Intel's (INTC +0.30%) Kirk Skaugen recently admitted in a recent investor conference that part of the reason that desktop demand was so weak in the first half of 2015 is that the company didn't bring its Broadwell processor family to traditional "tower" desktops. Instead, the company relied on warmed-over versions of its Haswell processor family that was introduced in mid-2013.
Although Skaugen stressed that the company will return to yearly refreshes of its desktop processors in order to avoid seeing the kind of demand slowdown that it did when it skipped Broadwell, I have to wonder if Intel might be forced to endure a similar "slowdown" with the launch of its Kaby Lake chips in 2016.
What will Kaby Lake really bring to the table?
According to Intel CEO Brian Krzanich on the company's most recent earnings call, the Kaby Lake family of processors will be built on the same 14-nanometer manufacturing technology that the just-launched Skylake family of chips are built on.
Further, Krzanich said that Kaby Lake will be "built on the foundations of the Skylake micro-architecture but with key performance enhancements." I believe that, from this statement as well as some of the other leaks and hints available on the Web, that Kaby Lake will be based on Skylake CPU cores paired with an updated graphics engine.
In that case, Kaby Lake should be a solid improvement in laptops, 2-in-1 devices, and all-in-one desktops; in other words, segments of the market where integrated graphics performance is important to customers.
But will it be an upgrade for traditional tower desktops?
A recent set of leaked slides from BenchLife.info indicate that Intel does, in fact, intend to bring Kaby Lake to traditional tower desktop form factors. In addition to the improved integrated graphics, I suspect that Intel will be able to achieve a modest CPU performance improvement by pushing the clock speeds higher, since the 14-nanometer process should be more mature by the time the chips launch (similar to what Intel did with its Haswell Refresh chips).
However, it's pretty easy to make the case that integrated graphics isn't all that relevant to buyers of traditional desktop towers. Indeed, I would argue that the kinds of buyers that didn't buy a system due to the lack of a Broadwell chip aimed at tower desktops are the same kinds of performance-hungry buyers (i.e., gamers) who will use stand-alone graphics processors.
If the Haswell Refresh led to muted tower desktop PC demand, I don't think Kaby Lake will fare too much better in the marketplace.
Will Skylake-E be Intel's desktop ace-in-the-hole?
Intel is expected to launch its next generation high-end desktop processor family known as Skylake-E at some point in 2017 (I'd bet on the first half, if not first quarter, of 2017). These chips should bring sizable performance improvements for high-end desktop buyers relative to the current Haswell-E processors as well as the Broadwell-E processors expected to arrive in the first quarter of next year.
However, unless Intel plans to start more aggressively emphasizing this platform for tower desktops (as I suggested here), this probably won't make much of a difference. Why? Well, Intel launched its previous high-end desktop chips, known as Haswell-E, at the end of 2014 and it wasn't enough to mitigate the slump in desktop chip sales that the company saw in the first half of 2015.
If Intel does get more aggressive with Skylake-E by introducing relatively high core count processors at lower price points than what we see now (the cheapest six-core Haswell-E processor retails for about $390), then this could help shore up demand for its tower desktop processors.
