MIT

A new AI controller modeled on biological synapses lets soft robotic arms handle novel tasks and unexpected disturbances without retraining, achieving over 92% shape accuracy under real-world disruptions.

Using multi-slice electron ptychography on a PMN-PT alloy, MIT-led researchers directly imaged a hierarchy of polar and chemical structures, finding that polarization regions are much smaller than leading simulations predicted.

The startup's chemically tailored boron polymer membranes already achieve the U.S. Department of Energy's 2031 efficiency target of 77 percent for proton exchange membrane electrolysis, while partner tests show a 60 percent reduction in operating costs.

MIT engineers have built an AI system called VibeGen that generates entirely novel protein sequences optimized for specific vibrational and flexing patterns, inverting the traditional structure-first approach to protein design.

Researchers from MIT and warehouse automation firm Symbotic have developed a hybrid deep reinforcement learning system that dynamically prioritizes hundreds of autonomous robots, preventing traffic jams and increasing package throughput by 25 percent in simulated warehouse environments.

MIT researchers present Wave-Former and RISE, two systems that combine millimeter-wave radar with generative AI to reconstruct hidden objects and entire indoor scenes through obstructions, achieving up to twice the precision of prior methods.

The MIT spinout plans to demonstrate a 1 MWh thermal battery in Bedford, Massachusetts this year, using graphite bricks heated to 2,400 degrees Celsius and liquid tin to deliver grid-scale storage at a claimed $25 per kilowatt-hour, a fraction of lithium-ion costs.

MIT physicists built the first terahertz near-field microscope and used it to directly image quantum oscillations of superconducting electrons in BSCCO, a result published in Nature.

The CMS Collaboration at CERN has detected wakes left by quarks speeding through quark-gluon plasma, providing the first direct evidence that the matter filling the universe in its first microseconds behaved as a nearly perfect liquid rather than a gas of free particles.

A DOE-backed proof-of-concept integrates cryogenic control chips directly into ion traps, potentially removing a key bottleneck blocking large-scale quantum computers.