Intel recently revealed that its upcoming Lunar Lake chips would be available this fall for Copilot+ AI PCs<\/a>, but the company waited until Computex to give us more technical details. For one, they’ll offer up to 48 TOPs (tera operations per second) of AI performance, thanks to an updated neural processing unit (NPU). In comparison, Intel’s previous Meteor Lake chips sported a 10 TOPS NPU, while AMD announced its Ryzen AI 300 chips <\/a>yesterday with 50 TOPS NPUs. The AI race is on, if you couldn’t tell.<\/p>\n Intel will once again have to play catch up in the AI PC space: AMD’s Ryzen AI 300 chips will be available in July alongside Qualcomm’s Snapdragon X Elite and X Plus. It’s unclear when, exactly, Lunar Lake systems will launch this fall. Still, for the Intel faithful, Lunar Lake appears to be a major upgrade. It will also sport a new Xe2 GPU, which will offer 80 percent faster gaming performance than the last generation, as well as an AI accelerator with an additional 67 TOPS of performance. (We’re still waiting to see how AMD’s new Radeon graphics will compare.)<\/p>\n Additionally, Lunar Lake will have revamped performance and efficient cores (P-cores and E-cores), and it should also deliver up to 40 percent lower system-on-chip power than Meteor Lake. Intel also says the chip features an “advanced low-power island” for efficiently handling background tasks. Clearly, both Intel and AMD are going to do whatever it takes to combat Qualcomm’s Copilot+ hardware. Those mobile chips are inherently more power efficient \u2014 they reportedly get over 20 hours of battery life on Copilot+ Surface devices (though we haven’t tested them yet).<\/p>\n When it comes to connectivity, Lunar Lake will offer the updated standards you’d expect: Wi-Fi 7, Bluetooth 5.4, PCIe Gen5 and Thunderbolt 4. (It’s strange that Intel isn’t committing to Thunderbolt 5 <\/a>yet, since it plans to launch that standard later this year.)<\/p>\n Given just how far off Lunar Lake actually is, Intel didn’t have more information to share about specific chip models or deeper specs. But judging from the company’s benchmarks, released during a media briefing ahead of Computex, it’ll should be faster than Meteor Lake while running Stable Diffusion.<\/p>\n<\/div>\n<\/div>\n<\/div>\n<\/div>\n