Apple took the wraps off its latest custom silicon on Tuesday, unveiling the M4 chip that will power the company’s new iPad Pro models. But the M4 isn’t just an incremental upgrade — it represents a major doubling down on artificial intelligence performance and capabilities.

At a special event, Apple’s executives gushed about the AI muscle of the M4. With a new neural engine capable of 38 trillion operations per second, they boasted it is “an outrageously powerful chip for AI.” That’s 60 times faster AI processing than Apple’s A11 chip from 2017, though it still falls short of the 45 TOPS (trillion operations per second) promised by Qualcomm’s upcoming Snapdragon X Elite chip for Windows PCs.

Still, Apple proudly claims the M4 can deliver desktop-grade AI performance with just a quarter of the power consumption. Backing up that AI prowess is a 10-core CPU split between four high-performance and six efficiency cores. Apple says this makes the M4 50% faster than the M2 found in the previous iPad Pro for CPU tasks.

The graphics capabilities have also been turbocharged, with a 10-core GPU supporting advanced rendering techniques like dynamic caching, mesh shading, and ray tracing. Plus, a revamped display engine quadruples the pixel rendering speeds over the M2. Apple’s use of cutting-edge 3nm process technology no doubt aids the performance and efficiency gains.

What’s particularly noteworthy is just how laser-focused Apple’s marketing has become on the M4’s AI capabilities. With each new M-series chip, the company has touted incrementally better AI performance as measured by those “trillion operations per second” figures. But the jump from the M3‘s 18 TOPS to the M4’s 38 TOPS is by far the biggest generational AI upgrade yet for its homegrown silicon.

Of course, as industry observers have cautioned, those TOPS numbers don’t tell the whole story. What ultimately matters is how well optimized the AI hardware and software stack is for the types of machine learning workloads being thrown at it. Apple’s in-house Arm CPU designs, unified memory architecture, and tight software integration could give it an edge over x86 Windows PCs when it comes to mobile AI performance and efficiency.

Apple’s focus on pushing the AI boundaries with the M4 is exquisitely timed. Just two weeks after this iPad launch, Microsoft is hosting its own AI event where it’s expected to take the wraps off new Arm-based chips and Windows laptops designed specifically for advanced AI workloads.

Microsoft has expressed confidence that its Arm partnership with Qualcomm — which includes access to the powerful AI Engine in the Snapdragon X Elite chips — will finally allow Windows PCs to beat Apple’s M-series machines, including the latest M3-based MacBook Airs, on both general CPU performance and AI acceleration.

Given the escalating AI arms race between Apple, Microsoft/Qualcomm, and others like Google and NVIDIA, consumers can expect to be the beneficiaries as advanced machine learning capabilities get baked into every new chip launch. With AI expected to play a pivotal role in everything from intelligent assistants to content creation to data processing, having powerful and efficient AI processing at one’s fingertips will be table stakes.

Apple has made a savvy move to get out in front and tout the AI prowess of its next-gen silicon. But you can be certain its Silicon Valley rivals won’t let Cupertino’s AI lead last for long.