In June, I argued that NVIDIA (NVDA 4.35%) does not need to continue developing cellular baseband technology given the new focus areas for its Tegra processor lineup. Then, in September, I dug deeper into the numbers to make the case that the financials of maintaining an internal baseband effort for NVIDIA's Tegra business did not even make sense.

In this article, I'd like to present more evidence that NVIDIA does not need its Icera division.

Guess whose LTE modem is inside of the Nexus 9?
According to AnandTech's Joshua Ho (via Twitter), the LTE-capable variant of the new NVIDIA Tegra-powered Nexus 9 tablet features a Qualcomm (QCOM 1.62%) MDM9x25 LTE modem -- not an NVIDIA stand-alone modem. This is important for a couple reasons.

First, this shows NVIDIA can successfully target devices that require high applications processor performance without offering an accompanying stand-alone baseband. In the low end and midrange of the market, integration is important for a variety of reasons, but since NVIDIA has been quite clear that it does not intend to play in those markets, that's not an issue here.

Next, and perhaps negatively for NVIDIA, this development shows that companies without a leadership solution in baseband will have a tough time winning the design. Note that NVIDIA has the Icera i500 stand-alone modem in its arsenal, but Google and HTC passed on it to go with the Qualcomm device for the Nexus 9.

It's hard enough to justify the baseband investment in the case in which every cellular-capable Tegra-powered tablet is also paired with an Icera modem, but when large designs using Tegra forgo the Icera, the investment makes very little sense.

Another data point to consider
Along with the fact that NVIDIA won the apps processor slot of the Nexus 9 but didn't win the cellular baseband, note that the company has yet to launch a follow-on to the Icera i500 baseband, which was announced in early 2013.

One major selling point of the unconventional software-defined radio approach that NVIDIA and Icera took with the i500 is that it was supposed to be possible to upgrade the modem to more advanced wireless standards over time. According to NVIDIA's website, the modem is upgradeable from category 3 LTE to category 4 LTE-Advanced, but this is pretty well behind the latest-and-greatest modems from the likes of Qualcomm and Intel (INTC 0.61%).

Failure to launch an i500 successor at the next CES in January would be, in my humble view, the final signal that NVIDIA has exited cellular baseband processors.

Would this be a bad thing?
Absolutely not. NVIDIA makes its money by selling high-performance graphics processors and the attendant software solutions. In its core high-margin GPU business, NVIDIA is unmatched, and its margins and overall profitability there are very attractive.

NVIDIA seems to still want to focus Tegra on automotive and higher-end gaming-focused tablets, and if it can make the economics of Tegra work, then that's fantastic.

Indeed, given that NVIDIA doesn't seem to need to be in the baseband business, and would have a hard time competing with Qualcomm, Intel, and MediaTek in the segment, I continue to believe that exiting basebands is actually key to making the Tegra economics work in the long run.