Nvidia Jumps on Revenue, Earnings Beat - Live Blog

Eric Jhonsa

Shares of Nvidia NVDA were rising close to 7% after-hours on Thursday after the chipmaker reported earnings and revenue beats for its January quarter.

Revenue of $3.11 billion beat a $2.96 billion consensus, while GAAP EPS of $1.53 beat a $1.34 consensus and non-GAAP EPS of $1.89 beat a $1.67 consensus.

For the April quarter, Nvidia expects revenue of $3 billion, plus or minus 2%, above a $2.85 billion consensus. The company noted that it cut its guidance by $100 million to account for the coronavirus' potential impact.

Eric Jhonsa, TheStreet and Real Money's tech columnist, is analyzing Nvidia’s January quarter earnings report, as well as its call with analysts, scheduled for 5:30 p.m. E.T.

Comments (69)
No. 1-50
Eric Jhonsa
Eric Jhonsa

Editor

Thanks for joining us.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia's call has ended. Shares are up 4.9% after hours to $284.00 after the company soundly beat January quarter estimates on the back of very strong Data Center (server GPU and hardware) segment demand.

Nvidia also issued April quarter sales guidance that topped consensus by $150M at its midpoint, in spite of the company cutting guidance by $100M to account for coronavirus-related uncertainty.

Eric Jhonsa
Eric Jhonsa

Editor

Huang with some closing remarks. Goes over the January quarter's Gaming and Data Center growth drivers (ray-tracing, training, inference, etc.). Says Nvidia will share more about its efforts at the GTC conference, which takes place in late March.

Eric Jhonsa
Eric Jhonsa

Editor

Question about automotive engagements.

Huang: I think the auto industry is struggling. But industry interest in redefining business models has never been greater. Autonomous driving is the key to that. The opportunity is very real, and automakers realize that.

Eric Jhonsa
Eric Jhonsa

Editor

Question about the sustainability of Nvidia's Data Center position, and how Nvidia is looking to stay ahead of rivals.

Huang: Our company has been focused on one thing: Accelerated computing. We're always looking at where system bottlenecks are for our GPUs, and trying to grow the number of applications we can accelerate. The number of AI applications that we're going to accelerate is really broad. As long as these applications continue growing, we're in good shape.

Eric Jhonsa
Eric Jhonsa

Editor

Question about inference growth. How do you see demand developing across various products?

Huang: Historically, inference was a small part of our business. But the demands of AI/deep learning models have changed that. Conversational AI models need to deliver instantaneous responses, and Nvidia has worked hard to bring down latency. You're going to see a lot more opportunities for us in inference.

Adds that at the network edge, GPU usage depends a lot on whether a solution needs to be software-defined/programmable or not. Predicts fixed-function AI chips costing just $0.50-$1.00 will become available, but adds Nvidia's GPUs will be needed for more demanding solutions requiring programmability.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia has slipped a little during the call, but is still up 5.2% after hours.

Eric Jhonsa
Eric Jhonsa

Editor

Kress: Gaming is usually seasonally down Q/Q in FQ1. This year, it will be a little more pronounced due to the coronavirus. We expect a low double-digit Q/Q decline for Gaming.

Eric Jhonsa
Eric Jhonsa

Editor

Question about FQ1 Gaming expectations, and desktop and notebook GPU expectations going forward.

Huang: End-user demand for gaming GPUs is great. Ray-tracing adoption is growing, and there's a big pipeline of ray tracing-capable games.

Also notes that notebook GPU growth and Nintendo Switch processor sales have changed the profile of Nvidia's gaming business, boosting sales but making the business more seasonal. FQ3 will be stronger than FQ4 due to seasonal builds.

Eric Jhonsa
Eric Jhonsa

Editor

Question about Nvidia's recently-announced cloud gaming partnership with Tencent.

Huang: Tencent is the world's largest game publisher. Transitioning to the cloud in China will be a long-term journey. Internet connections can vary in the region. But there's a big long-term opportunity.

Also takes a moment to talk up Nvidia's GeForce Now service. Argues its game selection and Nvidia's cloud experience are strengths. Says the initial reception has been "fantastic."

Eric Jhonsa
Eric Jhonsa

Editor

Question about how Nvidia got to $100M for its coronavirus-related guidance cut.

Kress: It's still quite early to figure out what the impact will be. The $100M cut is split pretty evenly between Gaming and Data Center. There could be an impact for both supply and demand.

Adds that China is about 30% of Nvidia's Gaming business and (though the percentage shifts from quarter to quarter) also an important contributor to Data Center.

Eric Jhonsa
Eric Jhonsa

Editor

Question about FQ1 Data Center expectations, and one about expectations for the attach rate for server GPUs in the coming years.

Kress: We expect Data Center sales to grow again this quarter. Our visibility looks good.

Huang: I believe every query will eventually be accelerated. Recommendation systems are very complex, and deep learning will be needed.

Also reiterates his view that slowing CPU performance advances will drive greater GPU usage, and that the programmability of GPUs will remain a major competitive advantage relative to fixed-function accelerators (i.e. ASICs).

Eric Jhonsa
Eric Jhonsa

Editor

Question about hyperscalers' share of Data Center sales.

Kress: Data Center sales grew for many different clients. Hyperscalers are still about 50% of sales.

Eric Jhonsa
Eric Jhonsa

Editor

Question about chip supplies in light of the supply constraints faced by TSMC (Nvidia's biggest foundry partner).

Huang: We're in pretty good shape in terms of supply. But it is true that industry supply is tight. We're relying on multiple process nodes and factories. We're working closely with customers and foundry partners. All of us have to do a better job of forecasting.

Eric Jhonsa
Eric Jhonsa

Editor

Question about cloud sales. What led growth to slow down last year, and what's leading it to rebound now? Will the business have this kind of lumpiness in the future?

Huang: Major breakthroughs in recommendation systems and conversational AI happened last year. AI models have grown tremendously in size. Microsoft just announced that it increased the size of an AI model from 7.5B parameters. Models are getting bigger, and there are more of them.

Also says that Nvidia is handling AI inference workloads for all of the hyperscalers, and that startups are now beginning to put their AI models into production with the help of public cloud instances.

Eric Jhonsa
Eric Jhonsa

Editor

First questions is about the factors driving Data Center growth.

CEO Jensen Huang uses the question to talk at length about the growth of AI-related workloads. Notes recommendation engines are a major growth driver, since they might have to analyze billions of pieces of data. Also the general growth of computing needs for inference workloads, training AI models, vertical customers and public cloud services, as well as rising interest in edge AI deployments.

Eric Jhonsa
Eric Jhonsa

Editor

The Q&A session is starting.

Eric Jhonsa
Eric Jhonsa

Editor

Kress now going over Nvidia's guidance. Reiterates that it's tough to gauge the coronavirus' exact impact right now, but Nvidia is choosing to be cautious.

Eric Jhonsa
Eric Jhonsa

Editor

Kress goes over the recent unveiling of Nvidia's Drive AGX Orin autonomous driving platform, which is much more power-efficient than its predecessor.

Eric Jhonsa
Eric Jhonsa

Editor

Kress going over some of the quarter's product announcements, including the rollout of the Magnum IO software suite (meant to address storage and I/O bottlenecks for multi-server GPU deployments).

Eric Jhonsa
Eric Jhonsa

Editor

Kress now going over server GPU engagements with enterprise clients such as energy companies, retailers, Domino's Pizza, etc.

Eric Jhonsa
Eric Jhonsa

Editor

Kress once more talks up how natural-language processing workloads such as voice assistants are fueling inference demand. Also notes content recommendation engines are an inference growth driver.

Eric Jhonsa
Eric Jhonsa

Editor

Notably, Tesla T4 GPU shipments (used for AI inference, among other things) were said to have risen 4x Y/Y. Training and inference sales both set records.

Eric Jhonsa
Eric Jhonsa

Editor

Says vertical industry growth was primarily driven by consumer Internet companies, although other verticals are also growing.

Eric Jhonsa
Eric Jhonsa

Editor

Now onto Data Center. Kress notes hyperscaler demand was fueled by both training and inference workloads, as well as GPU sales for cloud computing instances.

Eric Jhonsa
Eric Jhonsa

Editor

Kress notes the recent launch of the GeForce Now cloud gaming service (previously in beta). Highlights its ray-tracing support and (in contrast with Google Stadia) ability to let gamers use existing game licenses.

Eric Jhonsa
Eric Jhonsa

Editor

Kress: Gamers continued moving towards higher-end GPUs. 125 laptops featuring Nvidia gaming GPUs were available from retailers.

Eric Jhonsa
Eric Jhonsa

Editor

Highlights the growing use of ray-tracing in marquee games, as well as viewership growth for eSports.

Eric Jhonsa
Eric Jhonsa

Editor

Says channel demand was strong for Nvidia's entire gaming GPU stack.

Eric Jhonsa
Eric Jhonsa

Editor

Kress: We saw strong desktop and notebook gaming GPU end-demand.

Eric Jhonsa
Eric Jhonsa

Editor

CFO Colette Kress now talking.

Eric Jhonsa
Eric Jhonsa

Editor

IR chief Simona Jankowski is going over Nvidia's safe-harbor statement.

Eric Jhonsa
Eric Jhonsa

Editor

The call is starting.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia is holding onto its after-hours gains: It's currently up 6.4%. The stock is now within $5 of an all-time high of $292.76 (set in Oct. 2018).

Eric Jhonsa
Eric Jhonsa

Editor

Hi, I'm back to cover Nvidia's earnings call, which should start in a few minutes. Here's the webcast link:

deepzn2
deepzn2

thx champ

Eric Jhonsa
Eric Jhonsa

Editor

I'm taking a short break, but will be back to cover Nvidia's earnings call, which starts at 5:30 PM ET.

Shares are up 6% after hours after Nvidia comfortably beat January quarter estimates on the back of a 43% Y/Y increase in Data Center segment sales and issued above-consensus April quarter guidance in spite of cutting its outlook by $100M to account for coronavirus uncertainty.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia's quarterly revenue trajectory. Gaming sales have rebounded from a late 2018/early 2019 channel inventory crunch that followed a collapse in demand from crypto miners, but are still below peak levels.

Data Center sales have surged to new highs after dipping for a while last year amid a cloud capex slowdown.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia ended its January quarter with $10.9B in cash and $2B in debt.

For the whole of fiscal 2020, free cash flow was $4.27B (+36% Y/Y).

Eric Jhonsa
Eric Jhonsa

Editor

As previously signaled, Nvidia didn't buy back any stock last quarter. The company reiterates that buybacks will resume after the Mellanox deal closes.

Eric Jhonsa
Eric Jhonsa

Editor

Spending growth was restrained: Nvidia's operating expenses rose 12% Y/Y on a GAAP basis to $1.03B, and 7% on a non-GAAP basis (this excludes stock comp and some other expenses) to $810M.

For the April quarter, Nvidia is guiding for GAAP opex of $1.05B and non-GAAP opex of $835M.

Eric Jhonsa
Eric Jhonsa

Editor

One thing worth noting about Nvidia's Datacenter growth: It's happening at a time when Nvidia's flagship AI training/HPC GPU (the Tesla V100) is about 30 months old and reports have been swirling that the V100's successor will arrive soon.

While the V100 probably wasn't the only driver of the quarter's Datacenter growth (Nvidia has also been seeing strong demand for its Tesla T4 GPU for AI inference workloads), the growth does suggest Nvidia's competitive standing in AI training (a market it has historically dominated) remains quite strong in spite of growing competition.

Eric Jhonsa
Eric Jhonsa

Editor

Nvidia's FQ4 performance at a glance.

Eric Jhonsa
Eric Jhonsa

Editor

Inventory, which was elevated for much of last year, finished the January quarter at $979M. That compares with $1.05B three months earlier and $1.57B twelve months earlier.

Eric Jhonsa
Eric Jhonsa

Editor

Here's Nvidia's CFO commentary, for those interested.

Eric Jhonsa
Eric Jhonsa

Editor

Regarding the Mellanox deal, Nvidia says discussions with Chinese regulators are "progressing," and reiterates that it expects the deal to close in early 2020.

Eric Jhonsa
Eric Jhonsa

Editor

The "OEM and Other" segment, which among other things covers non-gaming GPU sales, saw revenue grow 31% Y/Y to $152M, beating a $128M consensus.

Nvidia attributes the growth to higher sales of entry-level PC GPUs.

Eric Jhonsa
Eric Jhonsa

Editor

The Professional Visualization (workstation GPU) segment continues seeing steady growth: Its sales rose 13% Y/Y to $331M, slightly beating a $328M consensus.

Automotive revenue, which covers infotainment processors and engagements for Nvidia's Drive autonomous driving platform, were flat Y/Y at $163M ($6M below consensus).

Eric Jhonsa
Eric Jhonsa

Editor

In line with its prior guidance, Nvidia says the gaming segment's notebook GPU and console SoC sales fell Q/Q, while its desktop GPU sales rose.

Eric Jhonsa
Eric Jhonsa

Editor

Gaming segment revenue, which covers sales of PC gaming GPUs and console processors, came in at $1.49B, a little below a $1.52B consensus.

Sales were down 10% Q/Q and (with sales depressed a year ago due to a channel inventory crunch) up 56% Y/Y.