Advanced Chart
  • Thu, Feb. 18, 9:22 AM
    | Thu, Feb. 18, 9:22 AM | 1 Comment
  • Wed, Feb. 17, 5:38 PM
    • Top gainers, as of 5.25 p.m.: IM +23.5%. AERI +15.3%. NVDA +7.4%BRCD +5.6%. SGY +5.2%.
    • Top losers, as of 5.25p.m.: JACK -19.1%. BKD -10.0%. ZEN -8.8%. HMY -8.3%. ARRS -5.1%.
    | Wed, Feb. 17, 5:38 PM
  • Wed, Feb. 17, 4:51 PM
    • In addition to soundly beating FQ4 estimates, Nvidia (NASDAQ:NVDA) is guiding for FQ1 revenue of $1.26B (+/- 2%), above a $1.23B consensus.
    • Top-line performance: As was the case in FQ3, strong gaming GPU demand (in spite of a weak PC market) was a major tailwind: Total GPU revenue rose 10% Y/Y to $1.18B, with gaming GPU revenue rising 21%. Quadro (workstation GPU) revenue rose 7% to $203M, and Tesla/GRID (HPC server/cloud) GPU revenue rose 10% to $97M.

      Tegra app processor revenue rose 40% to $157M, fueled by a 68% increase in automotive revenue to $93M (the result of both infotainment module Tegra sales and R&D contracts). The Intel deal once more produced licensing revenue of $66M.
    • Financials: Non-GAAP gross margin was 57.2%, +70 bps Q/Q and +100 bps Y/Y, and above a guidance midpoint of 57%. FQ1 GM guidance is at 57.5% (+/- 50 bps). Operating expenses rose 6% Y/Y to $445M (compares with 12% revenue growth). 4.3M shares were repurchased through an accelerated buyback. Nvidia ended FQ4 with $5.04B in cash and $1.41B in convertible debt.
    • NVDA +5.6% after hours to $29.20.
    • Nvidia's FQ4 results, earnings release, CFO commentary (.pdf)
    | Wed, Feb. 17, 4:51 PM | 7 Comments
  • Wed, Feb. 17, 4:22 PM
    • Nvidia (NASDAQ:NVDA): Q4 EPS of $0.52 beats by $0.20.
    • Revenue of $1.4B (+12.0% Y/Y) beats by $90M.
    • Shares +7.9%.
    | Wed, Feb. 17, 4:22 PM | 9 Comments
  • Tue, Feb. 16, 5:35 PM
  • Fri, Jan. 15, 12:35 PM
    • On a day the Nasdaq is off 4.1%, companies with strong PC exposure are seeing even bigger losses after Intel (down 9.2%) beat Q4 estimates and provided healthy 2016 sales guidance, but also reported a slowdown in server CPU sales growth and offered cautious earnings call remarks about Chinese demand and macro trends.
    • Intel's Client Computing Group (provides PC/mobile CPUs) only saw a 1% Y/Y revenue drop in Q4 in the face of weak PC demand. However, volumes were down 16%, with desktop, notebook, and tablet CPU volumes respectively falling 9%, 10%, and 33%. A 17% increase in ASPs minimized Intel's revenue drop. Earlier this week, IDC and Gartner respectively estimated PC shipments  fell 10.6% and 8.3% Y/Y in Q4.
    • The list of PC-related names seeing big losses includes Microsoft (MSFT -4.8%), Nvidia (NVDA -6.7%), AMD (AMD -8.1%), HP Inc. (HPQ -5.2%), Micron (MU -8.9%), Seagate (STX -6.4%), and Western Digital (WDC -6.4%). AMD reports on Jan. 19, Microsoft and Western on Jan. 28, and Seagate on Jan. 29.
    • In other Microsoft news, the company has carried out price cuts of up to 17% for Azure cloud computing services, following new price cuts from Amazon, and is reportedly looking to acquire Indian mobile ad network owner InMobi. Google was rumored to be in buyout talks with InMobi last year.
    • AMD, meanwhile, has begun shipping its Opteron A1100 (Seattle) ARM server CPU. Seattle, first announced in 2013, sports up to eight 64-bit ARM Cortex A-57 CPU cores, and enters a market also targeted by AppliedMicro, Cavium, and Qualcomm.
    | Fri, Jan. 15, 12:35 PM | 15 Comments
  • Thu, Jan. 14, 9:43 AM
    • Arguing fundamentals don't justify the company's current valuation following a 65% 2015 gain, Barclays' Blayne Curtis has downgraded Nvidia (NVDA) to Underweight, and cut his target by $2 to $23.
    • Among chip stocks, Curtis is more partial to Cirrus Logic, Cavium, and Skyworks. Nvidia, buoyed in recent quarters by strong high-end GPU demand for gaming PCs, currently trades for 21x an FY17 (ends Jan. '17) EPS consensus of $1.36.
    | Thu, Jan. 14, 9:43 AM | 15 Comments
  • Tue, Jan. 12, 5:56 PM
    • IDC estimates global PC shipments fell 10.6% Y/Y in Q4 to 71.9M, just a slightly smaller drop than Q3's 10.8%. Gartner is once more a little less downbeat, estimating shipments fell 8.3% to 75.7M.
    • For the whole of 2015, IDC estimates shipments fell 10.4% to 276.2M. The firm notes adding detachable tablets to its count would respectively provide ~6% and ~3% boosts to Q4 and 2015 growth rates.
    • The Q4 figures suggest Windows 10 (available as a free upgrade to Windows 7/8 users) and Intel's (NASDAQ:INTC) Skylake CPU launch provided only a modest sales lift. IDC: "The PC market continued to face persistent challenges [in 2015] from longer-PC lifecycles and competition from mobile phones and tablets, despite the slowing growth in those markets. However, economic issues like falling commodity prices and weak international currencies, as well as social disruptions in EMEA and Asia/Pacific that disrupted foreign markets were a larger factor for 2015."
    • IDC VP Loren Loverde still thinks PC replacement rates will pick up in 2016. "Commercial adoption of Windows 10 (NASDAQ:MSFT) is expected to accelerate, and consumer buying should also stabilize by the second half of the year. Most PC users have delayed an upgrade, but can only maintain this for so long before facing security and performance issues." Counting detachable tablets, IDC expects 1%-2% 2016 growth. Gartner forecasts a 1% decline.
    • Lenovo (OTCPK:LNVGY) remained the PC market's top player: IDC estimates Lenovo's shipment share rose 140 bps Y/Y in Q4 to 21.4%. #2 HP's (NYSE:HPQ) share rose 10 bps to 19.9%. #3 Dell's rose 70 bps to 14.1%.
    • Apple (NASDAQ:AAPL) and Asus (OTC:AKCPF) were tied at #4 with 7.9% shares - Apple's share rose 110 bps, and Asus' 90 bps. Apple's revenue share could be closer to 15%. Non-top 5 vendors, hurt by their relative lack of scale, collectively saw their shipment share drop 420 bps to 28.7%
    • Related tickers: NVDA, AMD, MU, STX, WDC
    | Tue, Jan. 12, 5:56 PM | 104 Comments
  • Tue, Jan. 5, 5:12 PM
    • In this year's torrent of CES week product announcements, Qualcomm (NASDAQ:QCOM) has unveiled the Snapdragon 820A automotive processor line. The chips target in-car infotainment systems, and feature specs competitive with Snapdragon SoCs aimed at high-end smartphones - a 14nm manufacturing process is used, as is Qualcomm's next-gen Kryo CPU core, Adreno 530 GPU, Zeroth A.I./neural networking tech, and an X12 LTE modem (600Mbps peak download speeds).
    • Qualcomm asserts the chips allow in-car infotainment systems to be upgraded via both hardware and software updates. They take aim at Nvidia (NASDAQ:NVDA), whose Tegra processors own a large chunk of the auto infotainment processor market, as well as Intel and Texas Instruments. Meanwhile, the Snapdragon 602A line, which was announced at CES two years ago, has been designed into select 2017 Audi cars.
    • Also announced: The Snapdragon X5, a low-cost 4G modem aimed at IoT devices. The chip supports 150Mbps peak download speeds, higher than those delivered by two previously-announced 4G modems meant for IoT hardware, and has a built-in ARM CPU.
    • Other announcements: 1) A smart home reference platform featuring the low-power/quad-core Snapdragon 212 processor. 2) A Bluetooth Smart SoC (developed by recently-acquired U.K. chipmaker CSR) that targets power-constrained embedded devices. 3) An expansion of Qualcomm's partnership with drug giant Novartis, through which Novartis will use Qualcomm's connectivity offerings to power its Breezhaler inhaler line.
    | Tue, Jan. 5, 5:12 PM | 15 Comments
  • Tue, Jan. 5, 4:38 PM
    • Mobileye (MBLY -9.6%) closed slightly above $36 after Nvidia (NVDA +1.6%) unveiled the Drive PX2, a second-gen processing module for driver-assistance and autonomous driving systems that's declared to be "the world's most powerful engine for in-vehicle artificial intelligence."
    • Nvidia: "DRIVE PX 2 can process the inputs of 12 video cameras, plus lidar, radar and ultrasonic sensors. It fuses them to accurately detect objects, identify them, determine where the car is relative to the world around it, and then calculate its optimal path for safe travel. This complex work is facilitated by NVIDIA DriveWorks, a suite of software tools, libraries and modules that accelerates development and testing of autonomous vehicles."
    • Nvidia also claims the PX2 can handle up to 24T A.I./deep learning operations/second (over 10x as many as the prior-gen Drive PX), thanks to specialized instructions, and that its GPU architecture can handle up to 8T operations/second for general floating-point operations.
    • The module contains two Tegra processors and two GPUs based on Nvidia's next-gen Pascal architecture. Nvidia asserts the original Drive PX has been used by 50+ automakers, tier-1 auto suppliers, developers, and research institutions for autonomous driving R&D.
    • Mobileye, which currently dominates the driver-assistance system market via modules sporting its EyeQ SoC (also uses A.I./deep learning algorithms), had a rough day. Shares sold off last March after Freescale unveiled a rival SoC, and last month after a leaked Elon Musk e-mail that suggested Tesla is open to discontinuing its Mobileye supplier deal (Tesla was quick to downplay the report).
    | Tue, Jan. 5, 4:38 PM | 4 Comments
  • Tue, Jan. 5, 4:46 AM
    • Early news from the Consumer Electronic Show: Nvidia (NASDAQ:NVDA) has revealed a new lunchbox-size super-computer for self-driving cars, saying Volvo (OTCPK:GELYY) will be the device's first customer.
    • The Drive PX 2 has computing power equivalent to 150 MacBook Pro computers, and can deliver up to 24T "deep learning" operations - allowing the computer to use artificial intelligence to program itself to recognize driving situations - per second.
    • Partnerships between automakers and Silicon Valley on self-driving technologies appear to be taking center stage at the annual tech conference.
    | Tue, Jan. 5, 4:46 AM | 9 Comments
  • Mon, Jan. 4, 1:54 PM
    • AMD is underperforming amid a market selloff on a day in which the company has unveiled (ahead of CES) its Polaris GPU architecture. Polaris, due in mid-2016, relies on a 14nm FinFET manufacturing process, and supports high dynamic range monitors, HDMI 2.0a, 4K h.265 video encoding/decoding, and DisplayPort 1.3. AMD also promises "industry-leading performance-per-watt."
    • Official details about Polaris remain limited. The architecture succeeds AMD's GCN 1.2, which underpins the company's R9 Fury GPUs (launched in mid-2015), and promises command processor, geometry processor, memory controller, L2 cache, and display engine improvements. AMD claims a Polaris-based card will consume just 85W of power to play Star Wars: Battlefront at 60fps on the medium preset, less than the 140W consumed by a card using Nvidia's (NVDA -2%) mid-range GTX 950 GPU.
    • The rumor mill has indicated Polaris-based GPUs (codenamed Arctic Islands) will support second-gen High Bandwidth Memory (HBM2), and include a flagship GPU (Greenland) featuring up to 18B transistors and supporting up to 32GB of memory. Arctic Islands/Polaris is expected to be used by PC/mobile GPUs, APUs, and semi-custom chips. There's also speculation Nintendo's next-gen NX console will feature an AMD semi-custom SoC relying on Arctic Islands.
    • Nvidia plans to counter Polaris/Arctic Islands with its Pascal GPU architecture, which will leverage TSMC's 16nm FinFET process.
    • Last week: MKM sees Nvidia, AMD benefiting from healthy GPU pricing
    | Mon, Jan. 4, 1:54 PM | 72 Comments
  • Fri, Jan. 1, 4:53 AM
    | Fri, Jan. 1, 4:53 AM | 27 Comments
  • Dec. 30, 2015, 7:11 PM
    • Hard-luck AMD has seend only "modest" Q/Q price declines for its high-end R9 GPUs, reports MKM's Ian Ing. He sees this boding well for the ability of AMD's GPU line to stay competitive competitive following its mid-2015 refresh.
    • Nvidia (NASDAQ:NVDA) is seeing "even more favorable pricing trends," though parts of its GPU lineup are up to a year old. "Some higher end cards are up 1% to 6% sequentially quarter-to-date (January fiscal), although we didn’t see NVDA’s U.S. retail up by the widely reported 15% in China due to alleged shortages seen in some industry blogs." Digitimes previously reported strong demand (particularly from China) had led to shortages and markups for some Nvidia-based graphics cards.
    • Ing sees several factors propping up GPU pricing. "Both NVDA and AMD are addressing the fast-growing gaming [addressable market] that benefits from increased leisure/ entertainment mindshare vs. traditional consumption, a strong season for AAA titles (STAR WARS Battlefront, Call of Duty, Black Ops 3), and the social and professional aspects of eSports growth."
    • Nvidia is coming off an October quarter (FQ3) in which its GPU sales rose 12% Y/Y to $1.11B, as a 40% increase in gaming GPU sales offset PC OEM, workstation, and HPC/cloud weakness. AMD, hurt by PC weakness and share loss, saw its PC CPU/GPU division's sales rise 12% Q/Q and fall 46% Y/Y in Q3 to $424M.
    • Looking ahead to 2016, AMD reportedly plans to launch a next-gen GPU line codenamed Arctic Islands in the summer. The GPUs are expected to use Samsung/Globalfoundries' 14nm manufacturing process and support second-gen High Bandwidth Memory (HBM2), with a flagship GPU (codenamed Greenland) featuring up to 18B transistors and supporting up to 1TB/s of memory bandwidth.
    • Nvidia, for its part, plans to launch its Pascal GPU architecture at some point in 2016. Pascal will use TSMC's 16nm manufacturing process, deliver (via HBM2) ~1TB/s of memory bandwidth, and support Nvidia's NVLink interconnect, which is promised to deliver much higher speeds than PCIe 3.0.
    | Dec. 30, 2015, 7:11 PM | 35 Comments
  • Dec. 22, 2015, 7:10 PM
    • An ITC administrative law judge (ALJ) has ruled Nvidia (NASDAQ:NVDA) infringes three Samsung (OTC:SSNLF) patents. The case will now go before the full commission.
    • The news comes eight days after the full ITC commission upheld an ALJ's ruling declaring Samsung and Qualcomm didn't infringe two Nvidia patents, and that a third patent (though infringed) is invalid. Samsung had countersued Nvidia before the ITC, and both companies are suing each other in Virginia.
    • Nvidia, which had argued Samsung's patents cover older chip technology that's no longer used: "We look forward to seeking review by the full ITC which will decide this case several months from now." One of Samsung's patents is set to expire next year.
    • NVDA -0.7% after hours to $32.70.
    | Dec. 22, 2015, 7:10 PM | 3 Comments
  • Dec. 14, 2015, 6:48 PM
    • Two months after an ITC administrative law judge ruled Samsung (OTC:SSNLF) and Qualcomm (NASDAQ:QCOM) didn't infringe two Nvidia (NASDAQ:NVDA) GPU patents and deemed a third patent that they were found to infringe to be invalid, the ITC commission has upheld the judge's findings.
    • Nvidia, whose Tegra processors directly compete against Qualcomm's Snapdragon processors and indirectly compete against Samsung's Exynos processors: "While today’s ruling ends this particular case in the ITC, we remain firm in our belief that our patents are valid and have been infringed and will look to appeal the ITC’s decision."
    • Samsung has counter-sued Nvidia in both a Virginia court and before the ITC; those cases remain pending. Each company has also sought to have the other's patents invalidated by the USPTO.
    | Dec. 14, 2015, 6:48 PM | 2 Comments