Nvidia stocks plummeted on Friday after the company released its earnings report. The company has, as of this writing, lost about 16 percent of its previous valuation. But as with AMD, what happens to a company after it reports earnings doesn’t always make sense from a technical perspective.
Nvidia’s stock has taken a hammering today for basically having a weak Q3. Revenue in Nvidia’s fiscal year Q3 2019 (Nvidia’s calendar runs a year ahead of the actual physical date) was up significantly compared with the same time last year, but flat in Q3 compared with Q2. Nvidia’s margin slipped a tiny bit, to 60.4 percent from 63.3 percent, but it’s the company’s inventory build that has analysts worried. Nvidia is currently holding $1.417B in product, up from $796M in January 2018 and $1.09B in Q2 2018. Nvidia blames this problem entirely on the decline of the cryptography market and has stated it will ship no new midrange cards to market through Q4 to give the channel time to work through the excess inventory build-up.
Wall Street clearly didn’t find these assurances very reassuring, given the damage they’ve inflicted to the company’s stock. At the same time, we have to note that Nvidia is fundamentally in excellent shape. We’ve made our opinions on the RTX family clear, but that doesn’t change the fact that Nvidia has the high end all to itself. It’s anchored extremely comfortably in that space at the moment, and up until yesterday, AMD didn’t have a midrange competitor that was cleanly sweeping its Nvidia counterpart.
While I don’t believe it played any part on the company’s share price slide, I can’t help but note that Jen-Hsun Huang fundamentally mischaracterized Turing in his remarks and QA. According to Jen-Hsun:
the Turing launch happened towards the end of the quarter, and it’s the biggest generational leap we’ve ever had… At every single price point it serves, it is substantially higher performance than the last.
This is an extremely stretched reading of the actual situation on the ground. Yes, the RTX 2070 is faster than the GTX 1080, the RTX 2080 is faster than the GTX 1080 Ti, and the RTX 2080 Ti is actually substantially faster, with a $500 price “upgrade” stacked on top. Since Nvidia decided to jack up the prices on these cards, none of them do a particularly good job justifying their pricing or price/performance ratios. That’s the entire problem. The marketing for these cards is writing checks that their actual performance can’t cash. And, as this graph from THG shows, the gaps between each same-priced GPU are not “substantial.” They’re anemic.
Investors, however, almost exclusively focused their questions on the channel inventory issue, which Nvidia said will take 1-2 quarters to completely clear. This has brought down their overall revenue estimates to $2.7B for Q4, which is significantly lower than what investors expected. Given that we heard rumors that Nvidia had to eat a substantial amount of inventory it’s not surprising to see this happen, and of course, there will be people who theorize that the problem is Turing, not Pascal. Nvidia’s official line is that Turing is selling very well and performing excellently, with no mention of the RTX 2080 Ti failures by themselves or analysts. 7nm ramps were not discussed, though Jen-Hsun specifically stated he wanted to have midrange cards in-market as quickly as possible. We have no advance information on Nvidia’s 7nm plans, but with Navi coming in 2019 and NV unlikely to want AMD to hold a process node advantage for any length of time, a 7nm Turing refresh seems more likely than not.
Now Read: Nvidia Confirms Some RTX 2080 Ti GPUs Are Defective, Promises Remedy, The First Battlefield V DXR Ray Tracing Benchmarks Have Dropped, and AMD Will Answer Nvidia’s Ray Tracing Technology, Eventually