Nvidia has released a preview of its Q1 2021 revenue figures and a statement indicating the GPU shortage that’s bedeviled the market since September 2020 isn’t going away any time soon.
The good news is, Nvidia expects to earn more than the $5.3 billion it previously predicted. Nvidia doesn’t mention it in their own note, but FY Q1 2022 revenue of $5.3 billion would dwarf the $3.08 billion the company reported a year ago. Pandemic-related demand, gaming replacement cycles, and cryptocurrency mining have boosted Nvidia’s revenue ~1.7x year-on-year.
And the bad news?
“Overall demand remains very strong and continues to exceed supply while our channel inventories remain quite lean. We expect demand to continue to exceed supply for much of this year,” said Colette Kress, Nvidia’s chief financial officer. “We believe we will have sufficient supply to support sequential growth beyond Q1.”
Kress’s remarks imply that GPU availability is unlikely to improve this year. While she does state Nvidia believes it can “support sequential growth,” it’s not clear if that means sequential growth in channel shipments, specifically, or sequential growth in quarterly earnings.
Nvidia has also stated it expects to report $150M in revenue from CMP cards, up from the $50M it initially forecasted. CMP cards are intended for mining and the cards currently available are based on the older Turing microarchitecture. The hope is that these older 12nm GPUs can relieve some of the pressure on 7nm cards and help gamers acquire them at something approaching reasonable prices.
Nvidia has made some interesting claims about Ampere’s sales performance relative to Pascal and Turing. The graph on the far left does not track the Steam Hardware Survey’s estimates for market share between the relevant GPUs five months after launch. Nvidia did not have a $500 MSRP launch card for Pascal, but the GTX 1070, 1080, and 1070 Ti ($339, $599, and $500 at launch respectively) all held a larger share of the market than the RTX 2070, five months post-launch. I’m not claiming Nvidia’s data point is wrong, but the Steam Hardware Survey’s historical data suggests something different.
The second graph also seems incorrect. Six months post-launch, Turing had 1.68 percent of the market while Ampere held 2.85 percent according to the Steam Hardware Survey. That’s a 1.7x gain, not a 2x gain, and it reflects the fact that Turing was abysmally priced and positioned. We didn’t call our review “You Can’t Polish a Turing” to highlight the excellent price/performance ratio of that GPU family. Also, Pascal adoption outpaced Ampere adoption during the same six-month period, with GTX 10xx GPUs accounting for 3.08 percent of the market, despite the fact that the 10xx family also suffered from its own availability problems.
Ampere is an excellent GPU family and the RTX 3000 series GPUs are great buys, at MSRP. But it sure looks like Nvidia is trying to make the adoption and availability situation look better than it is, at least as far as the retail channel is concerned. Outpacing Turing during the first six months of that GPU’s life is not an accomplishment. Not given that Pascal adoption was still surging through most of the same period thanks to fire-sale prices and poor positioning of the RTX 2000 series.
We’d be lying if we said any of this markedly affected Nvidia’s earnings, though. The company continues to rake in cash, hand over fist, and both gaming and data center markets have boomed in the past four years. You might not be able to polish a Turing, but you can clearly make a ton of money with one. Nvidia’s financial performance these past few years illustrates how critical the company has become to emerging AI and data center applications. Shortages are bad for customers, but they’re likely to be amazing for Nvidia’s bottom line through 2021.
We don’t expect any near-term improvement in the GPU market on Nvidia’s side and AMD GPUs remain similarly difficult to locate. Any hope for price declines seems to rest on the crypto bubble bursting once more.
Now Read:
0 Commentaires