Nvidia Is Taking Fresh Steps to Boost its Sales to Gamers
Nvidia (NVDA) just made a move to grow its sales to relatively cost-sensitive gamers. And it might be prepping yet another.
In January, Nvidia launched the GeForce RTX 2060, which at a $349 graphics card starting price was the cheapest GPU to date for its new Turing architecture. Today, the company followed that up by launching the GeForce GTX 1660 Ti, which carries a $279 graphics card starting price.
Unlike the 2060 and high-end Turing GPUs, the 1660 Ti lacks Nvidia's proprietary RT cores and Tensor cores, which are respectively used by games that support them to enable real-time ray tracing (a demanding graphics rendering technique that can produce photorealistic imagery) and (through a technique known as Deep Learning Super Sampling, or DLSS) the processing of AI/deep learning algorithms that can improve game performance. But for cost-sensitive gamers who are content to just rely on traditional graphics rendering, the 1660 Ti provides good bang for the buck.
In a 19-game test involving traditional rendering at resolutions ranging from 1080p to 4K, PC Gamer found the 1660 Ti to be moderately less powerful than the 2060, but also moderately more powerful than AMD's (AMD) Radeon RX 590 GPU, which is also priced at $279 and naturally doesn't pack RT or Tensor cores. AnandTech's tests, meanwhile, found the 1660 Ti to consume less power and run at lower temperatures than the RX 590.
The upcoming launch of AMD's Navi-architecture GPUs (they might arrive this summer) should put AMD on better competitive footing in the mid-range. But for now, Nvidia's mid-range lineup looks quite competitive, and (despite its recent sales pressures) the company remains dominant in the high-end gaming segment.
Separately, documents related to an HP Inc. (HPQ) gaming desktop and test results found in a graphics benchmark database suggest Nvidia might launch a GPU known as the GTX 1180. From the looks of things, the 1180 (assuming that's what it's called) would be very similar to Nvidia's RTX 2080 high-end GPU (it's priced at $699) when it came to traditional graphics rendering, but would lack the 2080's RT and Tensor cores.
With Nvidia having admitted that its Turing sales have been hurt by the fact that game support for the features enabled by its RT and Tensor cores remains a work in progress, and with high-end Turing GPUs priced above what high-end GPUs based on Nvidia's older Pascal architecture cost when they arrived in 2016, a cheaper alternative to the RTX 2080 could be well-received by some high-end gamers. That is, at least until more games supporting real-time ray tracing and DLSS arrive.
There is some risk that launching a high-end GPU that doesn't support ray-tracing and DLSS will give game developers less incentive to support the features. But the risk probably isn't too high, given the amount of support that now exists within the game industry for the technologies. And such an offering could make it a little easier for Nvidia to deliver on its guidance for a major second-half sales rebound.
Amazon's Logistics Efforts Are Becoming a Big Deal
Feb. 21, 2019 | 7:26 PM EST
Just as Apple still relies heavily on chips from third-party suppliers in addition to its own chip designs, Amazon still depends heavily on logistics services from the likes of UPS (UPS) , FedEx (FDX) and the USPS, and will likely continue to do so for a long time. However, just as Apple relies more and more on its own chip designs with each passing year, Amazon is steadily growing the portion of its shipping needs that are handled by its own logistics operations.
XPO Logistics (XPO) appears to be feeling the effects of this trend. Last week, XPO's shares tumbled post-earnings in part because the company disclosed its largest customer (widely believed to be Amazon) plans to stop using XPO to deliver parcels to USPS postal sort centers, from which the USPS takes care of last-mile delivery. Marc Wulfraat, the head of a logistics research firm, thinks Amazon most likely "figured out ways to reduce freight spend by leveraging their own assets and drivers so that they can replace XPO with a lower cost alternative solution provider."
Separately, online postage services firm Stamps.com (STMP) disclosed on Thursday (along with guidance that contributed to a 58% drop for its stock) that it's ending its exclusive partnership with the USPS, and that Amazon is one of the shipping partners that it's interested in working with.
"Currently, [Amazon is] operating 27 planes with contracts to have 40 planes by June of this year," noted Stamps.com CEO Ken McBride, while also highlighting Amazon's investments in a large Kentucky air hub and (though it's still in its early stages) the launch of the Shipping with Amazon (SWA) service, which provides last-mile delivery. "We are setting our corporate strategy assuming Amazon will be a big global player in shipping," he said.
Also of note: In its latest annual report, Amazon said for the first time that "companies that provide fulfillment and logistics services" are among its rivals. In addition to an air cargo fleet and SWA, Amazon's logistics efforts now include trucks, delivery vans and an ocean freight unit. There's also the Amazon Flex (sort of an Uber for package deliveries), through which contract workers are paid to handle last-mile deliveries.
On Amazon's Q4 earnings call, CFO Brian Olsavsky said Amazon can often do transportation work at a similar or better cost than third parties, and noted it can rely on its order and shipping data to make good decisions on where to invest. "We know where our demand is, we know where we're moving things between warehouses and sort centers, and by not involving third parties all the time, we found that we can extend our order [cutoff times]," he said.
There is one key difference between Amazon's logistics efforts and Apple's chip efforts: Whereas Apple's chip designs are only meant for Apple hardware, Amazon could conceivably provide its shipping services to third-party retailers, in much the same way that its warehouse-based fulfillment services are now offered to them. As McBride pointed out, such a move wouldn't be all that different from the rollout of Amazon Web Services (AWS), which started out by selling extra capacity on the data center infrastructure that Amazon had built to support its e-commerce operations.
Dropbox's Post-Earnings Selloff Looks ExcessiveFeb. 21, 2019 | 2:18 PM EST
From the start, Dropbox (DBX) has been a tough company for Wall Street to pigeonhole.
It's technically a cloud software/SaaS play, but one that depends heavily on consumers and small businesses for revenue rather than large enterprises; it spends a relatively small amount on sales and marketing; and it mostly relies on a giant, internally-developed, data center infrastructure rather than, say, Amazon (AMZN) or Google (GOOGL) .
The uniqueness of Dropbox's business model is arguably a key reason why the cloud storage and file-sharing software/service provider has traded at low multiples relative to many SaaS peers, and also why markets have been quick to sell off its shares in response to any moderately bad news.
The roughly 8% decline Dropbox is seeing on Friday following Thursday afternoon's Q4 report is the latest case in point. Dropbox beat Q4 estimates while reporting its revenue and billings rose 23% and 20% annually, respectively, and its paid user count rose by 400,000 sequentially and 1.7 million annually to 12.7 million.
The company also reported its average annual revenue per paying user rose by over $5 annually to $117.64, and -- though Dropbox says it doesn't expect any "material" revenue this year from its recent acquisition of e-signature software firm HelloSign due to M&A accounting quirks -- issued above-consensus Q1 and 2019 sales guidance. Its 2019 revenue guidance of $1.627 billion to $1.642 billion implies 18% annual growth at the midpoint.
So why is Dropbox selling off? The fact that billings (pressured by currency swings) were slightly below consensus might be playing a small role. But the bigger culprit, judging by the point in time yesterday when Dropbox plunged, is that Dropbox guided on its earnings call for its non-GAAP operating margin, which rose seven percentage points in 2018 to 12%, to be in a range of 10.5% to 11.5% in 2019. Dropbox also forecast its free cash flow (FCF), which was $362 million last year, would be in a range of $375 million to $385 million this year, which is below a $406 million consensus.
However, the main culprits behind Dropbox's margin and FCF outlooks are a set of planned 2019 investments, rather than any long-term margin or profit headwinds. The company plans to invest in developing and integrating HelloSign's offerings, and will be opening a new data center in the first half of 2019. In addition, though Dropbox took possession of a new San Francisco headquarters in mid-2018, the company is still building it out and won't be moving in until later this year, and until then will be recognizing rent expenses on both its old and new HQs.
Dropbox notes that if not for one-time expenses related to the new HQ, it would expect 2019 FCF to be in a range of $445 million to $465 million, even with its HelloSign and data center investments occurring. The company currently has an enterprise value (market cap minus net cash) equal to 19 times the midpoint of that FCF range, and equal to just 16 times a 2020 FCF consensus estimate of $537 million.
Given Dropbox's continued top-line strength amid ongoing competition from tech giants and enterprise software firms, as well as the continued health of key user metrics, its multiples look very reasonable at a time when many fast-growing software names are being granted much richer valuations.
To see Tech Check coverage from the previous trading day, click here.