SITALWeek

Stuff I Thought About Last Week Newsletter

SITALWeek #263

Welcome to Stuff I Thought About Last Week, a collection of topics on tech, innovation, science, the digital economic transition, the finance industry, Skee-Ball tickets, and whatever else made me think last week. Please grab me on Twitter with any thoughts or feedback.

Click HERE to SIGN UP for SITALWeek’s Sunday EMAIL (please note some ad blocking software may disrupt the sign up form; if you have any issues or questions please email sitalweek@nzscapital.com)

In today’s post: the benefits of analog AI chips; the best way to fight bundles is with better bundles; the motivation of Nvidia's Arm acquisition; comic relief; approaching a stock like Snowflake in a portfolio; the crowded tech trade; and lots more below...

Stuff about Innovation and Technology
High Fashion Twitch
Twitch continues to expand its array of live entertainment – this time reaching into the luxury fashion market. Last week, Burberry was the first to hold a live fashion show on Twitch with “runway” models wearing the 2021 Spring/Summer lineup. Why did Burberry choose Twitch over YouTube, Instagram, etc.? Twitch reckons their considerable reach with Gen Z is their main advantage for brands. And, their platform is designed for active, lean-in engagement rather than passive consumption. The show was, um, interesting?

Nasdaq is Cloud Bound
Nasdaq, which operates 28 equity and derivative exchanges around the world, will move all of their on-premises data centers to the public cloud over the next decade. Within about five years, Nasdaq will migrate several US and European exchanges to cloud providers like Amazon, Microsoft, and Google, which offer a key advantage of scalability over legacy IT systems. Nasdaq had previously moved some of their data warehouse to AWS, which proved critical to their ability to easily accommodate the higher volumes of 2020’s market volatility. But, it’s not just about scale: “Market customers will be able to develop new computer models more easily using cloud-based tools for data analysis, such as risk calculations that are computing intensive...‘Doing that with traditional on-premise hardware [is] limited,’” said Nasdaq IT chief Brad Peterson.

Analog AI Chips
Ambient audio surveillance devices, like Alexa and Google Home, use power-hungry, always-on digital signal processing chips (DSPs). Not only are the microphones always listening, the DSPs are constantly analyzing ALL the input for the appropriate wake word/phrase. With the rise in ambient surveillance of sounds and objects, there’s heightened demand for energy-efficient solutions, one of which could be analog-based AI chips. EE Journal writes about Reconfigurable Analog Modular Process (RAMP) devices, made by Aspinity, that can monitor audio or vibration sensors and run local inference against a model – all in analog – to weed out everything that’s not human speech before high-power processing. Here is a short video on how RAMP works. Aspinity claims that RAMP offers a 90% power savings that – when multiplied by the tens of billions of audio and vibration sensors in the future connected world – would really add up. EE Times posted an interview with Mike Henry of analog AI company Mythic, who explained that analog AI chips can function in SRAM and flash memory, which gives them instant boot and no need for power-hungry DRAM to perform inference workloads. Analog AI chips could work well in the nascent industrial IoT market, an example of which is Amazon’s rumored project Thor, which uses temperature, sound, and vibration sensors on industrial machines, such as the ones you’d find in a factory or an Amazon fulfillment center, combined with machine learning to predict maintenance and downtime.

USPS at Risk of Losing Lucrative Amazon Business
In 2019, the USPS delivered about 1/3 of Amazon’s US packages, garnering $3.9B in revenue (around $2.50/pkg) and roughly $1 in profit for each of the 1.54B packages. Amazon has recently indicated it might be delivering as many as two thirds of its own packages. FedEx no longer works with Amazon, and I’d loosely estimate that UPS delivers far less than 10% of Amazon packages in the US. It appears the USPS is highly vulnerable to Amazon’s ongoing delivery insourcing. Bloomberg reported last week that Amazon is opening up 1,000 more small delivery-hub warehouses in cities across the US and will increasingly rely on contract delivery agents for grocery orders.

An Apple Watch a Day to Keep Dr. Away?
The Singapore government is partnering with Apple to pay citizens to wear Apple watches for health benefits: “As part of the scheme, Singapore residents will be able to earn as much as S$380 ($280) in rewards and vouchers by completing goals and tasks set within the app. Goals can be accomplished by walking or doing other exercises like swimming or yoga, and the LumiHealth app will offer personalized coaching and reminders for health screenings and immunizations. Wellness challenges will nudge users toward making better food choices and improving sleep habits.”

Fight Bundles with Bundles
Spotify complained again last week after Apple rolled out some new bundles of music, video, news, and other miscellaneous stuff that no one particularly wants. But, that’s the paradox of bundles: by combining various options with varying degrees of quality and interests, you can create a positive sum value proposition. Rather than complain about it, I’d suggest Spotify get much more aggressive with bundling and partnering and create something that consumers really do want. Speaking of bundles, AT&T will offer phone service next year subsidized by advertising. Amazon took this tactic with its Fire and Kindle devices, but I haven’t heard about the effort in a long time, so I am not sure whether it was effective. The carriers have more data on you than anyone, and you’ve given them permission already to use it, so this could be a lucrative move for AT&T. Even more broadly (as I mentioned last weekcarriers are emerging as media bundling platforms, and advertising goes hand in hand with media as part of that value proposition. AT&T’s ad-supported HBO Max streaming service will launch next year and is likely to be part of this ad-subsidized phone service. Want to send a text? Just watch this ad first...

NFL’s Lackluster Debut 
Despite theories that lockdown-starved sports fans would clamor for the return of live sports, NFL viewership was down anywhere from 10-38% versus last year. In part, the problem is the 8% y/y decline in subscription-TV households in the US; but, I can’t help but wonder if the proliferation of new forms of content gaining share during COVID (that I wrote about last week) is also at play here. For now, we’ll see how live sports play out (ha). And, while I am risking even more confirmation bias about how interactive content is gaining share, here’s what Take-Two’s CEO, Strauss Zelnick, said to Protocol last week: “People realize that interactive entertainment is the standard bearer of the entertainment business. It is the most important entertainment business. It is the largest, it is the most rapidly growing, and it appeals to a very broad audience. It's become America's pastime and the world's pastime.” 

ARMing Nvidia
Jon Bathgate weighs in on the Nvidia-Arm deal: After several months of speculation, Nvidia announced the acquisition of Arm Holdings for $40B last week, which would mark the largest (and arguably most strategic) deal in the history of the semi industry if it closes. While the combination of Arm and Nvidia would touch every vertical across the semi industry, this deal is largely about the future of the data center, where Nvidia has quietly been escalating its support of Arm over the last 18 months, including with their announcement of GPU-accelerated Arm CPUs at last year's Supercomputing '19 conference. It likely became clear to Nvidia co-founder Jen-Hsun Huang that not only would Arm benefit from Nvidia's roadmap and R&D resources, but Arm will likely be significantly more valuable directly because of Nvidia's support for Arm, which Huang expressed in a conference call last week: “We’re going to accrue so much value to this architecture in the world of data center. Before we go and make that gigantic investment and gigantic focus, why don't we own it?”

It isn't totally clear that Nvidia will design their own Arm-based CPU for the data center (which they could have done with an architecture license rather than buying Arm outright). Huang seems more interested in continuing to diminish the role of the CPU in the data center, allowing GPUs and smart NICs to continue to take over the CPU's role, which he discussed in this interview highlighted in SITALWeek #243. There is also the added benefit for Nvidia that driving Arm in the data center will put continued pressure on top-rival Intel, and, to a lesser extent, AMD. (Color commentary from Brad: I’m a skeptic on this deal, but I think this might be the key point: from a game-theory perspective, accelerating Arm in the data center effectively kills Intel and AMD as future competitors for Nvidia.) We will see the ripple effects of this deal for the next decade across the industry – while Nvidia has pledged to support Arm's neutral "Switzerland" business model, it would not be surprising to see further accelerated adoption of RISC-V (see this post from Feb 2019 for more context), which has been gaining significant traction within the semi ecosystem in recent years. With Arm losing neutrality and facing one to two years of regulatory scrutiny and, thus, a more uncertain future, RISC-V gaining ground seems a likely outcome. What is clear from this acquisition is that Nvidia is in position to drive the future of the data center, taking the baton from Intel after 30 years of dominance. (More Color commentary from Brad: it’s noteworthy how much more negative the outlook for Arm is today than when Softbank acquired the company in 2016 – largely owing to the slowdown in mobile and threats from RISC-V; at the time of the Nvidia deal announcement, and excluding the potential earn out, Softbank will have a return on their investment of only ~5% (plus the FCF that Arm generated along the way) compared to a 100%+ rise in the Nasdaq.)

Miscellaneous Stuff
So. Many. Tickets.
Bankrupt children’s pizza and arcade chain Chuck E. Cheese is seeking bankruptcy court approval for $2.28M to destroy 7 billion of those little tickets you get from playing games to turn in for prizes. Boy, what I would have done with 7B Chuck E. Cheese tickets when I was eight...😭

How to Decrease Racial Inequality in Home Ownership
Redfin’s Glenn Kelman penned an editorial for Time Magazine’s Ideas site last week about the racial divide in US home ownership and possible solutions for addressing it. An inability to build wealth via asset ownership is a major contributing factor to inequality, and there are basic steps the US can take to help resolve it.

Making Chess Beautiful Again
DeepMind’s AlphaZero chess AI has come up with creative, new ways to play nine altered variations of the game, such as “no-castling” or sideways pawn movement. Historically, a chess engine’s only goal was beating their human opponent; however, for DeepMind, the endgame is to make chess more fun to play and watch. Chess champ Vladimir Kramnik, who worked with DeepMind on this project, said: “All in all it just makes the game more beautiful.” As I discussed in SITALWeek #260“Humans are tool-building machines. We’ve co-evolved with the tools we’ve created since the first stones were put to use.” It’s basically what we do – it’s what natural selection landed on for driving the evolution of our species. Other animals use objects in their environment, but (as far as I know) we are the only ones working on brain interfaces like Neuralink. Often, the tools we create make us better, more creative problem solvers, and nature loves problem solvers – it favors their reproductive success, but I digress...

Comedians Offer Respite and Vital Dose of Reality
NY Times interviewed Chris Rock ahead of his starring debut in the new season of Fargo on FX. Comedians are so important for shining light on the world around us, and among the many things we are missing during the pandemic are comedians like Rock who can stand up in front of an audience and tell us how stupid we’re all acting by comically pointing out the obvious. SNL’s 46th season will hopefully offer a significant comic respite from 2020 when it kicks off the first week of October. Lorne Michaels commented in Vulture magazine on the upcoming season, which will feature Jim Carrey as Joe Biden opposite Alec Baldwin as Trump in front of a live studio audience at Studio 8H. Michaels: “We came on in ’75, and the last helicopter of Saigon was ’75. And there was Watergate, of course. When I got here, the city felt abandoned and broke. But it was also a really exciting time to be in New York, and we were part of the rebirth. I don’t think things are nearly as bad now as they were then. But things go up and down and go through bad periods, and you just want to be able to express what you’re thinking about all of that, which comes from some place of thought." And he concluded: "All I feel is what I’ve always felt, which is it’s really important to get it right. And laughs are the clear indicator. That’s why the audience is so important. Because you just can’t come out and express your political opinions. There has to be something, something that gets close to the truth that you’re doing and that’s honest. And that’s where the laughs come from."

Evenstar Enya
Pitchfork reports on the surprising reach and influence of Enya, who has quietly and profoundly influenced countless musicians and genres of music.

Stuff about Geopolitics, Economics, and the Finance Industry
Arming Taiwan
The US is provoking China again by selling another $7B in arms to Taiwan. This latest deal would make $22B in arms sales to the insular semiconductor hotseat during Trump’s four-year presidency (vs. $14B during Obama’s eight years in office). The US is also blocking more imports of goods coming from forced labor camps in China, according to the WSJ

All Aboard the Crowded US Tech Train
A Bank of America fund manager poll indicates that owning US tech stocks is the most crowded trade on record for the survey.

❄️😵
Speaking of the crowded US tech trade, last week saw the IPO of multi-cloud data warehouse provider Snowflake. I don’t have a strong view on the company, except that it’s a good product that has solved a problem many have tried and failed to do before, and their success proves that customers have a desire to not get locked into native cloud apps, such as those offered by AWS, when cloud alternatives like Snowflake exist. Snowflake’s fully diluted market cap on its first day surpassed $100B (including the 100M+ warrants and options outstanding, which tend to be left out when you look up market values when companies are still loss making). Snowflake did $265M in revenues in its fiscal year ending January 2020, and they have been growing over 100% – most recently 121% to $133M in the July quarter. The following is by no means any sort of official estimate, but if you assume they maintain the impressive 120% growth rate this year and then double next year, you get to around $1.2B in revenues; the stock was, at one point after its IPO debut, trading at about 100x that revenue number when it hit $319/share. I saw some crazy stuff in the late-1990's stock market, and it’s likely I don’t remember everything that happened, but I don’t recall a legitimate business ever getting to that level of hope for the future (there were many zero-revenue IPOs, but it’s hard to put a multiple on zero!). That's not to say the company won't live up to the hope. Much in the way that people make fun of Warren Buffett or other value investors struggling to understand the technology shifts occurring in the global economy, I feel a little out of place when I see things like this in today's markets. So, you can imagine how my head exploded upon learning that Berkshire Hathaway invested $730M in Snowflake’s IPO (at the IPO price, which was about a $40B value, or 33x that $1.2B revenue guess) and more than doubled their money. That article quotes Buffett from a 2012 shareholder meeting: “I mean, the idea, that somebody is bringing something to market today, a seller who has a choice of when to come to market, and that that security, where there’s going to be a lot of hoopla connected with it, is going to be the single cheapest thing to buy out of thousands and thousands and thousands of businesses in the world is nonsense, you know.” Hey kids, get off my lawn!

All joking aside, there is a way to approach a stock like Snowflake in a portfolio: match the position size to 1) the predictions the valuation is forcing you to make, and 2) the range of outcomes around those predictions. A broad prediction to start with is: data will continue to explode and be collected/analyzed with AI and machine learning. It’s also a broad and safe prediction that databases over time will migrate to the cloud. And, it’s an increasingly broad prediction that customers will want those data stores to be cloud agnostic, i.e., not locked into a single cloud platform such as AWS. The size of the data market can be determined from various industry analysts; from there, you can make some assumptions regarding the future market share of Oracle, Microsoft, the public cloud platforms, and the host of other startups. Then, you ask the question: starting with Snowflake’s current valuation, what market share and margin structure will they need to achieve to be a good long-term investment? Betting on them meeting those metrics is likely to require a rather narrow prediction with a wide range of outcomes. Now, set the position size: the narrower the prediction and the wider the range of outcomes, the smaller the position size should be.

Disclaimers:

The content of this newsletter is my personal opinion as of the date published and is subject to change without notice and may not reflect the opinion of NZS Capital, LLC.  This newsletter is simply an informal gathering of topics I’ve recently read and thought about. It generally covers topics related to the digitization of the global economy, technology and innovation, macro and geopolitics, as well as scientific progress, especially in the fields of cosmology and the brain. I will frequently state things in the newsletter that contradict my own views in order to be provocative. Often I try to make jokes, and they aren’t very funny – sorry. 

I may include links to third-party websites as a convenience, and the inclusion of such links does not imply any endorsement, approval, investigation, verification or monitoring by NZS Capital, LLC. If you choose to visit the linked sites, you do so at your own risk, and you will be subject to such sites' terms of use and privacy policies, over which NZS Capital, LLC has no control. In no event will NZS Capital, LLC be responsible for any information or content within the linked sites or your use of the linked sites.

Nothing in this newsletter should be construed as investment advice. The information contained herein is only as current as of the date indicated and may be superseded by subsequent market events or for other reasons. There is no guarantee that the information supplied is accurate, complete, or timely. Past performance is not a guarantee of future results. 

Investing involves risk, including the possible loss of principal and fluctuation of value. Nothing contained in this newsletter is an offer to sell or solicit any investment services or securities. Initial Public Offerings (IPOs) are highly speculative investments and may be subject to lower liquidity and greater volatility. Special risks associated with IPOs include limited operating history, unseasoned trading, high turnover and non-repeatable performance.

jason slingerlend