"7th-generation Core" CPUs will launch in a few months, and that's all we know.

Intel's main Computex announcement was the launch of its high-end (and high-cost) Broadwell-E chips, but the company also made a passing mention of a couple of next-generation architectures for mainstream and low-end systems that will ship in finished systems by the end of the year.

The most significant of these two architectures is Kaby Lake, the replacement for Skylake. Kaby Lake breaks from the "tick-tock" schedule that Intel has followed for most of the last decade; that schedule has been replaced by something Intel calls "Process, Architecture, Optimization," in which it introduces a new process (formerly a tick), introduces a new architecture on that process (formerly a tock), and then tweaks the architecture without changing the process. Kaby Lake is an "optimization" and will be built on the same 14nm process as Skylake.

Intel has said very little about Kaby Lake, and aside from confirming that the CPUs will be called "seventh-generation Core" processors and that they'll definitely be shipping later this year, it didn't reveal much new information at Computex. Previous rumors and leaks point to expanded 4K video playback capabilities, including support for HDMI 2.0 and HDCP 2.2 and hardware decode support for 10-bit HEVC and VP9 videos. The processors should also be socket-compatible with Skylake, provided your motherboard OEM provides a BIOS update to add support. Rumors say the Kaby Lake launch will start with low-voltage Core i3/i5/i7 and Core m3/m5/m7 CPUs for laptops and convertibles first and come to desktops later—Asus is already showing off a Surface clone with a Kaby Lake CPU, suggesting that the chip is already sampling to Intel's partners. This bodes well for its availability.

We've seen some reports that some Kaby Lake chipsets will support full 10Gbps USB 3.1 speeds and Thunderbolt 3, but it's not clear what "support" means in this context—Intel won't tell us whether those features will be integrated into the chipsets themselves or if they'll be available via external controllers like they currently are. We'll cover it when the company is ready to talk more freely, but for now Intel isn't saying much about the processors aside from the fact that they exist and that they'll be out in a few months.

Intel also mentioned a lower-end architecture named "Apollo Lake," which is supposed to take the place of the Cherry Trail-based Celerons and Pentiums that ship in lower-end systems. We already know a fair amount about what this platform looks like—according to AnandTech, it's due to ship in systems later this year, and it includes new CPU and GPU tweaks, DDR4 support, and other advancements aimed at low-end systems.

Kaby Lake will be followed by Cannonlake, another tweak of the Skylake architecture made on a new10nm process. It was originally slated for this year but pushed back to the second half of next year. Intel had significant problems getting its 14nm process off the ground in 2014 and 2015, and hopefully the extra time will help Intel avoid the continual delays and staggered release schedule that both Broadwell and Skylake suffered from.