AMD has been artificially pumping up their stock of late so the question I have is; why? What is with the short term strategy? The only answer that I deem noteworthy is that AMD is trying to get acquired and there is a strong case emerging to support that. Also note that AMD CEO Lisa Su is currently the largest single AMD stock holder which may be an added incentive to sell the company.
During the last investor call, Lisa Su proudly said, “We achieved profitability this quarter, for the first time since 2014.” Unfortunately AMD could not have achieved that without selling a majority interest (85%) in their assembly and test operations to another firm to raise cash. As a result they recorded a significant profit on this sale ($127M after provisions for taxes). Without this one-time operations spin-out, the net operating results would have been a significant loss.
Also read: Why is AMD Stock Jumping?
A more recent stock pump is AMD releasing a questionable benchmark for their new Zen processor (coinciding with the Intel Developers Forum) that sent the stock on another run up. According to computer history, benchmarking CPUs is more of an art than a science. In this case AMD used an Open Source benchmark (Blender) which can easily be manipulated. AMD also did not disclose specifics of the configuration such as memory and they throttled (clocked down) the Intel CPU to match the Zen Part.
Bottom line: This is a classic case of benchmark smoke and mirrors. If AMD had nothing to hide they would have enlisted a third party to test the part and run a suite of benchmarks. To me this follows a pattern of deceptive behavior which should be of great concern.
AMD also hinted at yet another delay of Zen which was originally announced in 2015. AMD had hoped to get the 14nm Zen part into PCs for the 2016 back to school and holiday buying frenzy but now it looks more like it will be the 2017 back to school and holiday season. This delay will now put Zen against Intel’s planned 10nm PC launch next fall.
AMD did not mention their 10nm strategy but it would have been wise for them to use Samsung 10nm since their preferred wafer supplier, GlobalFoundries, will be skipping 10nm. The Zen part is currently being manufactured by GF using 14nm technology licensed from Samsung so it would have been a much easier transition to Samsung 10nm than say TSMC 10nm. Samsung 10nm will be in production in 1H 2017 so in theory a 10nm Zen part could have been ready for the 2017 fall shopping season if AMD was thinking ahead and had been on schedule.
This 10nm scenario is highly unlikely however since AMD recently modified their wafer agreement with GlobalFoundries to add 7nm. The new contract expires in 2020 and AMD will take a one-time charge of $335 million in the third quarter of 2016.
The other interesting piece of news is that Intel released their 14nm+ version which offers more than a 10% process performance increase. I was also told that it offers a cost reduction as well so this is what the AMD Zen part will be competing against. At 10nm Intel will again do multiple versions that should be well ahead of, and competitive to, GF 7nm which should keep AMD at bay.
Bottom line:The epic Intel vs AMD battle is back which will benefit the semiconductor industry but I wonder if this is yet another false start for AMD, one that could ultimately kill them?
Also read: AMD Zen and the Art of Microprocessor Maintenance