Nvidia shares fell steeply after the chipmaker gave weak fourth quarter revenue guidance when it reported third-quarter earnings after the close on Wednesday.
"After many years of near flawless performance, Nvidia finally stumbled," Deutsche Bank said.
Nvidia's stock cratered 19 percent in trading, pushing the stock even deeper into correction. Nvidia had already fallen to $202.39 a share at Wednesday's close, a plummet of more than 30 percent from its 52-week high of $292.76 a share.
"We were clearly wrong on the stock," Goldman Sachs said as the firm removed Nvidia from its list of conviction stocks. However, Goldman said it remains "Buy-rated on the stock as our view that Nvidia has access to one of the best growth opportunity sets in Semis and that it has a sustainable competitive lead within remains unchanged."
"The stock will likely not bounce back right away, given the severity of the miss," Morgan Stanley said.
Here's a wrap of all the major analyst opinions.
Reiterate Buy on best-in-class core biz ... Disappointing results/outlook as excess channel inventory impact turned out to be much more severe ($700mn) that we thought ($300-$350mn). We are not expecting a quick recovery, and it could take 1-2 quarters for NVIDIA (NVDA) to regain its growth/execution credibility. However, at the after-hours price of around $165/sh the stock is now 18x our CY20 estimate versus 20x-45x historical range. Ex crypto drag of nearly $1bn this year, we are still forecasting 14% sales growth in FY20/CY19 well-ahead of most semis that will see flat to declining sales ... Overall we lower our FY20/FY21 pf estimates by 16%/11% to $7.11/$8.90, and reduce our PO to $250 from $300 on 27x cash adjusted CY20 PE (lower than 29x prior, but still within historical 20x-45x range) ... we believe our model is adequately conservative and FQ4'19 will likely mark the trough.
We remove NVDA from the Conviction List with an updated 12-month price target of $200 (19% potential upside to the after hours share price of $168.45). Since we added NVDA to the Conviction List the stock is down 15.5% vs. the S&P500 down 1.4%. We were clearly wrong on the stock as we underestimated the magnitude of the channel inventory build in mid-range Gaming GPUs (typically ~1/3 of Gaming GPU revenue) and, to a lesser extent, the correction in game console SoCs ... we remain Buy-rated on the stock as our view that Nvidia has access to one of the best growth opportunity sets in Semis and that it has a sustainable competitive lead within remains unchanged.
After many years of near flawless performance, NVDA finally stumbled as the fall-off in crypto demand and the resulting ballooning of inventory impacted its quarter and more severely impacted the guidance. While the cuts are likely to be harsh, we expect the inventory adjustment to reset Gaming segment expectations to a meaningfully lower level and call into question what the true growth rate of Gaming was/is. Overall, we believe NVDA has a wide array of growth drivers (Datacenter, Provis, Automotive), however Gaming does not appear to be as compelling an example of growth as many previously believed. Consequently, until we have greater clarity on the post-crypto trajectory of the Gaming business, we maintain our Hold rating and cut our P/T to $190.
The stock will likely not bounce back right away, given the severity of the miss post management voicing confidence throughout the quarter that the litany of cautious data points did not signal a potential problem in gaming. We do see the next two quarters as artificially depressed, given the significant clearance of Pascal inventory, but the steepness of the snapback is dependent on sell through of the new Turing chips, where in the short term we are less optimistic than management, and the deceleration of the HPC/cloud business (which, to be clear, is still growing over 50% y/y) amid slowing cloud spending will raise concerns there as well.\
Aside from excess GPU inventory in the channel, core Gaming demand remains strong on continued blockbuster gaming titles/e-sports strength, notebook sales and launch of new games supported by Turing. Though management did not provide directional guidance per segment as it has done in prior quarters, with strong momentum in high-performance computing, datacenter acceleration and early adoption of inferencing products at cloud service providers, where spending remains robust (including in China), we are confident in NVIDIA growing its Datacenter business sequentially again in the Jan-Q. In Gaming, excluding the impact of excess channel inventory, we believe the team is seeing solid demand, and will likely see solid sequential growth for its RTX-20 series gaming platforms in the Jan-Q. Overall, with NVIDIA missing expectations over the past few quarters, we believe the company's guidance for the Jan-Q is prudently conservative and we expect the team to return to outsized growth as we move through FY20. Despite the near-term softness, our view on NVDA's secular drivers in Gaming and Datacenter is unchanged and we see continued growth in FY20 even with tough comps as the impact of cryptocurrency fully unwinds. We remain OW NVDA and would be buyers on weakness.
While we took down estimates nearly a month ago on gaming risk (LINK), guidance was even worse than expected due to ongoing channel inventory reduction for mid-tier Pascal cards. It is debatable how much gaming will snap back and when, but inventory resets like this often present buying opportunities – all else equal. While this may very well prove to be the case, 2019 still looks like a gaming transition year where RTX supported games will not broadly proliferate until C2H:19. In the meantime, data center – which was a little softer than expected (though still up >50% Y/Y) – faces a mixed narrative given some potential headwinds from hyperscale capacity digestion, some competition from AMD with a new GPU offering, and a rapidly evolving competitive environment for inference applications ... Longer-term, we also worry that NVDA could have a harder time moving to more advanced process nodes given the large die sized inherent in its architecture (AMD 330mm2 die size for new GPU v NVDA >800mm2), though this probably won't matter for a while. We cut our target from $260 to $190 based on the same 25x PE on lower EPS.
Led by excess channel inventory for the mid-range Pascal gaming GPU post a crypto demand hangover, NVDA offered a revenue guide for the January Q that was well below expectations - $2.7B vs. consensus of $3.4B – thereby driving a material reset to near-term estimates. Obviously, this is a clear setback and will weigh on shares, but in no way does this alter our long-term positive view on the company ... This will likely be a key area of debate – whether 1 Q or 2 Q reset – considering mgmt guiding gaming higher Q/Q into April, we think one Q. Bigger picture, we still think this is a double-digit grower, particularly with Turing launching now in the high end and the new "game-changing" architecture targeting the mainstream market likely sometime in 2019. Moreover, a sold out high-end Pascal Gaming channel along with sold-out high-end Turing RTX 2080 TI GPUs suggests enthusiast-class Gaming demand remains strong, likely driving demand for high-end Turing GPUs into CY19. The Turing launch is gaining momentum with Battlefield 5 content now available, Data Center growth is expected to continue with the recently launched T4 inference offering adding to the growth story, and Automotive/Pro Viz actually surprised to the upside. With shares offering significant upside to this new level, we reiterate our Outperform rating.
NVDA stock was down 17% after hours after missing Oct-Q sales by 2% and guiding Jan-Q sales 21% below Street/our expectations. Slower than expected inventory in the channel and company not shipping mid-range Pascal desktop GPUs are driving a sharp 30% Q/Q gaming correction vs our/Street flat/+5% expectations. Nvidia expects Apr-Q to be up Q/Q in gaming sales as underlying gaming market growth remains intact with no China impact. Nvidia is very bullish on next year growth based on market expansion and new product introduction. Company raised its buyback significantly to $7B with additional $3B to be returned over the next five quarters. Net-net, while crypto impact is disappointing, underlying growth in gaming, data center, and auto markets remains on track. As such, we remain buy-rated on the stock but are lowering our TP again to $244 from $270. We see CES in Jan as next catalyst for the stock.
OctQ missed, but the JanQ Rev/EPS outlook is 21%/36% below consensus due to high crypto GPUs in the channel. We forecast JanQ gaming GPU revs decline by 35% QQ as NVDA clears channel inventories. We lower our ests and price target, but model gaming revs resume QQ growth in the JulQ. The JanQ reset is a setback, but we continue to view NVDA as a top play on secular themes in AI, Gaming and A/V. We recommend buying the confession.
For the first time since upgrading NVDA to Buy in early 2014 ("Improving Execution With Beat And Raises Warrants Upgrade To Buy…"), we are moving to the sidelines, downgrading shares from Buy to Neutral. The catalyst is a meaningful F3Q19 results miss and much more substantial F4Q19 guide lower versus our cautious preview, and an adverse impact to FY20's Y/Y overall sales growth profile in our updated model ... Our FY19-21 EPS fall -12.4%, -15.6% and -14.6% to $7.26, $7.55 and $9.61, respectively, and on lower estimates and a lower target multiple our PT decreases from $240 to $190.
NVDA reported F3Q(Oct) Sales/EPS of $3.18B/$1.84, below consensus $3.24B/$1.93. More importantly, F4Q(Jan) outlook of $2.70B significantly disappointed vs. Street's $3.40B. Excess legacy mainstream Pascal (GPU) inventory led downside. Mgmt expects it will take two quarters to normalize channel inventory. We believe investors will be surprised by the magnitude and duration of gaming GPU excess and expect NVDA shares to trade lower near term. We see the mining-related miss as transitory. The response to NVDA's newly launched Turing, boasting DLSS, ray-tracing and ~2x performance vs. Pascal, has been bullish. High-end 2080Ti remains widely stocked out. DC revenues continue to grow 50%-plus Y/Y and NVDA's long-term secular growth thesis appears unchanged. Near term, we're reducing estimates and price target but maintaining our Outperform. Our target goes to $250 from $310.
On Thursday evening, NVIDIA reported significantly weaker than expected C3Q18 results and 4Q18 guidance as the "crypto-hangover" (hey Jen-Hsun, we coined the phase first) appears significantly larger than our sizable expectation. Indeed, guidance was -$700 million below Street expectations, as NVIDIA chooses not to ship mid-range product into the bloated channel, allowing it to clear ... While data center was a tad light, it was better than feared, given poor DC commentary from others this reporting period, and T4 commentary was very positive (we think this is NVIDIA's next big DC hit!). While we have been the crypto-GPU bears on the Street, admittedly we were not patient enough to let this unwind fully play out, and may have also underestimated the size of this Ethereum GPU bubble. That said, we are long-term bulls on NVDA as we believe in the A.I. inference opportunity, pro-viz upgrade cycle, and 7nm refresh coming this fall. While NVDA's report was unexpectedly bad, there is one thing we know… AMD will likely be worse (Street guiding AMD 1Q better than seasonal? And we believe crypto-GPU as a percentage of revenue is 2x NVIDIA's… yikes). Reducing our estimates and NVDA PT from $230 to $210.
OUR CONCLUSION – FRUSTRATING: Concerns (and now frustration) over a significant gaming channel inventory burn-off have materialized ... While we can appreciate that NVIDIA's weak F4Q19 outlook is impacted by a 1-2 quarter work-down of Pascal mid-range gaming card inventory in the channel (~$600M; assuming no sell-in in F4Q19 as crypto-related dynamics flush through the channel), couple with a seasonal decline in game console builds, we think investors will be frustrated by NVIDIA's comments exiting F2Q19 that: "…we [NVIDIA] see inventory at the lower-ends of our stack…inventory is well positioned for back-to-school and building season that's coming up on F3Q19…"
Bottom Line: Well, even if we model a strong double-digit growth in DCG next year, we think there is a high likelihood that NVIDIA will not grow next year. We are modeling for as such. The large shortfall in guidance due to a bloated channel due to crypto-currency is in sharp contrast to the comments around channel inventory from the company at the last earnings call. Our estimates and target price are going lower. We remain Market Perform rated.
We are Buyers on weakness. We had noted higher NVDA channel inventory pickup in SepQ and the Turing RT delay in mid-September (Link) early on, though we assumed investors' focus would be on Data Center. NVDA reported inline OctQ GAAP rev/EPS of $3.2B/$1.97, DC revenue up 58% y/y, Gaming up 13% y/y, and Automotive strong up 19% y/y. The JanQ guide was much weaker at $2.7B ($3.4B consensus) with the GPU channel inventory (as we had noted post our Asia trip) and with NVDA, we believe, kitchen sinking the quarter. Maintaining Buy, and adjusting estimates; PT to $245 (from $295) as Data Center, Automotive should drive growth.
We believe that NVIDIA could be facing challenges beyond just mid-range Pascal GPU inventory ... Guidance was worse than feared due to a sizeable buildup of gaming GPU inventory in the channel. NVIDIA guided Jan-19 revenue to decline -15% sequentially or $2.7 billion at the midpoint, driven primarily by ceased shipments of mid-range Pascal GPU's. This was below our/Street estimates of $3.4 billion. F4Q gross margin guidance of 62.5% was slightly below both consensus (63.1%), and our estimate (62.8%).
The surprisingly weak Q4 guide appears temporary. NVDA guided Q4 ~20% below consensus revs as the company halts ~1/3 of gaming segment sales to flush channel inventory built during the crypto enthusiasm in 1H18. This badly damages near-term revenue and profits, but Datacenter, Pro-Viz, and Automotive results support our structural growth view. Gaming resets our 2019 & 2020 EPS to $7.33 & $8.70 (from $8.18 & $9.59). PT goes to $237 (from $316) based on 30x (17x discount to rapid-growth tech peers) our CY20 EPS, discounted back 1 year. Buy.
Going forward, we think the focus will now shift to Data Center as gaming expectations are now reset due to crypto currencies and a product transition (Turing) which will unlikely ramp until around the Jul-qtr time frame. Net Net: we lower our price target due to the lower than expected results (PT to $260 from $310). Positively, gaming and Pro Visualization will likely be up q/q in January helping gross margins and offsetting the material gaming weakness.
Crypto related inventories are playing a significantly more serious problem for Nvidia, and in the near-term may mask strong gaming and data center momentum. Management is taking more dramatic action in the January quarter by not shipping mid-tier desktop GPUs until the channel normalizes (a $600 million hit), after last quarter's guide that included the notion of being masters of the channel. The Street will exact a penalty for the situation of course, however, our view, particularly after SC18 earlier this week, that Nvidia is set for many years of superior growth driven by data center, HPC, gaming, and autonomous driving vectors. We reiterate our Buy rating for NVDA, while taking our price target to $260, from $315.
via IFTTT
No comments:
Post a Comment