Nvidia has been aggressively hyping its Tegra system on a chip (SoC) since Tegra 3 came out in 2012. But despite the hype, manufacturers are cautious about putting the chip in their devices. Why can't Tegra find its mojo?
Once upon a time, Nvidia’s Tegra ARM system on a chip was poised to propel the company into the stratosphere.
Back in August of 2012 Nvidia posted a rock-solid quarter, beating analyst expectations. During an earnings call, Nvidia CEO Jen Hsun Huang said Tegra 3 was pushing the company’s profits skyward. Nvidia’s third Tegra chip was in the recently released Nexus 7, a tablet that was selling with a frenzy only seen before in Apple’s iWorld.
In January, Nvidia pre-empted Qualcomm’s scheduled keynote with its own event the night before to announce its next-generation Tegra. Audiences were wowed, as the chip itself was packed full of GPU cores, and Qualcomm was no doubt annoyed – everyone was murmuring Tegra instead of being focused on Qualcomm’s next-generation Snapdragon.
Then Nvidia’s own reality distortion field started to wear off.
Less than a month after CES, Digitimes’ sources in the Taipei OEM scene said that Tegra 4 wasn’t building interest. Digitimes went on to claim that only only Toshiba – not exactly known as a significant player in the mobile market – had expressed interest in building Tegra 4 devices.
Then, last week, Reuters reported that ASUS was considering dropping the Tegra 3 from the refreshed Nexus 7 – one of the few successful tablets the chip was featured in – in favour of a Qualcomm SoC to be unveiled at the upcoming Google I/O event in March.
While the Tegra 3 had powered the smash-hit Nexus 7, and the well-received but moderate selling Microsoft Surface, it didn’t get soldered into much more than that. In the smartphone world, the lack of integrated baseband made manufacturers hesitant about committing it to their handsets. It was placed in the much loved HTC One X (but only the international model), and the LG Optimus 4X, but neither of the phones made a critical impact in the market in 2012.
So why can’t Tegra find its mojo?
Tegra 3 was met with hesitation from smartphone makers because of the lack of integrated cellular baseband. Phone manufactures who wanted to place it in their devices had to pair Tegra with a separate baseband, which meant sourcing parts from multiple vendors. In the tablet world, Tegra seemed to draw more power than Nvidia claimed it would and initial benchmark scores seemed to disappoint.
In terms of Tegra 4, vendors may be hesitant to place it in their devices because of the SoC’s lack of modern API support. The SoC’s GPU is based on the NV4X architecture (GeForce 6/7 Series, fixed function shaders) which means it won’t support the likes of DirectX 11.0, OpenCL, and OpenGL ES 3.0. Granted, the other announced Tegra 4i (mind you, not A15 but 28nm quad-core A9) does have an LTE baseband but the lack of API support is a definite handicap.
Given that the market is already fragmented into Qualcomm Snapdragon, Samsung Exynos, and Apple A-series camps is there really room for a late comer who’s already had a few tries at the game? Maybe, but it will take a heroic effort.