conference date: November 5, 2009 @ 2:00 PM Pacific Time
for quarter ending: October 25, 2009 (third quarter fiscal 2010)
But I own a competitor, AMD.
Overview: Sequentially improved revenues brought a return to profitability and even a revenue increase from the year-earlier quarter.
Basic data (GAAP) :
Revenues were $903.2 million, up sequentially 16% from $776.5 million and also up 1% from $897.7 million year-earlier.
Net income was $107.6 million, up sequentially from negative $105.3 million, and up 74% from $61.7 million year-earlier.
EPS (earnings per share) were $0.19, up sequentially from negative $0.19, and also up 72% from $0.11 year-earlier.
Revenue expected up 2% sequentially [about $921 million]. GAAP Gross margin between 40 and 42%, with GAAP operating expenses expected around $305 million.
Continued supply constraints on 40 nm products.
There was "healthy market demand across the board. Tegra mobile-computing solution is shipping." The Microsoft Zune HD and Samsung M1 are Tegra based. "Today's most important computer problems are parallel."
Non-GAAP numbers given as: net income $110.3 million, down slighly from $111.4 million year-earlier. EPS $0.19, down slighly from $0.20 year-earlier.
GAAP results include a $25.1 benefit from insurance reimbursements that are not included in non-GAAP results.
There was "growing enthusiasm" for the Tesla platform for servers and cloud-computing.
Gross margin was 43.2% GAAP, 41.0% non-GAAP.
The new Fermi architecture for CUDA parallel computing was introduced. The Nexus tool is now integrated into Microsoft Visual Studio so Fermi CUDA can be written in Visual C++. Also launched RealityServer for streaming 3D applications. "Momentum behind parallel computing is unlike anything in the industry."
Cost of revenue was $511.4 million, leaving gross profit of $391.8 million. Operating expenses of $283.9 million included $197.9 million for R&D and $86.0 million for Sales, general and administrative. Leaving operating income of $107.8 million. Interest and other income was $2.4 million. Income tax expense was $2.6 million.
Core GPU business was particularly strong, up almost 25% sequentially. Desktop up 19%, notebook up 41% sequentially. There were supply constraints, particularly for 40 nm products. Professional business up 11% sequentially. 55 nm yields improved. Demand is outstripping supply.
Guidance has been low, how does that couple with supply constraints? Our forecasts are gathered from the field, from the OEMs and channel. We set wafer start plans from that. Demand has been more robust than anyone could foresee. With 12 to 14 week lead times, it has been difficult to keep up with unanticipated demand.
Chipsets? Primary product is MCP61 is only single-chip, lowest cost solution for AMD platform. AMD market is mostly mainstream desktop, and their CPUs lost share there. I don't think we lost share within the AMD market, we think AMD lost share to Intel. We have a dispute with Intel, so we cannot build next-generation Intel chipsets. Our chipsets exceed Intel's in capability, so we expect to do well, but we have no intention of building a DMI-based (next generation Intel) chipset.
Fourth quarter demand? Demand is really strong. Last quarter we were supply-constrained from the very first day of the quarter. We lost market share because of that. We were not able to fulfill several million units of demand at the end of the quarter. It could be because of Windows 7 and the aging of PCs leading to a need for refreshes. We expect to be supply constrained, the forecast assumes supply constraints.
When does the lack of DMI architecture support start impacting revenue? Revenues would peak in Q3 or Q4 next year, then it would ramp down.
Margin improvement details? Everything grew. 65 nm was a drag earlier in the year, we were all 55 and 40 nm products last quarter. Growth of professional division also contributed to margins, partly because we sell software components that add value to the chips. Royalty payments from consumer electronics market also have high margins.
There are lots of Tegra products being introduced next year, in a variety of formats including cell phones, tablets, automotive, etc. We see Tegra doubling every quarter off a small base. In a couple of quarters we will start breaking it out for you.
Supply constraint implications? Visibility is better when we are constrained, it forces customers to be clear with us about their needs. It does constrain our growth, we could have done $50 to $100 million more in revenues last quarter, and same in this quarter. It definitely contributed to share loss, 2 or 3 points worth.
We expect every major segment to grow rapidly in 2010.
When might 40 nm supply improve? Yields are improving, they are better today than at the beginning of last quarter. TSMC is focused on this. Yield enhancement is our number one strategic objective. Professional segment, however, is not supply-constrained.
Any sell through of previously-reserved inventory? No. We shipped everything we had in Q2 and Q3. Our guidance would be the same for this quarter whether or not it is a 14 week or 13 week quarter (and an explanation of why).
The Fermi architecture is revolutionary. People around the world are waiting for it. The competition's chip (AMD's) is just the old architecture plus DX 11. We will tell you about all the great graphics features when we launch. There are many derivatives of Fermi for specific markets.
MCP79 and MCP89 are going to differentiate products, not everyone is going to build exactly the same thing (using Intel chipsets).
NVIDIA Analyst Conference Summaries Main Page
NVIDIA Investor Relations page
Openicon NVDA main page