NVIDIA FY 2015 Q2 Financial Results

by Brett Howse on 8/10/2014 11:10 PM EST


Back to Article

  • Stephen Barrett - Sunday, August 10, 2014 - link

    I wonder why tegea is so popular with automotive Reply
  • A5 - Sunday, August 10, 2014 - link

    I think the Tegra chips are cheaper than similarly-powered competition, and power consumption isn't really a concern when you're connected to a giant alternator.

    They also got a pretty early design win with Audi, which may make it easier to get in with other companies. If nothing else, getting in the other VW brands would be a big slice of the market for those systems.
  • jjj - Monday, August 11, 2014 - link

    They have the Golf and the soon to be released Passat.
    As for why they got the wins ,they got into it early and JHH likes cars. Plus they got a big interest in using GPUs into robots,including self-driving cars, could become a huge segment for them. Others failed to notice the segment soon enough or have good enough products so they seem to have a bit of a head-start.
  • Dribble - Monday, August 11, 2014 - link

    I guess software- an arm soc is required but getting it plugged into all the cars systems, and running all the correct software is probably the hardest part. Nvidia are strong when it comes to writing drivers and providing support. Reply
  • Hrel - Tuesday, August 12, 2014 - link

    Everything A5 said is right, but to add to it. Tegra is a lot better at rendering graphics quickly than comparable solutions, and pretty much everything in a car's console is graphics based, not text based, because it needs to be easy to read and touch. So Nvidia has found a really strong niche application for their hardware. Hell, Tesla is using it too! Reply
  • twotwotwo - Sunday, August 10, 2014 - link

    You know, last round almost no one but Qualcomm (and, of course, Apple) got into highish-end phones in the US, because no one else had great LTE basebands and maybe because Krait/Adreno were pretty good and pretty power-efficient. (And even Apple's using a Qualcomm baseband.)

    And US phone wins may have had a halo effect; if you've done one Snapdragon integration and made one big contract with Qualcomm, the existing integration and relationship might make it easier to go with them for something else.

    Krait's probably no longer a factor because it sounds like the first 64-bit non-Apple ARMs will all be Cortex-A5x implementations. I wonder if basebands will quit working against everyone but Qualcomm, either because the Icera stuff works out for NVidia or because of acceptable licensed basebands or whatever. And, if US phones don't get more diverse, I wonder if NVidia or Samsung's chip division or Intel manage to get profitable-enough wins in tablets, non-US phones, mobile gaming, whatever to keep everyone investing.
  • jjj - Monday, August 11, 2014 - link

    When talking Tegra you got to remember that last year was very poor and they just launched the Shield Tab. Not sure if they accounted for any Shield Tab revenue in Q2 but the thing has high ASP. You got the device at 300$ and costly accessories so ASP will be some 330-350$. If they ship 200k tabs ,that's up to 70 extra millions At the same time a TK1 got to be some 25-30$ so the tab compensates for quite a bit of chip volume. Ofc hard to say how many tabs they will manage to sell,they do seem to be expanding to other markets fast (shipping in a few days in the UK) so 200k units in a quarter seems doable if they expand to enough markets.Could be even higher volume and then coupled with a few more wins and new GPUs, Q3 could end up quite a bit better than 1.2B. Reply
  • MartinT - Monday, August 11, 2014 - link

    Fascinating to see that more than half of their profits are the result of the Intel licencing agreement, while all their 'real' products merely eek out another 6x million on revenues of 1+ billion. Reply
  • jjj - Monday, August 11, 2014 - link

    That's not accurate.
    Operating income before interest and other was 163 million and income before tax was 154.6.
    Then again ,at least they make money unlike AMD that usually just breaks even on an operating level on GPUs.
    They also invest plenty in R&D , 337 millions in Q2, over 25% of revenue in R&D is quite a lot
    And Tegra is likely to be losing plenty of money for now on low volume and lower margins.
    How about Intel's mobile division 1.2B loss on 51 million of revenue in Q2.
  • MartinT - Monday, August 11, 2014 - link

    "That's not accurate. [...] income before tax was 154.6."

    That's true, thanks for the correction!
  • TheJian - Monday, August 11, 2014 - link

    Don't forget tegra has been losing money since inception. T1-4 were just test runs and delay tactics until K1. All they get from the first 4 is experience and name recognition (which is good, pretty much had to be done) and a gaming platform for android. From K1 on though, desktop gpus change the game and in house cpu for xmas devices puts them on a level field with Qcom from here on out. Everything that goes with desktop gpus comes to mobile (drivers, game devs already using it in the same year, feature support same as desktops, cuda, physx, full OpenGL 4.4 etc etc) and the rest will have trouble matching all of this. Qcom has zero experience in gaming and drivers, and devs have pretty much zero experience with their gpus. NV/AMD have 20yrs of that stuff (which is why I hope AMD gets in this arm stuff soon! I'm not talking servers, they should have went mobile first), so both have a gpu advantage for quite some time. Now that K1 will be in nexus 9 and likely other gaming oriented devices, I think they'll quit losing money and start finally adding to the bottom line with tegra over the next few quarters after it ramps.

    So their gpus are making more than you think, it's just that tegra is stealing income. It's worth it though, because now GPU is taking over the modem as gaming becomes something we all want to do on mobile (90% of revenue on googleplay is games, apple's in the 80's, amazon mid 60's etc). With google now pushing games with AndroidL+AEP and trying to close the gap between desktop and mobile, we'll start to see the NV/Google partnership bloom. We saw the start of this at Google's IO event, as everything there was on K1. Others will have to adopt K1's gpu IP or lag as S805 shows already. S810 is off the shelf so that won't help either (that will hurt batter/cpu and their gpu won't catch NV either). NV will rule the tablets for gaming for at least a year or more now until we see Qcom's answer after S810 (likely well over a year away and will miss the next Nov nexus also IMHO).

    Denver will be in HTC Nexus9 and qcom has no Nov answer. Next year you'll have a 20nm maxwell (for Nov nexus I mean) to contend with and as I said whatever is after S810 will be too late to make it into a Nov launch. The IN HOUSE 64bit from Qcom will probably start to ship about a year from now or a tad later, making it a few months shy of 4.5-5mo for putting it into a product, that would end up a late Dec/Jan product in retail and google won't accept that :) They want xmas sales, wisely IMHO, so launch in Nov. I'd expect a shield rev2 handheld soon or for xmas (denver if they wait to rev2 for then) as it's a no brainer and should basically just drop in to the T4 shield housing with little change. Also matches the 1440x810+4GB specs we've heard, which clearly are not the shield tablet's specs. This would be a slight bump from the old shield and massive gpu upgrade of course. Qcom said S810 is 6 months after S805, and in house 64bit 6 months or more after S810. Anandtech thinks 2nd half 2015 for devices with Qcom in house cpu, but I'll believe that when I see it. Even the change in their roadmap landed them around least 6-8 months behind NV's 64bit denver arriving in nexus in Nov. Considering it really does look like they got caught with their pants down on 64bit, I'd be pretty surprised if they can get into a Nov Nexus next year with the IN HOUSE chip.
  • Sushisamurai - Monday, August 11, 2014 - link

    Is it me, or did I just lose a year by reading this article?

    If we're talking about 2015, it should be forward guidance, fiscal results should be 2014
  • Sushisamurai - Monday, August 11, 2014 - link

    But seriously, please fix your dates in this article, as you reference YoY gains from 2015 and 2014, pitting "this fiscal quarter" from 2016 and 2015...

    Time for an editor anand...
  • Black Obsidian - Monday, August 11, 2014 - link

    I'm not sure what your complaint is, but it sounds like you're unaware that many companies operate in a fiscal year that's ahead (sometimes by as much as a full year) of the calendar year. These results being for nVidia's Q2 2015 fiscal year is not a typo. Reply
  • Sushisamurai - Monday, August 11, 2014 - link

    Oopsies. That was my honest mistake there. My bad and no offence there. I'm actually invested in Nvidia, but the news/websites I read in terms of financials and reporting "simplify" the dating terminology, so they reference the current fiscal year (Q2 2014) vs last year (Q2 2013). I had to actually look up Nvidia's investor's site to see that they report it as Q2 2015. Wtf... Reply
  • Brett Howse - Monday, August 11, 2014 - link

    NVIDIA's fiscal year seems to always cause this kind of confusion, but the dates are correct. FY 2015 for NVIDIA began on January 27th 2014. That's why I specifically said it was for fiscal year 2015, but even so, it's easy to think it's a mistake. Reply

Log in

Don't have an account? Sign up now