Brain-like chips may make PCs 1 million times more efficient

Future computers may be modeled to simulate and resemble the human brain, and however futuristic that might sound, this technology might be closer than we think.

A new scientific report shows that a new material could potentially help in the production of PC chips that will be 1 million times more efficient than current technologies, and this could happen before the end of this decade.

A new brain-like computer chip.
Image source: KTH

Scientists from the KTH Royal Institute of Technology in Stockholm, Sweden and Stanford University teamed up to research the subject and published their findings on the KTH website. According to their research, this future, ultrapowerful brain-like PC could become reality thanks to a new material that has been developed.

The material in question is a titanium carbide compound called MXene. Using this new material to make memory components could potentially bring the scientists’ vision to life, making computers that simulate the human brain not just real, but also commercially viable, in the near future.

The scientists found that electrochemical random-access memory (ECRAM) components showed a lot of potential when made using 2D titanium carbide. Capable of both storing and processing data, ECRAM would cut out a lot of processor and memory communication by performing both tasks at once.

The report describes the way ECRAM works, referring to it as “a component in which switching occurs by inserting ions into an oxidation channel, in a sense similar to our brain, which also works with ions.”

A render of a head resting on a computer chip.
Digital Trends Graphic

Developing such a computer with brain-like qualities could create a machine with outstanding A.I. capabilities, and according to the scientists, it would be easy enough to manufacture with the use of MXene. The assembly process would be the same as complementary metal-oxide semiconductor (CMOS) wafer assembly and would involve integrating layers of 2D material on silicon.

In a statement made to TechRadar Pro, Max Hamedi , KTH associate professor, talked about the power these new computers could possess. The memory components made with MXene have been faster than any previously shown.

“We will be able to fabricate special purpose computer blocks [in, say, 5 to 10 years] where memory and transistors merge, making them at least 1,000 times more energy efficient than the best computers we have today for A.I. and simulation tasks [some calculations even show 1 million-fold energy efficiency for certain algorithms],” said Hamedi.

These future computers will require at least five years of trials, but one of the hardest parts is already over: The breakthrough. If things continue down this road, we may be seeing unprecedented computing power within the next few years.

Editors’ Recommendations