Intel (INTC 2.19%) is finally taking steps that its scrappy underdog competitor Advanced Micro Devices took over a decade ago: It will be creating some separation between its chip manufacturing and chip design departments. Ultimately, there will still be one Intel versus the case AMD and GlobalFoundries, which is now an entirely separate entity, with the chipmaker being fully released into the wild in 2012. (GlobalFoundries became a separate entity in 2009, and AMD sold off its last equity stake in it three years later.)
For now, Intel's move will give investors an uncomfortable look at the ugly financials of its Foundry Services (IFS) manufacturing business. The good news is that, eventually, things should improve for IFS. But it will take time, and the picture won't be pretty anytime soon.
Why Intel wants a business shake-up
After years of rapid changes in the semiconductor industry, as well as some internal missteps at Intel, the company is struggling. It lost its crown as the top chip designer and manufacturer, and it's losing lots of money.
The company has been shuttering various parts of its business that are ancillary to its current goals, namely, catching up in processor technology and building a sustainable and growing chip manufacturing division. Big challenges lay ahead, though, including the wooing of big chip design houses like Broadcom that have been reluctant to pledge explicit support for IFS -- at least thus far.
So while Intel rebuilds, it's also splitting up the reporting and organization of what has for decades been a fully integrated business, one that designs its own chips, manufactures them, and then sells them to device assemblers (like PC and laptop makers). It's a business model that worked wonders. It was profitable in a bad year and really profitable in a good year. But not anymore. Based on an initial and very generalized financial model Intel provided, the separately managed manufacturing division will operate at a thin gross profit margin (just 9% after subtracting cost of goods sold) and negative operating margins (approximately negative 18% after selling and admin costs).
What changed? Manufacturing the most advanced chips in the world has gotten exponentially harder, not to mention more expensive. And while Intel sat happily atop its empire, an army of chip designers advanced semiconductor technology and tapped third-party foundry Taiwan Semiconductor Manufacturing to handle the heavy chipmaking lifting. As TSMC's capabilities increased along with its design peers, it was able to eat those higher costs of advanced manufacturing. Semi-stagnant Intel was not.
Based on expectations for next year, IFS will generate $20 billion in revenue, the vast majority of it derived from Intel's chip design departments. In contrast, TSMC has reported over $75 billion in sales over the last year, and it's ridiculously profitable.
Data by YCharts.
What's the plan now?
To drive more accountability, and to eventually give the chip design departments more flexibility in tapping other chipmakers for help (Intel's upcoming Meteor Lake chips are, in part, being made by none other than TSMC), Intel's mostly internal manufacturing financials will be reported separately from Intel's chip design departments.
And as Intel manufacturing pursues cost-savings opportunities, and courts new third-party customers, the expectation is it will return to profitability. There's no timeline on that return to profitability, though, nor did Intel announce a big new customer acquisition that shareholders have been waiting for. CFO David Zinsner did say during the presentation that "we continue to be confident that we will sign up our first Intel 18A external foundry customer this year."
When can Intel declare victory?
I've said numerous times in recent years I believe the eventuality is that Intel will fully split in two, much like AMD did. There could be massive benefits for Intel manufacturing to be fully freed from chip design, giving the two very different businesses flexibility in pursuing their own fate.
Times have changed drastically. Today, the world's most advanced chips are made with collaboration between different parties that each specialize in a separate step of the process, from design to chipmaking tools to final manufacture.
At any rate, it could make sense for Intel to first improve its flagging manufacturing arm before letting it go (don't sell low, sell high, as the investing phrase goes). For the time being, Intel's reorganization to give investors better insight into progress doesn't mean much for the stock. It's still far too soon to declare victory. There are other great stocks to buy in the semiconductor industry. Keep any bet on an Intel turnaround small given the long and steep uphill battle ahead.
