AMD Technology Falls Short… Really? – Seeking Alpha

Posted: August 25, 2017 at 3:58 am

A recent piece on Advanced Micro Devices (NASDAQ:AMD) presented a case that as the company's technology lags behind that of Intel Corporation (NASDAQ:INTC) and Nvidia Corporation (NASDAQ:NVDA), and that it wont be able to compete effectively once the mining spike subsidies. Well, we beg to differ. Lets dive into the details.

The bear thesis was based on the premise that AMD finished in the second place on both the Ryzen and Vega fronts.

First of all, "being second best isnt the problem at all. AMDs success is driven from closing the gap between CPU technologies - it isn't based on beating Intel's technology. To that end, the company deployed 14nm process in order to close the technology gap.

In the past, AMDs technology problem stemmed from being on a lagging process node. That problem is now solved amid GlobalFoundries push into advanced processing technologies powered by the likes of Samsung (OTC:SSNLF) and IBM Corp. (NYSE:IBM).

There is no point to inking pages in discussing the performance of Ryzen, Epyc and Threadripper. There is a lot of material around the web covering the performance aspect. The point is that AMD successfully closed the performance gap between its technology and the technology of Intel. This will lead to market share gains going forward.

The author goes on to discuss Ryzen by comparing it to the failure of Bulldozer. Intel sure does take the crown with single-threaded performance. However, discarding Ryzen just because Bulldozer failed in multi-threading is a faulty argument.

Bulldozer was based on a 32nm process, which went to compete against Intels 22nm Ivy Bridge. Even if Bulldozer had been efficiently designed, node lag would've put AMD at a significant disadvantage. Thats the key difference between Bulldozer and Ryzen.

AMD is at parity with Intels process technology, not to mention Jim Keller and the infinity fabric. Note that Jim Keller was also the lead architect for the original Athlon 64. He was also involved in designing Apple's (NASDAQ:AAPL) A4 and A5 processors. Both Jim Keller and infinity fabric were absent in Bulldozer. That was the reason Bulldozer didnt scale well on multiple cores.

To review, Ryzen isnt anything like Bulldozer. It uses infinity fabric to utilize more cores efficiently. More importantly, AMD's process technology is at par with Intels offerings. Intel had the edge with its process lead. That lead is now lost. Even the new Coffee Lake processors are based on 14nm process.

All in all, Ryzen cant be compared with Bulldozer at all. The competitive environment is altogether different with factors like level process technology and infinity fabric at play. The author has put forward the argument that multi-threading is AMD's excuse. This doesn't seem to be true. AnandTech has the following to say about multi-threading:

"AMD pulls ahead when anything needs serious threads by a large amount, and most of the time the memory arrangement is not as much of an Achilles heel as might be portrayed. If a user has a workload that scales, AMD is bringing the cores to help it scale as wide as possible.

Then, there's price

"Intels i9-7900X is priced at $999, the exact same level that AMD has pitched its flagship 1950X the latter of which proved almost 50% quicker in the benchmark tests", notes Techradar.

Another argument put forward by the author was that multi-threading might not be attractive in the consumer market. Well, "more cores at a cheaper price" is a potential selling point for AMD. Apart from enthusiast and gamers, the average consumer will treat this as positive news.

Moreover, AMD isn't primarily focused on the consumer market. Its devoted to take market share from Intel in the enterprise space. Lisa Su had frequently touted the growth of AMD's enterprise market. We can see that AMD is being successful in that arena.

China is becoming the playground for AMDs datacenter market. Tencent (OTCPK:TCEHY) and JD.com (NASDAQ:JD) followed suit as Baidu (NASDAQ:BIDU) announced deployment of Epyc in its data centers. Baidu attracts the most traffic from China. JD is 7th in traffic rankings. And everyone knows Tencent as the company that controls most of the Chinese mobile games market. AppAnnie puts Tencent as the No.1 company in terms of its revenue-generating ability.

With all due respect, multi-threading isnt an excuse. Its a real threat to Intels dominance.

According to the AMD bear:

"Vega 64 is such a complete flop, it seriously jeopardizes the company."

Well, we have several reservations about that. First, Vega cant be a flop even assuming sub-par performance. The cryptocurrency craze is actually driving the demand for AMDs GPU products.

"While management wasn't specific on how much, the GPU revenue upside was driven by cryptocurrency applications, noted Stifel's analyst Kevin Cassidy while commenting on the companys second-quarter earnings release.

However, bears are fast to point out that Vega isnt the best mining solution, according to the recent data. Vega 64 mines up to 37 MH/s, as compared to 29MH/s mined by GTX 1070. Despite this lead, power requirements make Vega unfavorable for mining. However, this can change with firmware and mining software updates. Even if Vega stays sub-par in mining, gamers will buy Vega given the unavailability of other cards.

Moreover, Vega 64 matches GTX 1080 in performance. Vega 56 has an ~8% performance advantage over GTX 1070. Price-wise, AMD is a better option. Vega 56 is a $100 cheaper than GTX 1070, and it performs better than the latter.

Power consumption is the bear's key argument. Gamers usually dont care about power draw. Back in the day, price and performance were the main factors I considered while buying the GeForce 8800 Ultra.

When it comes to gaming cards, the pecking order is price and performance, followed by power.

AnandTech puts it like this:

"Having priced the RX Vega cards aggressively enough that at least on a price/performance basis, theyre competitive with NVIDIA if not enjoying a small lead."

How is Vega a threat for AMDs business model given the fact that the company generated around 46% of its revenue from the enterprise, embedded and semi-custom segment during the second quarter of 2017? Further, the rest of the revenue isnt just from GPUs. The computing and graphics segment also includes desktop, notebook processors and chipsets. Therefore, even if Vega flops, AMD can still stay a going concern.

Again, Vega isn't a complete flop, nor does it jeopardize AMDs business.

Another argument against AMD is the burning of cash. The author notes:

"Worse still, AMD lost money last quarter, despite massive demand from currency miners.

Strictly speaking, AMD reported negative cash flow. It didnt lose money in an economic sense. Its accounts receivable balance increased by a staggering $303 million during the first half of 2017, indicating increased demand for the companys products. Therefore, fixating on cash flow might not be fair in the case of AMD.

(Source: AMD Q2 2017 earnings release)

Increased receivable balance indicates healthy demand, which is deducted while calculating cash flows. The company adjusted $303 million in its cash flow statement during the first half of 2017. But this will reverse once debtors pay those receivables. In an economic sense, it wouldnt be right to say that AMD is losing money.

AMD has closed the technology gap, thanks to 14nm process technology and Jim Keller's designing capabilities. Lagging technology doesnt affect the bull thesis at all, as the thesis was based on narrowing the technology gap. If AMD were to beat Intel, it should be in Intels place in terms of market cap. Anyhow, second best is a moot point as far as the investment thesis is concerned.

Regarding multi-threading, AMD is a serious threat to Intel in the datacenter market. Chinas market is already showing signs of AMD's product adoption. Ryzen is completely different compared to Bulldozer. Bulldozers failure wont be reflected in Ryzen. Thanks to cryptocurrencies, AMDs GPU side has the time to make amends. The balance sheet is indicative of healthy business coming the companys way.

In effect, AMDs technology does fall short, but not short enough.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.

Additional disclosure: This publication is for informational purpose only and reflects the opinion of Focus Equitys analysts. This opinion doesnt constitute a professional investment advice. Our senior technology analyst compiled this research piece. Focus Equity is a team of analysts that strives to provide investment ideas to the U.S. equity investors.

More:

AMD Technology Falls Short... Really? - Seeking Alpha

Related Post