Nvidia’s mistakes catch up to the bottom line

Huge self-inflicted financial hole for Q2

Nvidia world iconNVIDIA HAS A curious take on the causes of their latest financial meltdown, one that doesn’t seem to mirror what is happening in the rest of the industry. The almost 20% drop in expected revenue announced at the last minute seems to be a largely self-inflicted wound.

If you read the release here, Nvidia states what we all knew was coming, that their Q2 revenue would crater.  Nvidia says it will be down from an expected $950-970 million to $800-820 million. That is a nearly 20% drop, and it comes less than a week before the end of the quarter. It is astounding that that Nvidia financial folk didn’t realize this before, even more so if you know their channel problems. It is almost like the company was trying to mislead investors, but such an honest company would never do that.

Officially, the tanking in the consumer GPU business happened because of economic weakness in China and Europe. In the GPU world, that means consumers buy cheaper machines without discrete graphics. Since Nvidia triumphantly abandoned the chipset market, they no longer have any parts to sell to these customers. This is not money that will ever come back to them until they can convince the market to up ASPs by significant margins. When was the last time that happened in tech again?

More importantly, and this will be a continuing theme of the article, AMD and Intel saw no such weaknesses. In fact, DAAMIT’s ATI division saw continued product shortages only on their highest end cards, not a sign of weakness at all. When one company in an industry sees problems, and the rest do not, is that a problem caused by prevailing economic conditions?

The other official problem was increased memory costs, something that hits low end and low margin products especially hard. Nvidia has warehouses full of obsolete low end parts, so this one seems a little more plausible. At first.

If you look a little deeper, you see two problems. First, ATI showed no such weakness, in fact, it is quite the opposite. Their HD5000 line has had an unusual price increase, and it has been sustained in all but one specific product SKU for three quarters. In the graphics card industry, this is unheard of.

To make matters more problematic, Nvidia’s entire line from $198 MSRP down, is obsolete. Last September, the first DirectX 11 graphics card was released, and by early 2010, ATI had a top to bottom DX11 lineup. Nvidia’s cheapest DX11 card costs $199, and prices go up from there, a vanishingly small market segment in terms of units. This means they can only compete for 10% or less of the market with non-obsolete parts.

If you look at inventory, that obsolete product pile starts to show up in painful ways. At the end of Q1, ATI was selling everything they could make, only shortages of wafer starts by TSMC capped sales. If ATI could have gotten more wafers, they would have sold more cards. Nvidia on the other hand had inventory build up.

Nvidia blamed this on packaging times and related issues, but a quick check shows that the value inventory spike is likely more than the value of the entire run of parts that spike was blamed on. Curious. Given the vagaries of the Nvidia Q1 conference call where that blame was laid, it is unlikely that there will ever be clarity on this point.

A much more likely explanation is that the obsolete products mentioned above, all but the top two products from Nvidia at the end of Q1, simply stopped selling. To fix this lack of sales, Nvidia looks to have tried the same trick that failed so miserably the last time they faced such a problem, stuff the channel with lower-end G200 parts.

If you recall, before the launch of the GTX470 and GTX480, Nvidia forced retailers to buy several junk cards for every non-obsolete card they wanted. This is called channel stuffing, and from what we gather, it lead to an almost total stoppage in orders later in Q2. Nvidia knew this would happen, but they did it anyway to forestall Dear Leader from having to face the analysts with bad news during the company’s Q1 call. Once again, the hope was that some miracle would save them.

The last time Nvidia tried this was during the 65nm to 55nm transition. The company promised that all sales by a certain point would be 55nm, and they achieved that with fire sales and dumping. The same backlash hit Nvidia a few quarters later, orders stopped, and their results tanked. In an eerily familiar fashion, their competitors did not suffer the same fate that time either. I guess this is Santa Clara financial slang for ‘economic weakness in Europe and China’.

Once again, the same failed trick is having the same miserable results. The channel is stuffed, distributors and retailers are not buying more, impending new products only heightened the worry for retailers, and sales tank. If it was anything else, like prevailing economic conditions, ATI and Intel would be having the same problems. Surprisingly, they simply are not.

To make matters worse, sources in both Santa Clara and the far east tell SemiAccurate that Nvidia is canceling wafer starts on 40nm. ATI can’t get enough, the industry is on massive allocation, and Nvidia is giving up some of those precious wafer starts. Ironically, these excess wafers will probably go to ATI, a company who’s products do not seem to be affected by those darn prevailing economic conditions or RAM fluctuations. Imagine that!

In the end, the Nvidia earnings warning has two causes, an utter lack of competitive products, and ham-handed channel stuffing. If there are any macro economic conditions that would account for their almost 20% drop in revenue, it would have been reflected in AMD or Intel’s earnings. Both companies reported weeks ago, and both had unusually strong sales. So much for prevailing conditions or seasonality.

Nvidia will likely try to spin analysts towards a future facing direction during their Q2 conference call in two weeks. The problem there is that the company lacks anything to make that future bright. Apple just dumped Nvidia like we said a year ago, the chipset revenue is gone for good like we said even further back, and the main volume product lines they are selling are obsolete.

With the GF106 and GF108 launches in about a month, some of this will be rectified, but the whole GF100/104/106/108 line is unlikely to ever be financially viable. Nvidia has no laptop parts coming this year, and is unlikely to have any competitive mobile parts until late 2011 when TSMC gets 28nm parts out in volume. To top it off, ATI will have a refreshed HD6000/Southern Islands line out before the end of 2010, upping the competitive bar.

Nvidia has no real future in consumer GPUs, products that account for approximately 60% of their income. By the time the company has a competitive architecture again, huge swathes of the market will be obsolete due to CPUs with integrated graphics. While there may be blips from here on out, Nvidia’s future is not looking viable, much less bright.S|A

The following two tabs change content below.

Charlie Demerjian

Roving engine of chaos and snide remarks at SemiAccurate
Charlie Demerjian is the founder of Stone Arch Networking Services and SemiAccurate.com. SemiAccurate.com is a technology news site; addressing hardware design, software selection, customization, securing and maintenance, with over one million views per month. He is a technologist and analyst specializing in semiconductors, system and network architecture. As head writer of SemiAccurate.com, he regularly advises writers, analysts, and industry executives on technical matters and long lead industry trends. Charlie is also available through Guidepoint and Mosaic. FullyAccurate