Advanced Micro Devices (AMD)
Our previous two articles looked at why AMD would want to not directly own its fabrication assets, so we’re not going to go too far into that here. In short, it is very expensive to operate semiconductor manufacturing lines, hard to increase their capacity, and incredibly expensive to move to a smaller process (e.g: 60nm to 45nm). As demand for AMD products changes, they can either have unused capacity (fab lines with nothing to do) or not enough capacity (can’t make as many parts as they can sell). Doing this in the same market as larger competitor Intel, leaves little room for error.
By spinning off their fabrication assets into something they are (for the time being) calling The Foundry Company, AMD can focus on designing products while not only getting, but freeing up a whack of cash with which to do so. That isn’t to say they are turning their back on the manufacturing industry: AMD will own just over 44% (this just in: make that 34.2%, more details about the deal in the next article) of the new company money-wise, and half of it voting wise. The Foundry Company (which I will just call TFC, with apologies to my fellow Toronto Football Club supporters) will be a consolidated subsidiary, meaning whatever happens to it, happens to AMD, but AMD only feels 34% of it.
The advantage isn’t just that AMD is sharing the risk and expenses with other investors (Mubadala and ATIC, more on them below) but that much like AMD can now focus on designing products, TFC will have separate management and employees focused on manufacturing not just AMD parts, but offering contract foundry services to other semiconductor designers. TFC will have the opportunity to grow, even if AMD demand is flat, by making chips for customers in all kinds of tech industries, from cheap toys to high-end medical equipment. And 34% of that business could help see AMD through periods where their own chips aren’t doing too well.