Intel Lunar Lake “Core Ultra 200” CPUs To Offer Over 100 AI TOPs & 3x NPU Performance, 40 Million AI CPUs Shipped By 2024 End
Intel has announced that its Next generation Lunar Lake “Core Ultra 200” CPUs AI will offer more than 100 AI TOPS to boost PC speed.
Intel Lunar Lake “Core Ultra 200” CPUs feature over 100 TOPs of combined NPU and GPU AI performance.
Intel's next-generation Lunar Lake CPUs will be a huge deal for Chipzilla as it strengthens its position in the AI PC segment. The company has already launched its first-generation NPU codenamed Meteor Lake in its Core Ultra 100 series, and the next-generation Arrow Lake and Lunar Lake chips will take things to the next level.
During this Vision 2024 eventIntel showcased some more AI performance metrics of its Lunar Lake CPUs, which will be branded under the “Core Ultra 200” series (Core Ultra 200V to be exact). Lunar Lake chips will be an advanced packaged SOC that includes new IPs and new technologies such as an on-package LPDDR5 memory solution.
Will use up to Lunar Lake CPUs. Battlemage iGPU (Xe2-LPG) with up to 8 cores and 8 Xe cores featuring Lion Cove P-Cores and Skymont E-Cores on Compute Tile. The chip will also feature an all-new NPU that delivers up to 3x the performance over existing chips like the Core Ultra 100 “Meteor Lake”.
Talking about the AI performance of its Lunar Lake “Core Ultra 200” CPU, Intel said the chips will offer more than 100 TOPs which is a huge increase from Meteor Lake which offers around 30-35 TOPs. It is a combined form factor that includes CPU, GPU, and NPU.
Meteor Lake's rated NPU at 3x is going to be about 45 TOPs which leaves about 50 or so AI TOPs for the Xe2-LPG Battlemage iGPU. Battlemage is going to be a huge architectural upgrade over the Alchemist and Alchemist+ iGPUs featured on the Meteor Lake and Arrow Lake CPUs, so the benefits are expected and it's good to see Intel consolidate TOPs for all core IPs. is similar to AMD and Qualcomm. have been doing too. Qualcomm has featured 75 TOPs for its Snapdragon X Elite CPU while Strix Point is going to be around the same ~100 AI TOPs given its 3x performance increase over Hawk Point. Thanks to the updated XDNA 2 design.
We recently said what kind of exodus The AI PC segment may have inspired chipmakers to focus on the NPU side of things more than traditional CPU and GPU enhancements.. While chips are getting faster across the board, there's still a case of “what could have been” if AI hadn't been a big thing and less effort had been put into making NPUs faster. GPUs are still the way to go when it comes to native AI processing, and while having an NPU isn't necessarily a bad thing, prioritizing resources over AI can mean we shift to the CPU and GPU. Can see fewer upgrades.
It remains to be seen how the consumer market reacts to the AI PC segment in the coming years, but Intel says it expects to ship 40 million AI CPUs by the end of 2024 and 100 million units by 2025. There is a plan.
2024 AI PC platforms
| Brand name | Intel | AMD | Intel | AMD | Qualcomm | apple |
|---|---|---|---|---|---|---|
| CPU name | Lunar Lake “Core Ultra 200” | Ryzen 9050 “Strix Point” | Meteor Lake “Core Ultra 100” | Ryzen 8040 “Hawkpoint” | Snapdragon X Elite | M3 |
| CPU architecture | x86 | x86 | x86 | x86 | arm | arm |
| CPU process | Intel 20 A | 4nm | 7nm (Intel 4) | 4nm | 4nm | 3 nm |
| Maximum CPU cores | 8 Cor | 12 Cor | 16 Cor | 8 Cor | 12 Cor | 16 cores (MAX) |
| NPU architecture | NPU | XDNA 2 NPU | NPU | XDNA 1 NPU | Hexagonal NPU | In the house |
| Total AI tops | ~100 tops (45 tops NPU) | ~100 tops (45 tops NPU) | 34 TOPS (11 TOPS NPU) | 38 Tops (16 Tops NPU) | 75 tops (tops) | 18 tops |
| GPU architecture | Battlemage Arc Xe2-LPG | RDNA 3+ | Alchemist Arc Xe-LPG | rDNA 3 | Adreno GPU | In the house |
| Maximum GPU cores | 8 Xe-Cores | 12 compute units | 8 Xe-Cores | 12 compute units | TBD | 40 cover |
| GPU TFLOPs | TBD | TBD | ~4.5 TFLOPS | 8.9 TFLOPS | 4.6 TFLOPS | TBD |
| Memory Support (Maximum) | LPDDR5X | LPDDR5X-7500 | LPDDR5X-7467 | LPDDR5X-7500 | LPDDR5X-8533 | LPDDR5-6400 |
| Availability | 2 AH 2024 | 2 AH 2024 | Q4 2023 | Q1 2024 | Mid 2024 | Q4 2024 |

