Nvidia (NASDAQ:NVDA) was downgraded one more time by another small firm, a move that once again caused its stock to correct by about 3%. BMO Capital Markets said it expects graphics cards dropping 16% Q/Q in calendar Q1, a steeper decline than the seasonal 6% average. The culprit seems to be channel inventory, which is bloated.
As a reminder, Pacific Crest suggests a 30% to 40% (q/q) reduction in Chinese sell-through, and a 20% q/q reduction in US sell-through. One can argue, however, that NVDA is not just about gaming and graphic cards, and it still has the auto and datacenter segments that will provide for ample growth.
While I do not discount the importance of the datacenter and auto segments, the truth is that revenue from these two segments are not enough to compensate for a slowdown in graphic card sales.
Table made by me with data from NVDA
As you can see from the above table, gaming represents almost 60% of total revenue. Not only that, but revenue increase from gaming, on a percentage basis, was about the same as datacenter revenue increase. Which means that the main driver of growth for FY'17 was gaming, and not all the other segments that occupy most of the headline news these days.
Fellow Seeking Alpha writer Alex Cho wrote several days ago:
Hence, the inventory burn from Q4'16 likely has negative pass-through impact on both Q1'17 and Q2'17 results. We believe that Nvidia's results will be partially supported by higher ASPs, but unit contribution drops by approximately 24% y/y, and 20% y/y for Q1'17 and Q2'17, respectively.
Therefore, we're now revising our consumer graphics/OEM revenue lower to reflect McConnell's channel supply chain checks. Hence, our estimates on graphics revenue is revised lower from $1.268 billion and $1.002 billion to $883.13 million and $713.82 million for Q1'17 and Q2'17.
Our full-year CY'17 estimate for both consumers graphics, and OEM is revised from $5 billion to $3.956 billion, which implies NVDA's graphics revenue in consumer/OEM will decline by 10.4% y/y versus $3.956 billion (prior year).
Now if Alex's calculations are confirmed, then this changes everything for NVDA going forward. It basically means that the only source of meaningful revenue growth for NVDA in CY'17 is the datacenter segment.
Yes, the auto segment also will grow, but not enough to make an impact on NVDA's overall revenue mix. I do not expect the auto segment to provide the revenue volume needed to make a very big difference to Nvidia's revenue mix until 2020.
While we will have to wait for more color from NVDA when it reports its Q1'Y18 results on March 9th, we also have to ask ourselves why NVDA might have such a shortfall this year.
One answer for the shortfall might be intense competition from Advanced Micro Devices (NASDAQ:AMD). I am not going to get into a graphic card comparison; however, it is the only logical explanation. Yes, both companies have introduced many new cards over the past several quarters; however, AMD has put on the heat with a product mix that is priced lower.
If so, might this growth shortfall for NVDA's gaming segment continue through next year? Honestly, I don't know, and it's anyone's guess, but it is a possibility.
In any case, I put some numbers together of my own.
In the table above, I take NVDA's last year revenue per segment numbers and modeled 0% growth for gaming, 20% growth for professional visualization, 100% growth for datacenter, 20% for auto and 0% for OEM and IP.
To my surprise, the figure that came out is the exact same figure that analysts think NVDA might do for FY'18 also. However there is one difference. Have analysts modeled lower revenue for gaming, as the channel inventory investigation from BMO Capital Markets suggests? Also, have analysts actually modeled 0% growth for gaming for FY'18? I doubt it.
What all this means is that NVDA will probably not grow by 16% for FY'18 that analysts are forecasting, but less. It also probably means that FY'19 12% growth forecasts also will be lower.
When you add all this up, it probably means that price targets for NVDA going forward will come down by a lot, if the gaming revenue shortfall due to a bloated inventory is confirmed on the company's next conference call. And remember, I am being very generous expecting the datacenter segment to grow by 100%.
NVDA currently has a trailing P/E of 38 and a 12-month forward P/E of 28. You have to ask yourselves if any stock that will grow by 16% is worth such a valuation. For me it isn't.
However, please also note the above P/E metrics do not take into account that the gaming segment might exhibit zero growth for CY 2017. Because if BMO's channel investigation is correct, growth will be a lot less than 16%. And if I am correct, analyst price targets will come down like a rock.
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.