Our connected devices are hiding a big secret. They use energy—a lot of it. Every time you use your phone, your computer, or your smart TV to access the internet, you’re sending data requests to warehouse-sized buildings around the world, full of hundreds of thousands of servers. These data centers are among the most energy-intensive systems on the planet, representing approximately 10% of global electricity generation (though more conservative estimates put it at 3%).
Yet we’re still blindly making classic computers—and they’re getting bigger and even more energy dense. China is home to the most energy-intensive supercomputer in the world, the Tianhe-2 in Guangzhou. This machine uses about 18 megawatts of power, and is expected to be succeeded by the exascale Tianhe-3, which will only further increase this extraordinary level of energy consumption. For reference, the average hydroelectric dam in the US produces approximately 36 megawatts of power.
This is just one reason why quantum computing is key to the future. In addition to holding the potential to solve some of the world’s most computationally challenging problems, quantum computers use significantly less energy, which could lead to lower costs and decreased fossil-fuel dependency as adoption grows. (Disclosure: I’m also the CEO of a quantum-computing company.)
Unlike classical computers, which use binary bits to encode information as 1s or 0s, quantum computers work using qubits. Thanks to the “weirdness” of quantum mechanical properties, qubits can represent both 1s and 0s at the same time, allowing quantum computers to find optimal solutions that classical systems cannot, all while using less energy.
Here’s why: For a quantum processor to exhibit quantum mechanical effects, you have to isolate it from its surroundings. This is done by shielding it from outside noise and operating it at extremely low temperatures. Most quantum processors use cryogenic refrigerators to operate, and can reach about 15 millikelvin–that’s colder than interstellar space. At this low temperature, the processor is superconducting, which means that it can conduct electricity with virtually no resistance. As a result, this processor uses almost no power and generates almost no heat, so the power draw of a quantum computer—or the amount of energy it consumes—is just a fraction of a classical computer’s.
And then there’s the price. Most modern classical supercomputers use between 1 to 10 megawatts of power on average, which is enough electricity to meet the instantaneous demand of almost 10,000 homes. As a year’s worth of electricity at 1 megawatt costs about $1 million in the US, this leads to multimillion-dollar price tags for operating these classical supercomputers. In contrast, each comparable quantum computer using 25 kilowatts of power costs about $25,000 per unit per year to run.
Businesses are constantly looking for a competitive advantage, especially in an era of shrinking margins and fierce competition. In the case of computing, they’re looking for better, faster, or more efficient ways to solve problems than a classical computer. In the future, most quantum applications will utilize hybrid computing, which is a combination of classical and quantum computing that will provide a workable alternative to this unsustainable status quo—one that unlocks new commercial applications while dramatically curbing energy usage and costs.
With hybrid, the hard parts of commercial computing that aren’t suitable for existing classical systems can be sent to a quantum processing unit and returned to a classical computer. High-energy portions of hybrid applications will be run on quantum computers—often through the cloud—while the low-energy pieces are reserved for classical. Hybrid computing means utilizing the best of both the quantum and classical worlds, and lowering the barriers for companies of all sizes to get started using quantum computers.
Thanks in part to hybrid computing, early quantum applications are already being used in industries including automotive, manufacturing, and finance. Volkswagen is using quantum computers to build early applications that will be able to optimize public transportation routing in cities around the world. DENSO, a leading auto-parts manufacturer based in Japan, has reported that it can reduce gridlock and improve efficiency of autonomous robots on its factory floors with the help of an application built with a quantum computer.
Quantum computing is showing signs of early benefits today, but there’s more to do before we see fully practical deployment of quantum computing in production. We need continued buy-in and investment from both governments and businesses to achieve widespread adoption. We also need to train and develop the next generation of expertise and talent in the quantum workforce. Finally, we need to continue breaking down barriers to using quantum computers with affordable, flexible cloud access and developer-friendly software and tools.
Quantum computers hold the promise to solve today’s toughest business problems and impact the bottom line for companies in virtually every industry. They’re also a key tool we can use to combat the looming threat of ever-growing energy use of classical computing. Businesses are already starting to feel the pressure to get their heads in the quantum-computing game, but the impetus goes beyond innovation and technological competition for a single company. It extends to a collective goal: ensuring our world’s computing power doesn’t outstrip our planet’s ability to support it.
Correction: The previous headline for this piece “We’ll run out of energy in 20 years if we don’t switch to quantum computing” overstated the threat to global energy production. The headline has been updated to better reflect the article text. In addition, the article has been updated to more accurately explain the costs of electricity generation and use.