Analyst Conference Summary


Advanced Micro Devices, Inc.

conference date: July 16, 2015 @ 2:30 PM Pacific Time
for quarter ending: June 27, 2015 (second quarter, Q2)

I own AMD stock
Forward-looking statements

Overview: As expected, another lousy quarter for AMD. Time to abandon hope? Even the predicted seasonal uptick for Q3 is unlikely to get AMD back in the black.

Basic data (GAAP):

Revenue was $942 million, down 8.5% sequentially from $1.03 billion, and down 35% from $1.44 billion in the year-earlier quarter.

Net income was negative $181 million, down sequentially from negative $180 million, and down from negative $36 million year-earlier.

EPS (earnings per share) were negative $0.23, flat sequentially from negative $0.23, and down from negative $0.05 year-earlier.


Revenue for Q3 is expected to be up sequentially from Q2 by 3 to 9 percent. Cash to end at $700 million, partly due to an inventory build. No longer believe can be profitable in the second half. Looking to restructure to reduce costs, which will entail restructuring charges.

OEMs are still very cautious, especially for back-to-school season.

Conference Highlights:

The EESC (Enterprise, Embedded, Semi-Custom) segment had strong sequential revenue growth, and channel business grew, but plunge in PC processor sales more than offset. OEM APU sales were particularly dismal.

Non-GAAP results: net income negative $131 million, down sequentially from negative $73 million and down from $38 million year-earlier. EPS of negative $0.17, down sequentially from negative $0.09 and down from positive $0.05 year-earlier. Gross margin 28% down from 32% sequentially.

GAAP gross margin was 25%, down sequentially from 32%. A charge of $33 million to drop 20 nm in favor of 16 nm FinFET was particularly damaging to margins.

Computing and Graphics segment revenue of $379 million was down 29% sequentially from $532 million and down 54% from $828 million in the year-earlier quarter. But ASPs were actually up. Late in the quarter notebook sales were particularly down. Expects mobile Carrizo platforms coming to market in the second half to restore mobile sales, and the channel inventory has been rebalanced. Initial demand for the new GPUs has been good.

Enterprise, Embedded and Semi-Custom segment revenue of $563 million rose 13% sequentially from $498 million, and was down 8% from $613 million in the year-earlier quarter. The sequential increase was mainly due to custom SoCs. The y/y decrease was due mainly to lower sales of server chips. The $33 million node transition charge was attributed to this segment. Believe Q3 will be the annual peak due to inventory build for game consoles. Won a major new SoC design in the quarter, but it is confidential.

After the quarter ended AMD introduced the first GPU using HBM, die-stacked high-bandwidth memory, as well as a full family of new Radeon R7 and R9 graphics cards. New APU variants ready for Windows 10 are also being introduced.

HP and Samsung introduced new thin client designs based on G-series processors.

AMD has been demonstrating its AR-based server chips, the Opteron A1100 series.

Cash and equivalents (including marketable securities) ended at $829 million, down $77 million sequentially. Debt was $2.27 billion. Cash flow used in operations was $58 million. Capital expense was $17 million.

No debt should come due before 2019, and interest is now at a reduced rate.

GAAP cost of sales was $710 million, leaving gross profit of $232 million. Research and development expense was $235 million. Marketing, general and administrative expense $134 million. Amortization $0 million. Restructuring $0 million. Leaving an operating profit of negative $137 million. Interest and other expense was $43 million. Tax $1 million.

AMD employee headcount ended at 9,469, down sequentially from 9,583 and down from 10,300 year-earlier.


Minimum cash balance goal, ability to raise cash? Minimum target is now $0.6 to $1 billion, but we could manage the building with less.

Semi-custom pipeline? We have been active with various customers. Last year announced 2 new designs that would start to ramp in second half of 2015. New design win is significant. There are other potential wins based on combining CPU and graphics IP.

GPU refresh, Fury X and Fury out of stock? Initial ramp has been as expected. Fury just launched this week, and Nano will launch in August.

Channel inventory? In desktop and AIB we grew sequentially while draining downstream inventory. We believe we are about done with that process. We are watching China because of macroeconomic issues.

Seattle timeline? Will continue to launch in second half of 2015.

Wafer supply agreement? We have taken about $0.4 billion against the $1 billion commitment. We are working with GlobalFoundries to rework that given the demand side. Does not see any financial implications at this time.

Semi-custom low gross margins, will the new wins have the same issue? We want to be strategic around op ex and our long-term roadmap. Computing and graphics IP are critical to our plans. The markets for semi-custom range in gross margins. The goal is to return to profitability.

Intel's push-back of 10 nm road map, your node map in general? FinFET is maturing nicely and is part of our 2016 roadmap across our markets. We have just tapes out a couple of FinFET designs. But our focus is on our design architecture. Zen should be very competitive. If the node difference is shrinking that would help AMD.

We aim to continue to reduce op ex, and we will share the details when they are final.

Segments in Q3? We expect both segments up, with semi-custom leading. Graphics should be up off a low Q2.

HBM effect on margins, extent of technology lead? We are the first in GPUs and targeted the enthusiast segment. Manufacturing has been as expected, so the family should be strong in the second half of the year. We aim to grow graphics market share. It is important to lead in the enthusiast segment.

We started several designs in 20 nm and decided that given the trade-offs, FinFET would be better. The portfolio should move into FinFET next year.

New semi-custom design? It can take from 24 to 36 months to see significant revenue. It should have a $200 to $500 million lifetime revenue. This is a third, not a rehash of the 2 already announced.

Virtual reality? We are very excited about the virtual reality market, which uses both CPU plus GPU technology. Today it is mostly developer systems. It should be a growth driver for our graphics business.

$850 million inventory in Q3? It is to support second half sales, in particular chips for gaming consoles. Inventory should shrink after Q3.

On our interest payments the cash goes out in Q1 and Q3, even though on an accounting basis it is spread over all quarters. That is another reason cash will be low at the end of Q3.

OpenIcon Analyst Conference Summaries Main Page

AMD Investor Relations page

Openicon Main AMD page


More Analyst Conference Pages:


Disclaimer: Our analyst summaries may include both our condensations of statements made by company representatives and our own analysis. They are not covered by any warranty. We cannot guarantee anything said by company representatives is true. We 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