This week is a remarkable week, with the world premier of Black Panther causing social media to blow up, while Florida saw a mass school shooting that saw 17 dead, all while the Olympics are happening in the most tense geopolitical region on the planet.
And yet, I choose to stay away from all those topics, and focus on the news that only a handful of you care about. Because I’m strange like that.
So let’s begin.
Google’s TPUs in Beta
This week saw Google’s famed TPUs launch in beta, which, in and of itself, is interesting. It’s said to be optimized specifically for machine learning applications, much like an ASIC processor, outperforming GPUs performing the same task. This last point is interesting, particularly since Nvidia just reported record earnings based on cryptomining and AI operations.
It’ll be interesting to see just how this development impacts GPU sales for enterprise customers, particularly in the AI space. But there are a couple of caveats. First, Nvidia is a hardware supplier, while Google is offering its TPUs as a cloud solutions platform. This difference in business model means that Nvidia’s sales will be impacted if Google is able to soak up the demand that Nvidia’s customers currently satisfy. That’s all contingent upon how many TPUs Google is able to supply, on top of efficiency metrics in terms of operations per second, or terraflops.
Second, developers who already use TensorFlow don’t have to do anything to take advantage of Google’s TPU servers. This implies that non-TensorFlow users will need to modify their code to be compliant. Whether the time and effort required to make the conversion is worth the apparent advantages of Google’s servers is something we have to wait and see, especially since Google’s charging $6.50 per cloud TPU hour, as opposed to $1.46 per cloud CPU hour charged to access standard Tesla P100 GPUs.
Third, Nvidia’s Volta cores are said to be able to run at 120 terraflops, which is rather competitive with Google’s reported 180 terraflops for their TPUs. This makes the math even more interesting, especially if Nvidia based servers keep costs in line with today’s prices.
Lastly, most, if not all, of the world’s top cloud service providers operate on Nvidia hardware. This includes Amazon and Microsoft, two juggernauts who have the marketing and sales budget to steam through whatever obstacle Google can muster up. As I mentioned before, the competition isn’t between Nvidia and Google. It’s between Google and Nvidia’s customers.
Snapchat’s New Lure
I never knew this, because I rarely, if ever, use Snapchat, but ostensibly Snapchat didn’t have an analytics function for its creatives and content producers. That’s one area that the company addressed this week. It’s largely a push to bring back influencers to the platform, many of whom left Snapchat for Instagram, which offers similar services.
As I’ve stated many times before, I’m bearish on Snapchat for many reasons, prime among which is that I don’t understand the value proposition of the company — for both end users and advertisers. But adding an analytics feature to this platform is definitely a boon, especially in convincing advertisers of Snapchat’s value.
This should help advertisers find creatives who cater to their target demographic better than in previous years, allowing them to segment creatives based on industry, area of interest, and demographic data. The gamble seems to be that if Snapchat allows creatives to widen their audience on the platform, advertisers will get enticed to bring business back to the platform, which will, in turn, show more influencers that they can monetize their Snapchat feeds, which will, ultimately, create a positive feedback loop.
They’re also hoping this drives up the total number of users on the platform, which is their primary metric for stock value on the market today.
Will this work? I’m bearish, but less bearish about this than most other things Snapchat has done. First, this offers transparency that creatives long desired from the platform. It allows them to know what’s happening, and plan around facts and figures, as opposed to intuition and gut feeling. Second, it puts them on a level playing field with competitors, who have long offered similar services. This should mean that, at least for influencers, Snapchat is no longer the underdog. Third, and this ties in to the previous point, is that it lowers the cost of entry into the Snapchat ecosystem. By offering this dervice themselves, it eliminates the need to rely on costly, third party solutions.
But, once again, will this work? Hopefully it does. I like this move more than any other the company has made. But, whether it can help Snapchat drive up DAU figures, and ultimately become profitable, is another question. It’s a good first step though.
The Future is Now Old Swiss
Apple outsold the Swiss in number of watches sold in the final, holiday quarter of 2017. This sounds remarkable, and it is, because one is a high tech juggernaut known for beautiful scenery and remarkable architecture, and the other is Switzerland, a bloody country.
But the sales figures don’t tell the whole story.
Bear in mind, Apple Watch prices pale in comparison to the ultra-expensive offerings of the Swiss. In fact, for Swiss watches, $3,000 is considered a “mid tier offering”. $10,000 is considered “mainstream” while the high end can command prices in the millions. So, are the Swiss in dire straits? Probably not.
But it is looking bleak.
This latest story is telling of a larger industry trend — people aren’t buying watches like they used to. Most people check their phones for the time, and, if they need a watch, they’re finding alternatives in smartwatches (a la the Apple Watch), and fitness trackers. The value proposition of the Swiss watch is falling, and it’s a problem the industry has known about for a while.
The solution for the industry, for the longest time, was China, and other developing markets. The rising middle class, and the appetite for prestige, made these markets primed for the taking by the Swiss, especially for the heavyweights, like Omega and Rolex, who have long positioned themselves as proxies for success. Those markets are now maturing, and are looking to consolidate their experiences, instead of looking for specialized gadgets to serve their needs.
So smartwatches are a definite threat to the Swiss watch industry. But, what people fail to understand are two things — the industry’s resilience, and the bigger threat that’s been threatening Swiss watch making. First, the industry is remarkably resilient. In the 70’s and 80’s, with the dawn of Seiko’s Astron quartz watch, the Swiss watch industry underwent heavy restructuring, and rebranding. Consumer brands became luxury brands; independent makers consolidated under one roof; marketing became the single largest expense. And the industry survived. In fact, they’ve actually thrived, expanding into luxury goods at large, becoming huge conglomerates, and economic engines for Switzerland.
But this second point is a bit more worrisome. Cheap, Chinese, mechanical watches. Companies like Sea-gull now have offerings that look, and function, as well as Swiss offerings, at less than half the price. Couple this with the immense marketing budgets targeted at the Chinese market, the single largest market for luxury watches by a mile, and it’s hard to say smartwatches are a bigger threat. What’s more is that the Chinese don’t see this as just another industry, but as a source of national pride — a showcase for their precision manufacturing prowess, akin to the ball point tips for ball point pens.
Smartwatches are wonderful fabrications of the IT age. But, in the grand scheme of things, they’re not as wide spread to be considered a threat for Swiss watch making. But the trend is something to be taken to heart — people aren’t buying Swiss watches like they used to. And it doesn’t look like they’ll be returning any time soon.