About a year ago, I said that "it seems inevitable that NVIDIA (NVDA -2.94%) will wind down or sell" its cellular baseband operations. By the time CES 2015 rolled around, it was clear to me that NVIDIA was more or less finished in the cellular baseband market as by that point, it hadn't updated its cellular baseband offerings for nearly two years.
Color me unsurprised, then, to see that NVIDIA just announced that it will wind down its cellular baseband operations during its second fiscal quarter of 2016.
Why is NVIDIA winding it down?
In the press release announcing the wind-down, NVIDIA said that it had originally bought a cellular baseband developer known as Icera in order to "engage the smartphone revolution with a leading integrated application processor and modem platform."
However, NVIDIA says that it has "reshaped its strategy" in order to "focus on high-growth opportunities in gaming, automotive and cloud computing applications." These opportunities don't require integrated applications processors and basebands, which means there's not much reason for NVIDIA to continue to invest in developing its own baseband technology.
What will NVIDIA do about modems going forward?
Former NVIDIA senior director of investor relations Chris Evenden once said at an investor conference that connectivity technology was a key component of NVIDIA's automotive strategy. After all, NVIDIA is out to make cars smarter, and in today's world, it's well known that "smart" devices are also often "connected" devices.
In its press release announcing the wind down of Icera, NVIDIA says that its pre-existing Icera 4G LTE modem "meets the company's needs for the next year or more." However, NVIDIA also says that in the future, it will "partner with third-party modem suppliers."
Will NVIDIA sell it, or will it be wound down?
NVIDIA said that it is "open to a sale of the technology or operations." At this point, I don't think that there would be much customer interest in the operations themselves; I doubt that the revenue base nor the margins on that revenue base were high enough to attract such interest.
NVIDIA did claim that its Icera division had some pretty interesting technologies, but one concern here is that NVIDIA's latest LTE modem supports only category 4 LTE speeds. It's not clear what the company had in the pipeline, but NVIDIA's competition, such as Intel (INTC -4.54%) and Qualcomm (QCOM -1.35%), both plan to be shipping modems with much more advanced category 10 capability later this year.
My guess is that NVIDIA will ultimately end up winding down the operation, rather than selling it. If there were no buyers for Broadcom's (NASDAQ: BRCM) cellular division, I don't think there will be any buyers for NVIDIA's, either.
How does this affect the NVIDIA investment thesis?
It has been pretty clear for a while that NVIDIA was "done" with cellular baseband development, so I don't think that news is likely to take investors following the story closely by surprise.
The question that I would have for management at this point is what the company plans to do with the likely meaningful annual cost savings associated with shuttering this division. Will NVIDIA let it flow through to the bottom line (and potentially to investors via a dividend increase or buybacks), or will it take the savings and reinvest it in other areas of its business?
I'm sure that NVIDIA will lay it all out during its earnings call scheduled for May 7.