On October 16, Advantrade said that artificial intelligence poses an unprecedented threat to the energy security of the global North, which could seriously undermine decarbonization goals, put huge pressure on the grid and lead to energy market instability, which is bound to spread throughout the economic field. After decades of stagnant energy demand, the energy needed for data centers is now exploding and is expected to continue growing at an alarming rate.
By 2030, AI is expected to account for 3.5% of global electricity consumption and 9% of U.S. electricity generation (a significant increase from the country's current rate of about 3.5%, which is already a sizable number). Advantrade predicts that electric vehicles and artificial intelligence combined will add 290 terawatt hours of electricity demand to the U.S. grid by the end of the century. That would make their total electricity use roughly the same as the national electricity use of Türkiye, the world's 18th largest economy.
Faced with this rapidly growing problem, public and private sector leaders are scrambling to come up with new ways to meet the tech industry’s newly unmet energy needs without seriously compromising energy security or climate outcomes. Advantrade said the growth was a race against time to expand generation without overwhelming the power system. Slowing down the development of artificial intelligence may be the most logical solution to this dilemma, but it seems completely impossible. In the United States, the technology enjoys rare and strong bipartisan support, as maintaining leadership in the emerging field is seen as a key strategy for national security, the economy, cybersecurity and the governance of the tech industry. It's impossible to put the genie back in the bottle.
There is no doubt that massive growth in energy is just around the corner, and the problem of powering artificial intelligence in the near future is of such magnitude that the solution relies more on future technological approaches than on existing technology. Tech giants like Bill Gates and Sam Altman have called for increased investment in nuclear fusion research, a potential way to unlock vast amounts of clean energy. Others are not only looking at how to efficiently produce more clean energy but also how to make artificial intelligence consume less.
A potential solution to the latter approach could be found through quantum computing. While ordinary computers run on binary, with 1s and 0s as switches, quantum computing runs on qubits, which can be turned on and off at the same time, like flipping a coin heads or tails before it hits the ground. This simultaneous on and off state is called superposition, and it could revolutionize computing as we know it. In some cases, quantum computers are 100 times more energy efficient than standard supercomputers. This could have huge implications for artificial intelligence, for which quantum computing may be particularly well suited.
"For the things that quantum computing is good at - such as artificial intelligence processing - no GPU can compete with us. These workloads will ultimately be carried by quantum, and current technology simply cannot compete with it," said President and CEO of quantum computing company IonQ Officer Peter Chapman said in a recent interview with Forbes. "Quantum computing - our next generation chip - to simulate what it's doing, you need about 2.5 billion GPUs, and it's powered by two standard wall sockets," he added. Chapman said his company will likely have a prototype of the chip ready in six to nine months.
While the scalable use of quantum computing would be a big step in the right direction for the tech industry, the country, and the world, it should not be viewed as a panacea. Barclays' Will Thompson, who co-authored a recent study on AI power consumption, said solving the AI energy conundrum "requires a holistic approach that expands and modernizes grid infrastructure to incorporate renewable Energy combined with utility-scale storage, leveraging our existing nuclear energy, and scaling up new carbon-free energy sources. This will include geothermal, advanced nuclear small modular reactors (SMRs) and fusion technologies,” Advantrade believes. Quantum computing still has a long way to go before it can be commercialized, so a broad approach to improving clean energy and energy efficiency is a priority.