Analyst Conference Summary


conference date: October 14, 2015 @ 2:00 PM Pacific Time
for quarter ending: September 26, 2015 (second fiscal quarter 2016, Q2)

Forward-looking statements

Overview: Low demand, lower revenue and profits. But within guidance, and new products are selling well.

Basic data (GAAP):

Revenue was $527.6 million, down 4% sequentially from $549.0 million and down 13% from $604.3 million in the year-earlier quarter.

Net income was $127.3 million, down 14% sequentially from $147.7 million, and down 26% from $171.5 million year-earlier.

Diluted EPS (earnings per share) were $0.48, down 13% sequentially from $0.55, and down 23% from $0.62 year-earlier.


Fiscal Q3 2016, December quarter revenue expected up 3% to 7% sequentially, partly due to an extra week in the quarter. 69% gross margin. Operating expenses around $230 million. Other expense $7 million. 265 million fully diluted shares. 12% tax rate.

Full fiscal year capital expense will be $40 million, not the previous estimate of $25 million.

Conference Highlights:

Xilinx's 16 nm Zynq UltraScale+ multiprocessor SoC shipped a quarter ahead of schedule.

Xilinx announced a strategic collaboration with Qualcomm for ARM server platforms.

In the quarter wired and wireless sales increased, but industrial & defense sales decreased 12% sequentially. Broadcast down.

New product revenue is expected to accelerate in the December quarter. Zynq is leading sales growth. 28 nm is expected to grow significantly. 20 nm had good Q2 growth and should also grow strongly in Q3. 16 nm sales began in Q2. There are indications that Q3 will mark the end of the bottom of the sales cycle.

Revenues by end market: Communications and Data Center 41%; Industrial, Aerospace & Defense 41%; Broadcast, consumer and automotive 18%.

Revenue by product:

43% New products: UltraScale, Virtex-7, Kintex™-7, Artix™-7, Zynq™-7000, Virtex-6, Spartan-6
25% Mainstream products: Virtex-5, Spartan-3 and CoolRunner™-II
28% Base products: earlier Virtex, Spartan-II, Spartan, CoolRunner and XC9500
4% Support products: Configuration solutions, HardWire, Software & Support/Services

70.1% gross margin, up sequentially from 70.0%.

Cash, equivalents and long-term investment balance was $3.3 billion. $1 billion long-term debt and $581 million current debt. Operating cash flow was $133.7 million. Depreciation $12.5 million. Capital expenditures $5.4 million. $100 million of stock was repurchased. Stock based compensation expense was $26.7 million. The dividend payment required $80.2 million.

Revenue by geography: North America 30%; Asia 40%; Europe 20%; Japan 10%.

Cost of revenues (GAAP) was $157.6 million, leaving gross profits of $369.9 million. Operating expense total was $216.8 million, consisting of: research and development $130.2 million; selling, general and administrative $84.8 million; and amortization $1.8 million. Leaving operating income of $153.2 million. Interest and other expense was $9.2 million, and the income tax provision was $16.7 million.

The dividend of $0.31 per share will be payable on November 24, 2015 to holders of record on November 5, 2015.

In 2016 Zynq sales are expected to ramp in the automotive segment. 28 nm revenue should grow for the balance of fiscal 2016.

Backlog was up at the end of the quarter. Believes Chinese revenues will grow in December quarter.


Effect of extra week in December quarter? Because it is a holiday quarter, ending January 2, the number of extra shipping days is lower, but we get a full week of extra expense.

Bottom of cycle for Xilinx? We expect some recovery in communications in China, but not a big bump. We are already off the bottom in wireless and improving slowly. Our 11% new product growth rate indicates newer wired, automotive and industrial designs are ramping. Aerospace and defense declined as certain programs ended. Also, inventory has been reduced in China, so more ordering is taking place.

Inventory is pretty lean now. The end-of quarter build up was specific to one customer. It is always difficult to predict wireless sales, but we believe inventory of FPGAs in China factories has been worked down.

Wireless infrastructure future? The next major surge will be 5G, we are already in prototypes, but it will be years before the deployment of 5G. We are seeing 4G deployment in India and other smaller markets. So we don't see a return to the March 2014 peak, rather a steady improvement over time. In retrospect it was an artificial peak, based on an inventory build because each customer prepared for a higher share of the market than they actually received.

Qualcomm partnership? Would start with cards having both our chips on them, then will move to higher levels of integration. We have a tight relationship with Qualcomm. We believe the datacenter market is still an emerging market. We cannot disclose the timeline with Qualcomm, but we have other partners as well. It will be a couple of years before revenue is significant.

Server opportunity size? It is an emerging market. It is early, it is likely a hundreds-of-millions market. It could be a billion dollar market. But we will be competing with Intel, and the applications vary greatly.

x86 processor approach to this datacenter acceleration market? Power is the biggest challenge to integration. The current market uses a high performance processor and a high-end FPGA, both being high power users. So putting them both on a chip would be difficult. We believe we can be in the Intel server market.

16 nm plans? The first device goes through a lengthy evaluation phase. When we are ready we can do a full rollout. We got high functionality much quicker than expected on the current device. We will roll out more devices faster than in previous generations. It will be a broad and deep family of offerings.

The Zynq product is allowing us to grow into new areas, so it is an expansion play for us.

Are you getting questioned by customers re the Intel Altera merger? The newer cloud servers no longer have needs that match general purpose processors. Making the CPU faster does not give a significant return, and multiple cores also have limiting factors. Power requirements are the biggest limitation. FPGAs give better power results, depending on the application. Customers are excited about the move to FPGAs. Intel wants to control the FPGA rollout. But customers have an independent desire to control their destinies, and may look at non-x86 Intel based solutions.

We expect a 30% increase in automotive revenue this fiscal year, led by advanced driver assistance capabilities. But we have slots in a variety of applications including mobile looking cameras and radar. We are very confident about our ability to compete with Mobileye (MBLY).

OpenIcon Analyst Conference Summaries Main Page



More Analyst Conference Pages:



Disclaimer: My analyst call summaries may include both our condensations of statements made by company representatives and my own analysis. They are not covered by any warranty. I cannot guarantee anything said by company representatives is true. I try not to make errors, but it is possible. Before making or terminating an investment you should always verify any factual basis of your decision.

Copyright 2015 William P. Meyers