The Looming Energy Challenge

The conversation surrounding the growth of artificial intelligence (AI) often emphasizes advancements in large language models (LLMs) and data capabilities. Investors frequently track things like GPU sales, noting growth like Meta's commitment to acquiring 350,000 H100 GPUs from Nvidia by year-end, surpassing Nvidia GPUs purchased by Microsoft, Google and Amazon combined over the same time period. While these metrics provide a snapshot of short-term developments (within the next year), the mid-term (next three years) and long-term (next five years) perspectives reveal a more critical challenge: securing adequate electrical power to sustain AI's future growth. The power requirements for AI data centers are expected to increase significantly over the next five years. By 2030, the power consumption of data centers in the United States alone is projected to reach 35 gigawatts (GW), nearly double the 17 GW consumed in 2022. This surge is driven by the substantial computational power and cooling capabilities needed for AI and machine learning. Because of this, the technology sector faces a significant power challenge as it invests billions into AI advancements. Generative AI models, with their intensive computational processes, demand far more power than other technologies. Although Nvidia and other GPU producers are contributing to the solution by producing GPUs with higher performance per watt, the sheer volume of GPUs produced globally and deployed in data centers underscores the magnitude of the power requirements. Leading technology companies are proactively planning for an AI-driven future by investing heavily in large-scale, predominantly renewable energy sources. * Google has invested in Kairos Power, a company developing high-temperature fluoride salt-cooled reactors. This partnership aims to leverage advanced nuclear technology to create cleaner energy solutions. * Microsoft, through its founder Bill Gates, has a significant interest in TerraPower. TerraPower is developing advanced nuclear reactors, including the Natrium reactor, which aims to provide safer, more efficient nuclear energy. In 2021, TerraPower selected Kemmerer, Wyoming, for its advanced nuclear reactor demonstration plant. * Amazon itself has not directly invested in nuclear power; however, it benefits from partnerships with energy providers who are exploring nuclear options as part of a diversified clean energy portfolio. * Meta has signed long-term agreements to support the construction of solar projects, such as the 330 MW solar projects with Adapture Renewables in Illinois and Arkansas, and the 349 MW Kelso Solar Project in Missouri. These investments are integral to the long-term strategies of these tech giants, essential for supporting AI's growth potential. However, these companies can only influence power requirements to a limited extent. The U.S. currently grapples with an aging power grid infrastructure, highlighted by events such as the 2021 Texas winter storm, which exposed vulnerabilities in the natural gas supply. Supporting an additional 35 GW of power demand requires substantial infrastructure, including transmission lines, substations and energy storage systems. The increase in demand due to AI data centers, which could reach this level by 2030, highlights the need for significant upgrades and expansions in the electricity grid Because of this and in addition to the work that's being done in the private sector, the U.S. government is actively investing in grid modernization. The Department of Energy (DOE) leads the Grid Modernization Initiative, committing substantial funds to research and development. This initiative aims to create a resilient, reliable, and flexible power grid capable of integrating all electricity sources more effectively, enhancing grid security and bolstering U.S. competitiveness in the global energy economy. The Biden administration's allocation of $20 billion for grid modernization represents the largest investment of its kind in U.S. history. In summary, as AI continues to grow, ensuring sufficient electrical power remains a critical issue. While technological advancements in AI and GPU production are notable, the focus must also include substantial investments in renewable energy and grid modernization to support the AI-driven future. Forbes Business Council is the foremost growth and networking organization for business owners and leaders. Do I qualify?