"If you pay attention to what the biggest tech companies are saying about AI demand, you'll notice a common thread: they're running short of compute capacity. That means the large language models underpinning AI products today need even more data centers to be trained and for inferencing, and therefore, they need more power. Against that backdrop, energy efficiency has suddenly become a critical priority for semiconductor manufacturers."
"PowerLattice, a startup founded by veteran electrical engineers from Qualcomm, NUVIA, and Intel in 2023, claims to have developed a groundbreaking approach that reduces the power needs of computer chips by more than 50%. On Monday, the startup emerged from stealth with a $25 million Series A funding round led by Playground Global and Celesta Capital, bringing its total funding to $31 million."
""This is the hard stuff: How do you get power into the device? There are very few teams and people that can do it," said Pat Gelsinger, general partner at Playground Global. "We have assembled what I'd argue is the dream team of power delivery." As the former CEO of Intel, Gelsinger carries significant authority in the semiconductor world, which makes his participation a powerful stamp of approval for PowerLattice."
Big tech firms report shortages in compute capacity, driving demand for more data centers and higher power consumption for training and inferencing large language models. Energy efficiency has become a top priority for semiconductor manufacturers. PowerLattice, founded in 2023 by engineers from Qualcomm, NUVIA, and Intel, developed a power delivery chiplet that brings power closer to processors and cuts chip power needs by more than 50%. The startup raised $25 million in a Series A led by Playground Global and Celesta Capital, totaling $31 million. Pat Gelsinger endorsed the team's expertise. TSMC is producing PowerLattice's first chiplet batch for testing.
Read at TechCrunch
Unable to calculate read time
Collection
[
|
...
]