For many years, chip giant Intel's (INTC 0.02%) internal chip manufacturing technology had been a source of strength. The company's teams consistently developed more advanced technologies than what the competition could and brought those technologies to market earlier than they did.

Over the last five years, though, the company's manufacturing technology has gone from an asset to a liability. Intel's competitors are moving at a much faster pace than Intel is, something that has eroded the chip giant's advantage in chip manufacturing over the years.

An Arm executive holding a wafer of Intel 10nm test chips.

Image source: Intel.

Beyond that, Intel's manufacturing stumbles have led to significant product delays and cancellations, stifling the company's product-level innovation and weakening its competitive positioning. 

Put bluntly, I'm still shocked by how poorly things have been going for Intel's manufacturing group. 

This leads me to think that Intel should, at the very least, consider spinning off or shutting down its internal manufacturing operations and rely on third parties for chip manufacturing work going forward. 

A question I'd like to provide my thoughts on is this: When should Intel throw in the towel on its own manufacturing efforts?

If 7-nano flops, it should be game over

Intel's struggles began with its 14-nanometer technology, which was delayed by about six months. Not only was the technology late, but by the time Intel brought it into mass production, the yield rates were so low that the company's gross profit margins were negatively (and significantly) impacted. 

In both 2016 and 2017, Intel reported significant cost reductions as it has improved the yield rates on that technology. 

Although things didn't go very well with 14-nanometer, they've been downright disastrous with 10-nanometer. Back in late 2013, Intel told investors that it expected its 10-nanometer technology to be ready for mass production by the end of 2015, presumably for product launches in the first half of 2016. 

Now, after multiple delays, Intel is talking about first 10-nanometer product availability sometime in the second half of 2018 (though I expect these will be low-volume parts shipped solely for the sake of Intel claiming they met their targets). 

Things have gone from bad at the 14-nanometer generation to much worse during the 10-nanometer generation. The trend, unfortunately, isn't encouraging. 

What I would propose to Intel's management and board of directors, is the following strategy: If the development of Intel's 7-nanometer manufacturing technology doesn't go much more smoothly than the development of 10-nanometer, then the company should view this as a sign that things are so badly broken within the company's manufacturing organization that it's best to just abandon the effort altogether. 

Making such a decision wouldn't be easy, especially given how intertwined Intel's product development groups and its manufacturing group are, but management and the board of directors need to make pragmatic decisions that maximize Intel stockholder value.

If Intel can improve its product execution by building its chips on third-party manufacturing technologies, then I'd strongly advocate for Intel to do that. Intel makes its money by selling finished products to consumers, not manufacturing technology, so the focus needs to be on what manufacturing strategy ultimately allows Intel to win in the marketplace.