Intel’s next-gen chips for laptops, known as Lunar Lake (Core Ultra 200), won’t just major in power-efficiency, but will also provide a powerful performance for AI workloads in the best ultrabooks.
We could have guessed this was the case – as Intel was always going to advance considerably on the AI processing that current-gen Meteor Lake is capable of – but at the Vision 2024 event, Intel gave us exact figures for Lunar Lakes AI capabilities (measured in TOPS, or trillions of operations per second).
To put things in perspective, Meteor Lake sports an NPU capable of 10 TOPS, but Lunar Lake will treble that to 45 TOPS (exactly the threshold for Microsoft’s definition of an ‘AI PC’ it should be noted).
However, that’s not the full story, as Lunar Lake will be capable of over 100 TOPS when the AI processing power of the CPU and integrated GPU are taken into account alongside the NPU accelerating things.
As regards the additional 55 TOPs which comes from the CPU and GPU, we don’t know how that breaks down, as Intel didn’t elaborate on this. But it’s a safe bet that the majority of the work here is being done by the graphics solution, which is Battlemage (Intel’s next-gen series of GPUs, our first sighting of which could well be in Lunar Lake).
Analysis: Next-gen AI performance will be a close-fought battle
Intel’s Lunar Lake CPUs are expected to debut later this year – although volume production likely won’t happen until 2025 – and big things are expected of the laptop chips. That’s both in terms of raw performance and also with the various tricks hidden up the sleeve of this silicon – plus this talk of 100+ TOPs will only inflame matters further. But how does that figure compare to the rival notebook silicon on the horizon?
Well, notably there’s the Snapdragon X Elite that’s going to debut before Lunar Lake, with laptops carrying this SoC set to arrive in June. We already know that Qualcomm’s chip will offer 45 TOPS from its NPU, which is exactly the same as Lunar Lake.
However, it should be noted that overall – including the CPU plus GPU, as well as NPU – the Snapdragon SoC will reach 75 TOPS, which is considerably lower than the mentioned 100 TOPS for Lunar Lake. And the other downside of an ARM-based chip is the restrictions on the apps you can use (or at least, how fast software that isn’t written for ARM will run under emulation, with Windows on ARM).
What about AMD? As Tom’s Hardware, which spotted Intel’s revelation, informs us, AMD hasn’t put a figure on AI performance with its next-gen Strix Point chips for laptops. However, Team Red did say that Strix Point – with a next-gen XDNA 2 NPU – will offer up to three times the generative AI performance of current silicon, which is 16 TOPS. So, that puts the NPU for Strix Point at around 48 TOPS, just a hair faster than Intel and Qualcomm in theory – but it’s not much of a difference.
All of these NPUs will be in the same ballpark, but in overall AI performance, Intel looks to have the better of Qualcomm – and we don’t yet know exactly how Strix Point will pan out in terms of what grunt the CPU and GPU will add to that NPU. We’re betting it’ll be a close race all-round, though, and we must remember that AI tasks aren’t everything either – general performance under Windows 11 is more important, naturally.
