Analyst Conference Summary

conference date: October 19, 2016 @ 2:00 PM Pacific Time
for quarter ending: October 1, 2016 (second fiscal quarter 2017, Q2)

Forward-looking statements

Overview: Strong y/y growth.

Basic data (GAAP):

Revenue was $579 million, up 1% sequentially from $575.0 million and up 10% from $528 million in the year-earlier quarter.

Net income was $164 million, up 1% sequentially from $163.0 million, and up 29% from $127 million year-earlier.

Diluted EPS (earnings per share) were $0.61, flat sequentially from $0.61, and up 27% from $0.48 year-earlier.

Guidance:

December quarter (fiscal Q3) revenue is expected about flat sequentially. 69% gross margin. Op ex $245 million. Other expense $2 million. Tax rate 14%. Increase in op ex is mainly due to new tape outs.

Conference Highlights:

Moshe Gavrielov, CEO said "We remain confident in our long term growth strategy and ability to generate strong cash flows, which is the basis for our commitment to shareholder return.  As a result, we are initiating a more deliberate repurchase program to complement our long-standing dividend program." 5 of 8 end markets grew in the quarter. Zynq product sales increased 25% sequentially.

The May 2016, $1 billion repurchase authorization will be utilized in "a more deliberate manner over the next several quarters." Also aims to grow the dividend continuously.

Xilinx is accelerating operating expense to extend its leadership position.

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

AMD, ARM, Huawei, IBM, Mellanox, Qualcomm, and Xilinx, joined forces to bring a high-performance open acceleration framework to data centers. The companies are collaborating on the specification for the new Cache Coherent Interconnect for Accelerators (CCIX).

16 nm UltraScale+ has shipped nine unique products to nearly 200 customers. Spartan, Artix and Zynq families are being expanded to "a wide range of applications including Embedded Vision." Baidu is using Xilinx FPGAs for machine learning in data centers.

Revenue by product type:

46% Advanced products: UltraScale, Virtex-7, Kintex™-7, Artix™-7, UltraScale+ (these are at 28 nm, 20 nm, and 16 nm)

54% Core products. So all the older, standard products.

69.6% gross margin up sequentially from 69.2% driven by product mix.

Cash, equivalents and long-term investment balance was $3.7 billion. $1 billion long-term debt and $0.6 billion was current debt. Operating cash flow was $184 million. Depreciation $11 million. Capital expenditures $11 million. $100 million of stock was repurchased. Stock based compensation expense was $30 million. The dividend payment required $84 million.

Revenue by geography: North America 33%; Asia 38%; Europe 20%; Japan 9%.

Cost of revenues (GAAP) was $175.9 million, leaving gross profits of $403.3 million. Operating expense total was $226.5 million, consisting of: research and development $141.8 million; selling, general and administrative $83.5 million; and amortization $1.2 million. Leaving operating income of $176.8 million. Interest and other expense was $1.2 million, and the income tax provision was $11.5 million.

$0.33 dividend payable to stockholder of record on November 8, 2016 will be paid on November 23.

Q&A:

On target for 6% growth guidance for fiscal year? Right now we see fiscal Q4 with communications firming and growth in consumer, automotive.

Baidu, are you sitting next to Intel processors? This is a couple of year process, potentially $200 to $300 million per year by 2020. This is not part of the Q4 upswing.

Share buy back "more deliberate" language? Past was opportunistic. Intent is to buy over next several quarters.

16 nm node vs. Intel 14 nm timing? We taped out in June 2015, shipping product, positive feedback from customers about functionality. New products shipping faster than planned, more on way. We are clearly a year ahead of Intel, more if you count functionality.

Automotive growth drivers? Historically was entertainment side. Zynq is driver assistance. Now advanced driver assistance is growing, with 85 models to be delivered. Early 16 nm revenue was driven by Zynq. We are working on 7 nm to move towards autonomous driving in 2020 and beyond. There are a host of new players in the market. It will happen over 10 years with limited deployments at first.

Decline in legacy products, is it a normal rate? The decline in older products is because we did not do well at 40 nm at the high end. So that is declining at a fast rate. But the last 3 nodes we have executed well so in the future we should have a tail wind. We have dominant market share at 28, 20, and 16 nm and have a 6% annual growth rate target.

Share count down of flat? Should go down. Keep in mind impact of convertibles on diluted count.

Zynq revenue in dollars? 6% to 8% of overall revenue.

Industrial segment visibility, long term growth? It is a mixed category with a lot of customers. We have good visibility with large customers. Test and measurement, we know we are doing extremely well. Growth prospects are defined by the markets our customers serve. Joint Strike Fighter has been a good program for us.

5G timing? Market has changed quite a bit. The standard is not stable yet. All major players want to be first to market for 5G. We believe our 16 nm portfolio is being used in nearly all prototypes. Mass deployment won't start until 2020, but some 4.x deployments before that.

Wireless/Wireline confidence Q4 and Q1? In wired we have seen the design wins beginning to ramp. There is cap ex uncertainty. Wireless demand in China and India is being reported by our customers.

Onshore cash? Of $3.7 billion, about 2/3 is permanently reinvested offshore. Half that has a tax provision, plus the 1/3 already in the U.S. So about 2/3 is available in the U.S. without book tax implications.

Communications revenue growth before the 5G rollout? 4G has peaked. But deployments are still taking place around the world, including India. Plus there is densification. Expect wireless to continue to be choppy, but wired will grow and is less volatile.

Data center workloads v. GPUs? It is a fast-emerging market, with a lot of applications. GPUs have the edge for floating point and parallel work, including vision. FPGA dynamic configuration is an advantage for some applications. Our issue is abstraction to enable non-FPGA programmers to use our products.

OpenIcon Analyst Conference Summaries Main Page

Search

More Analyst Conference Pages: