ORLANDO, Fla. — As energy availability and cost become a constraint on technology usage and enterprise AI applications increase computing needs within organizations, IT managers must plan to address the issue. .
At Gartner IT Symposium/Xpo 2024, which concludes today, the power needed to power growing IT workloads was an undercurrent, especially given AI’s increasing energy usage. The annual event also featured discussions about emerging technologies such as agent AI and the future of IT, where multiple developments such as quantum computing and neuromorphic computing could arrive within the same time frame. . Coordinating deeper layers of technology resources.
However, speakers at the conference pointed out that energy consumption is an immediate problem that is unlikely to be resolved anytime soon.
Jensen Huang, president and CEO of Nvidia, said recent advances against Moore’s Law, which states that CPU performance effectively doubles every two years with minimal additional cost, have led organizations to He said the company is far behind in terms of the amount of data it processes. The slowdown in earnings growth started a decade ago and is now “to the point where it doesn’t exist,” Huang said in his keynote speech. As a result, computing costs and energy usage skyrocket.
“If computing demand continues to grow exponentially, but general-purpose computing performance does not increase, we should see inflation in computing costs and, very importantly, inflation in computing energy. ” said Huang. “In fact, energy consumption is increasing exponentially.”
Huang, whose company is a leading supplier of GPUs, said accelerated computing addresses energy efficiency. “You have to use acceleration wherever possible,” he said.
Gartner analyst Darryl Plummer pointed to a different approach during a discussion of the consulting and market research firm’s 2025 strategic forecasts. He said Fortune 500 companies will shift $500 billion from energy business investments to microgrids by 2027. Combating “chronic energy risks and AI demands”. Microgrids provide an independent energy system to meet the needs of a company or group of companies.
“Essentially what we’re saying is we’re not producing enough energy, and the energy we produce can’t be delivered through weak distribution infrastructure that isn’t modernized enough,” Plummer said. said.
Companies that use large amounts of electricity are starting to “generate their own energy” and share that power with close corporate partners through microgrids, he added. “This is something you should be thinking about today if you’re a big consumer of energy.”
In fact, energy consumption is increasing exponentially. Jensen Huang Nvidia President and CEO
Plummer said cloud data center operators are already buying “small nuclear power plants,” noting that Microsoft is moving ahead with plans to buy power from the Three Mile Island nuclear power plant in Pennsylvania. Nuclear power is one energy source the company and other hyperscalers are considering tapping into as generative AI’s power needs continue to grow.
Companies may need to pursue energy risk mitigation strategies for several years before their energy risks are mitigated. Gartner analyst Frank Buytendijk predicted that it will take five to 10 years to realize the “promise of energy-efficient computing.”
Buytendijk, who spoke about the future of computing at the symposium, said some nascent technologies may emerge during that time.
“Quantum computing, neuromorphic computing, and optical or photonic computing can be much more energy efficient and often much faster,” Buytendijk said. “It’s going to take time, but we’ll get there. The idea of energy-efficient computing is a little further down the road.”
John Moore is a TechTarget editorial writer, covering CIO roles, economic trends, and the IT services industry.
