 |
|
|
|||||
When you ask ChatGPT a question, how does it come up with the answer? Most people dont give much thought to whats going on behind the screen or how graphics processing units (GPUs) make the AI magic happen in data centers across the country. GPUs are extraordinary, processing exceptionally large data sets and handling highly intensive computations involving billions of simultaneous calculations. Modern GPUs can train neural networks that enable AI breakthroughs that were once the subject of science fiction films. Thanks to GPUs, AI is being used to more accurately diagnose diseases, accelerate drug discovery, detect fraud, improve transportation efficiency, improve customer service, protect our national security interests, and provide personalized education experiences. Not surprisingly, the parallel processing architecture that makes GPUs excel at training AI requires an immense amount of energy. Thats why, according to Goldman Sachs Research, the power usage of data centers will grow by 160% by 2030. Our country is at a critical inflection point. The surge in energy required to power GPUs presents both a challenge and an opportunity. The United States power grid and service modelsbuilt in the pre-AI eraare not equipped to meet this demand. Today, data centers only account for about 2% of U.S. power consumption. But as AI models grow, data center development will outpace how quickly new electrical energy supply and transmission can be built without a shift in strategy. Seize the AI moment This moment marks an opportunity for utility providers, communities, and the technology industry to come together to turn AI into a source of long-term economic prosperity and growthrather than let it create a chasm. GenAI has the potential to transform the U.S. economy, with some estimating it could increase U.S. labor productivity by 0.5 to 0.9% annually. Investing in our AI infrastructure will ensure American communities can attract and retain highly skilled talent, while accelerating scientific advancements. As global powers invest in AI, the right infrastructure is also necessary to ensure our governments and communities can protect critical infrastructure. So how do we seize this opportunity? We need to rethink our approach to building infrastructure and delivering power. Today, electric utilities have a duty to serve demand as it comes. In laymans terms, that means anyone can say, Im building something over here and I need more power for it, and the utility has to add that request into their queue for service. That can be unfair when some customers are asking for massive amounts of power, which could impact the entire community. In some cases, speculative developers demand power in the early stages of a project, prompting utilities to plan and invest in infrastructure for a data center project without backing that may never come to fruition. These speculative developers operate under the assumption that they can put the pieces of the puzzle in place to later sell to a data center developer or operator, a prospect which may never transpire. When those projects fall through, the local community is left holding the bag. Build a more sustainable approach Its possible for data centers to access the energy they need without driving up energy costs for everyone. To achieve this, the industry needs to work with utilities in a collaborative way to close the gap between supply and demand. Data centers can and should take on more of the risk (and costs) of building infrastructure and transmission, which will require regulatory support. Data center developers generally have easier access to capital, fewer restrictions when it comes to buying land and faster decision-making capabilities than regulated utilities. That means, were able to contribute financially to the development of transmission lines and substations. Complex regulatory processes often stand in the way of this today. Paving the path for microgrids, which enable data centers go off grid power and even return power to the grid during peak demand periods, can turn data centers into a grid resource rather than a drain on supply. The path forward AI represents unprecedented potential for human advancement. Whats always been done wont work in the era of AI and electric-everything. American utilities are going to need to invest $50 billion to build new generation capacity for data centers alone by 2030. To achieve this, energy leaders are going to need to think out of the box and explore ways to closely collaborate with industry. Its the only path forward for long-term, resilient growth and a more equitable AI-powered future. Chris Crosby is CEO of Compass Datacenters.
Category:
E-Commerce
The modern international food trade currently plays a significant role in breaking food insecurity in parts of the world. Although innovative, the downsides are the high expense of transit, supply chain vulnerability, and potential for environmental harm. The alternative? Helping farmers worldwide to successfully and sustainably overcome crop stressors such as insects that take a toll on crops grown to feed the populations close to them. AI can help make a positive difference. There are an estimated 570 million farms of varying sizes globally, a number not expected to expand significantly. The worlds population, however, is predicted to grow (from almost 8 billion to nearly 10 billion by 2050) requiring farmers to produce 50% more food on those 570 million farms, a significant problem, unless there is change. Farmers already lose 20-40% of their crops annually to weeds, pests, and disease; changing pressures caused by climate change and resistance are exacerbating the situation. Many current treatments to fight weeds, pests, and disease were developed over 30 years ago and today struggle against resistance. Farmers are resourceful and eager to find solutions but pests are nimble and persistent. North Americas corn rootworm, for example, has adapted to crop rotation. And Asias barnyard grass mimics rice plants to evade hand-weeding. Innovation is desperately needed to provide solutions. So, how can we ensure there is sustainable, more localized, food production in a world where farmers face these challenges? Artificial intelligence (AI) is emerging as a promising tool to address these challenges by efficiently turning data into actionable recommendations and solutions. Take a page from pharmas AI book AI is revolutionizing many industries, from education to energy. While its still early in its impact, improving crop yields and bringing agriculture into the future is shaping up to be another industry AI is poised to disrupt. AI can help with on-field tasks like efficiently powering a sprayer. It can also help with in-lab tasks like accelerating the discovery of safer, more effective crop protection solutions. Consider the pharmaceutical industry. AI is set to transform drug discovery by enabling the rapid development of vaccines and treatments to protect global health. Today, scientists can create annual vaccines for evolving viruses like COVID-19 and influenza while also advancing therapies for other common and rare diseases. AI is beginning to equip scientists with tools to navigate the vast diversity of chemical space, prescreening them for efficacy and then rapidly identifying the most promising molecules in the fight against disease. By analyzing data from sources like DNA-encoded chemical libraries, AI helps scientists pinpoint potential candidates from billions of options. Machine learning models further expand this exploration, unlocking chemical diversity from ultra-large, make-on-demand libraries. Now, apply that approach to agriculture, because crops get sick too. In humid regions, fungi threaten yields, while in arid climates, insect infestations can destroy fields within days. AI-driven innovation in crop protection can help address these challenges with the same urgency and precision as in pharma research and discovery. Sifting through genetic and chemical datasets can help farmers tackle evolving threats faster and more accurately. Traditional crop protection discovery is slow and expensive, typically taking more than 13 years to bring a product to market. AI-informed research can likely cut discovery time in half and ultimately generate higher quality leads. A better way to support farmers Beyond crop protection discovery benefits, the responsible use of AI in agriculture has the potential to transform global food systems, making them more sustainable and resilient in the face of challenges. With AI, farmers gain powerful tools to not only safeguard their crops but also to enhance overall productivity and sustainability. AI can help farmers anticipate and respond to the effects of climate change, such as altered growing seasons, pest invasions, and extreme weather events. With predictive tools, farmers can make informed decisions about crop rotation, pest control, and irrigation, leading to improved outcomes while navigating unpredictable conditions. Farmers can also harness AI to improve productivity while minimizing environmental impacts. AI-driven solutions can allow for precise monitoring of soil health, real-time weather analysis, and efficient resource use, ensuring that farmers apply water, fertilizers, and pesticides only where and when they’re needed most. This reduces waste, lowers costs, and mitigates the negative effects of overuse on the environment. The long-term potential of AI in agriculture lies in its ability to boost productivity for farmers and foster more sustainable food systems that can feed a growing global population while preserving the health of the planet. Looking to the future AI is already making meaningful strides in revolutionizing agriculture and the potential is enormous. Enhancing crop protection and boosting productivity without the need for farmland expansion is only the beginning. The potential for breakthroughs is vast, with new solutions on the horizon that could significantly transform agriculture and drive further progress. By continuing to innovate and integrate AI into agricultural practices, well reach new levels of sustainability and efficiency, ultimately creating a more resilient and productive food system to support the world. As these technologies develop, ongoing research, ethical considerations, and farmer education will be critical to ensure AIs responsible integration into agriculture. Jacqueline Heard,PhD, MBA is cofounder and CEO of Enko.
Category:
E-Commerce
The Chinese AI company DeepSeek has put the AI industry in an uproar. Denied the most powerful chips thought needed to create state-of-the-art AI models, DeepSeek pulled off some engineering master strokes that allowed the researchers to do more with less. The DeepSeek-V3 and DeepSeek-R1 models the company recently released achieved state-of-the-art performance in benchmark tests and cost much less time and money to train and operate than comparable models. And the cherry on top: The companys researchers showed their workthey explained the breakthroughs in research papers and open-sourced the models so others can use them to make their own models and agents. The main reason DeepSeek had to do more with less is that the Biden administration put out a series of restrictions on chip exports saying that U.S. chipmakers such as Nvidia couldnt ship the most powerful GPUs (graphics processing units, the go-to chip for training AIs) to countries outside the U.S. This effort started in October 2022, and has been updated and fine-tuned several times to close loopholes. Biden released an executive order shortly before leaving office further tightening restrictions. DeepSeek apparently played by the rules. It made do with H800 chips the U.S. allowed Nvidia to sell in China, instead of the more powerful H100 that U.S. tech and AI companies use. With less powerful chips, the researchers were forced to find ways of training and operating AI models using less memory and computing power. The DeepSeek models use a mixture of experts approach, which allows them to activate only a subset of the models parameters that specialize in a certain type of query. This economizes on computing power and increases speed. DeepSeek didnt invent this approach (OpenAIs GPT-4 and Databrickss DBRX model use it), but the company found new ways of using the architecture to reduce the computer processing time necessary during pretraining (the process in which the model processes huge amounts of data in order to optimize its parameters to correctly respond to user queries). In DeepSeek-R1, a reasoning model comparable to OpenAIs most recent o1 series of models (announced in September), DeepSeek found ways of economizing during inference time, when the model is thinking through various routes to a good answer. During this process of trial and error, the system must collect and store more and more information about the problem and its possible solutions in its context window (its memory) as it works. As the context window adds more information, the memory and processing power required leaps up quickly. Perhaps DeepSeeks biggest innovation is dramatically reducing the amount of memory allocated to storing all that data. In general terms, the R1 system stores the context data in a compressed form, which results in memory savings and better speed without affecting the quality of the answer the user sees. DeepSeek said in a research paper that its V3 model cost a mere $5.576 million to train. By comparison, OpenAI CEO Sam Altman said that the cost to train its GPT-4 model was more than $100 million. Since the release of DeepSeeks V3, developers have been raving about the models performance and utility. Consumers are now embracing a new DeepSeek chatbot (powered by the V3 and R1 models), which is now number one on the Apple ranking for free apps. (However, that success has attracted cyberattacks against DeepSeek and caused the company to temporarily limit new user registrations.) For the past two years, the narrative in the industry has been that creating state-of-the-art frontier models requires billions of dollars, lots of the fastest Nvidia chips, and large numbers of top researchers. Across the industry and in investment circles that assumption has been challenged. As a result, Nvidia stock fell nearly 17% Monday as investors question their assumptions about the demand for the expensive GPUs. And its all happening because a small shop of Chinese researchers knew theyd need some big engineering breakthroughs in order to create state-of-the-art models using less than state-of-the-art chips.
Category:
E-Commerce
All news |
||||||||||||||||||
|
||||||||||||||||||