Analyst Conference Summary

conference date: August 11, 2016 @ 2:00 PM Pacific Time
for quarter ending: July 31, 2016 (Q2, second quarter fiscal 2017)

But I own competitor AMD
Forward-looking statements

Overview: Excellent results show ability to work outside the PC sales slump. Beat prior guidance by a good margin.

Basic data (GAAP) :

Revenues were $1.43 billion, up 9% sequentially from $1.31 billion, and up 24% from $1.15 billion in the year-earlier quarter.

Net income was $253 million, up 29% sequentially from $196 million and way up from $26 million year-earlier.

EPS (earnings per share) were $0.40, up 21% sequentially from $0.33, and 8x the $0.05 year-earlier.

Guidance:

For fiscal Q3 expects revenue within 2% of $1.68 billion. Gross margins 57.8% GAAP and 58.0% non-GAAP; operating expenses near $530 million GAAP, $465 million non-GAAP. Tax rate 21%. Capital expenditures $35 to $45 million.

Conference Highlights:

The dividend of $0.115 will be paid on September 16, 2016 to shareholders of record as of August 25, 2016.

Colette Kress, EVP & CFO, led the call.

Deep learning computations using GPUs has become an important driver of growth.

GPU segment revenue was $1.20 billion, up 11% sequentially from $1.08 billion, and up 25% y/y from $959 million.

Tegra segment revenue was $166 million, up 4% sequentially from $160 million and up 30% from $128 million year-earlier.

Other Segment revenue was $66 million, constant sequentially and y/y.

In addition to the segments, NVDA reports revenue by a number of market platforms:

Gaming revenue was $781 million, up 14% sequentially from $687 million, and up 18% from $660 million year-earlier. In the quarter introduced four members of its Pascal GPU family, led by GTX 1080. [WPM: there was probably built up demand in the quarter, and also some slowness in Q1 in anticipation]

Professional visualization (Quadro) revenue was $214 million, up 13% sequentially from $189 million, and up 22% from $176 million y/y.

Datacenter revenue was $151 million, up 6% sequentially from $143 million, and up 110% from $72 million y/y. Introduced Tesla M10 and Tesla P100. Microsoft Azure is now using Nvidia GPUs computing to cloud customers.

Automotive revenue was $119 million, up 5% sequentially from $113 million, and up 68% from $71 million y/y.

$163 million OEM and IP business, down 6% sequentially from $173 million and down 6% from $174 million year-earlier. This was due to declining PC demand.

Non-GAAP numbers: Net income $313 million, up 19% sequentially from $263 million, and up 65% from $190 million year-earlier. EPS $0.53, up 15% sequentially from $0.46, and up 56% from $0.34 year-earlier. 58.1% gross margin. Excludes $58 million in stock-based compensation and other non-cash and one-time items.

Cash and equivalents balance was $4.88 billion, up sequentially from $4.75 billion. $184 million cash from operations. $33 million was used for capital expenditures. $151 million free cash flow. Repurchased shares using $9 million. $62 million was used for dividends. Short-term debt $1.43 billion.

GAAP cost of good sold was $602 million, leaving gross profit of $826 million. Operating expenses of $509 million consisted of $350 million for R&D and $157 million for Sales, General and Administrative expense, and $2 million for restructuring. Leaving operating income of $317 million. Interest and other expense $0 million. Income tax expense $64 million.

Q&A:

Jen-Hsun Huang, CEO, answered the questions.

Datacenter computational applications or deep learning? Three markets: HPC, grid (graphics virtualization), deep learning. Most growth is coming from deep learning. HPC is still growing, but relatively stable. Grid is growing fast, but on a small base.

Yield improvement in Pascal? 40 nm was challenging, so we changed how we approach new nodes. So we are more ready for new processes, and we executed well on 16 nm FinFET with Pascal. We will continue to refine yields, but it is too early to predict yields and margins long term.

Q3 end markets with outside growth? All our businesses are growing, driven by a focus on key markets like deep learning and automotive. We have been investing heavily in deep learning for about five years.

Gaming? Our installed base is about 80 million. Only about one-third have even upgraded to Maxwell. Pascal is a bigger leap than past introductions, and includes VR. The PC gaming market has changed. The production quality is much higher than in the past, partly because the architecture of consoles and PC systems is now compatible. Excited that the next generation of game consoles will have a 2X boost. That means designers will be able to make games that require higher capability GPUs on PCs.

Usual meanderings about the wonders of e-sports.

Datacenter competitive landscape? GPUs made deep learning practical. Neural nets can be modeled on a GPU so neurons communicate with each other. It is not just the chip, it is the architecture, software, tools, etc. We have put all of our investment behind one architecture that works for all deep learning areas.

Usual meanderings about the wonders of autonomous driving.

Live video is a driver of datacenter traffic and computation. I believe every datacenter is going to be GPU accelerated. Energy efficiency is vital, and there is still a lot of workload that requires CPU. Intel CPUs with Nvidia GPUs is the way to go.

Sub $250 price point? We have taped out and ramped every Pascal GPU, but we have not introduced every one.

VR headset companies are doing a great shop getting devices out there. Content is being generated. We also expect practical applications in medicine, architecture, and design. We see a lot of VR adoption ahead. Our Pascal is a better renderer than the competitor's [AMD's] because it has multi-projection and real world physics simulation. "Our position in VR is really great."

OpenIcon NVIDIA Analyst Conference Summaries Main Page
NVIDIA Investor Relations page
Openicon NVDA main page

company page: Nvidia

Search

More Analyst Conference Pages: