Between computers, smartphones, tablets, wearables, and the Internet of things, the number of networked devices around the world is growing rapidly, and all those devices need energy, even if they’re not doing anything. That could be a problem.
A new report from the International Energy Agency, an intergovernmental organization dedicated to ensuring reliable and clean energy, says that the electricity demand of networked devices around the world in 2008—420 terawatt-hours—was equal to that of France; in 2013 the demand surpassed that of Canada, reaching 616 terawatt-hours. By 2025, the report projects, networked devices will account for 6 percent of global electricity demand at 1,140 terawatt-hours. As much as 80 percent of that demand will be used just to maintain a network connection, keeping devices ready and waiting.
“In their current state, network-enabled devices carry an inherent paradox,” the IEA report says. “They have enormous potential to deliver diverse efficiencies across many sectors and services, yet they fall far short of their own potential to be energy efficient.”
Simply minimizing power consumption in standby mode, according to the IEA, could reduce the electricity demand of the world’s devices by 600 terawatt-hours annually by 2020.
Indeed, researchers are working on ways to ensure that devices are as efficient as possible while in standby mode–which is to say, they could be off.
The working memory in today’s devices stores bits of information in capacitors by electrically charging or discharging them. That information is lost if the power supply is cut off, making the memory volatile. (Our computers do also use nonvolatile flash memory, but it is too slow to function as working memory and is used only for mass storage.)
With an alternative approach, known as spintronics, information would be stored in the spin states of electrons, a quantum-mechanical phenomenon that can persist even when power is shut off (see “A Spintronics Breakthrough”).
Back in 2001, a Japanese researcher named Koji Ando coined the term “normally off computing” to describe spintronics-based devices that, unlike our “normally on” devices, would completely turn off between operations—even, for example, during the milliseconds of downtime between keystrokes as someone types. When Ando, a researcher at the National Institute of Advanced Industrial Science and Technology in Japan, first proposed the idea, spintronics was too far behind existing memory technology in performance to be commercially competitive. But academic researchers and companies such as Qualcomm, Toshiba, and IBM have been plugging away at it, and Ando and colleagues concluded this spring in the Journal of Applied Physics that the technology could be usable as the basis of some memory functions in computers within five to 10 years.
How AI is reinventing what computers are
Three key ways artificial intelligence is changing what it means to compute.
These weird virtual creatures evolve their bodies to solve problems
They show how intelligence and body plans are closely linked—and could unlock AI for robots.
A horrifying new AI app swaps women into porn videos with a click
Deepfake researchers have long feared the day this would arrive.
Surgeons have successfully tested a pig’s kidney in a human patient
The test, in a brain-dead patient, was very short but represents a milestone in the long quest to use animal organs in human transplants.
Get the latest updates from
MIT Technology Review
Discover special offers, top stories, upcoming events, and more.