IBM (NYSE:IBM) recently showed off working test chips manufactured in the company's upcoming seven-nanometer manufacturing technology. This has seemingly led some to believe that this is "bad news" for Intel (NASDAQ:INTC), which has traditionally led in bringing advanced manufacturing technology into high-volume production ahead of its peers.
Although this achievement is interesting and makes for great headlines, I'm not convinced that Intel shareholders should be particularly "worried" about this announcement. Let me explain.
The gap between research and high volume manufacturing
There's a significant difference between producing a single test chip and going into mass production. For example, back at Intel's 2013 investor meeting, the individual responsible for running Intel's Technology and Manufacturing Group, Bill Holt, said that Intel had test chips of its 10-nanometer manufacturing technology up and running.
And yet, it's looking like volume production of Intel's first 10-nanometer processors won't begin until sometime in 2016 for what's increasingly looking like a 2017 product rollout. That's a roughly three-year gap from when Holt said Intel had test chips on 10-nanometer running -- and the test chips may have been floating around even earlier -- to economically viable high-volume manufacturing.
So there's very clearly a significant time gap between "we have test chips in the lab," and "we're ramping production for sale of products."
A brief history lesson
Let's turn our attention back to IBM. Back in 2008, IBM and Advanced Micro Devices (NASDAQ:AMD) said that they were "first" to the 22-nanometer manufacturing technology by demonstrating a 22-nanometer SRAM cell. Intel, at the time, had "only" shown off a wafer of SRAM cells built on its then-upcoming 32-nanometer manufacturing technology.
Which company actually made it to high-volume manufacturing on the 22-nanometer manufacturing node first? Unsurprisingly, that would be Intel. The company went into high-volume production of its 22-nanometer FinFET manufacturing technology in the fourth quarter of 2011, and product was on the shelves by April 2012.
Products based on IBM's own 22-nanometer manufacturing technology -- the POWER8 processor -- didn't become available in systems until 2014.
Intel isn't unassailable, but it's very hard to beat
Intel is by no means perfect, even in the world of manufacturing technology. The company's delays in ramping its 14-nanometer manufacturing technology to acceptable yield rates, as well as the rumored delay of its 10-nanometer technology, highlight that ramping these technologies into high-volume manufacturing can bring unforeseen challenges.
However, I would argue that Intel is still solidly ahead of the pack in terms of ramping advanced manufacturing technologies at good yields. TSMC isn't expected to start high-volume production of its own 16-nanometer FinFET Plus technology until this month; and in a number of key metrics, I believe TSMC's process is behind Intel's 14-nanometer technology.
Samsung, which has been fiercely vying for high-end foundry business, managed to beat TSMC in ramping its own 14-nanometer technology, but started production later than Intel did. I believe that Samsung's 14-nanometer technology is also behind Intel's 14-nanometer technology in several key metrics.
I don't think Intel is invulnerable to competition; but at the same time, I'm not ready to buy into the notion that the rest of the world has caught up to, let alone leapfrogged, Intel as far as logic chip manufacturing technology is concerned.