Welcome to Stuff I Thought About Last Week, a collection of topics on tech, innovation, science, the digital economic transition, the finance industry, lightning, 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: Aggregating health wearable data with AI; vertical integration in the analog-to-digital shift; the “software is eating the world” memo failed to arrive for many industries; more context on Intel and its path forward; how to think about valuations on high-growth software companies – a lesson in expected returns and range of outcomes; and lots more below...

Stuff about Innovation and Technology
Skin Temp Trumps Internal Temp
Connected ring maker Oura released some interesting insights on the value of temperature data for predicting health. Health wearables measure skin surface temperature, which, surprisingly, may be more useful than measuring internal temperature. The body works hard to keep your core temp between very tight bands of 1-2 °F (0.5-1 °C), and it uses your skin to help accomplish that. As a result, your skin temperature can vary by 27 °F (15 °C) in a day. Oura measures skin temp every minute, and new study results validate the readings as being useful for a variety of potential health indicators. Deriving the most powerful future health benefits from wearables may lie in aggregating data from multiple wearables, your phone’s sensors, and third-party devices (like connected blood pressure cuffs) and running it all through a local AI inferencing model on your phone (for privacy) to detect potential problems. And, in the more distant future, imagine a smart toilet with sensors sending data to your phone-based AI health hub to diagnose more troublesome diseases like cancer.

Alto Luxury Taxi
Alto is a Texas-based ride-hailing company that operates its own vehicles, employs drivers directly, and provides benefits. In addition to providing on-demand rides, the company will also deliver food and run errands. The cost, according to D Magazine, runs around double (or more) similar rides from Lyft and Uber. Alto is expanding to Los Angeles from Dallas and Houston. I was not previously familiar with Alto, but it seems more accurate to describe the startup as a modern, luxury taxi service rather than a ride-sharing company. A common pattern that we see when a legacy business model goes digital is that vertical integration is key to creating the highest value platform. Uber and Lyft are horizontal demand aggregators with decentralized drivers using their own cars for ride sharing and third-party restaurants for food delivery. That model has so far created a lot of consumer surplus, but it has not necessarily created a lot of value for drivers or profits for the platforms themselves (and mixed results for restaurant profits). Although Alto is small (100 drivers and 60 SUVs in Dallas), I’ll be interested to watch their progress as a more vertically-integrated digital platform. In the food and grocery delivery space, I’ve long argued vertical integration and new business models (e.g., dark kitchens and purpose-built grocery fulfillment centers) will have the best chance of success, and it’s worth considering how many horizontal gig platforms would be much more successful as vertically-integrated service providers. For an extreme example of the value of data in vertical integration to improve service quality, checkout how Picnic uses genetic algorithms to schedule trucks.

It’s All Software these Days
It’s been nearly ten years since Marc Andreessen explained that “software is eating the world”; but, even with a decade to prepare, most analog, Industrial-Age businesses remain completely naive or willfully ignorant. The WSJ ran an article on the problems VW has had with software: “After years of development, Volkswagen decided in June last year to delay the launch and sell the first batch of cars without a full array of software, pending a future update, which is now scheduled for mid-February. Tens of thousands of ID.3 owners will have to bring their cars in for service to have the new software installed.” One of VW’s mistakes was to outsource its software development to parts supplier Continental, which is itself an analog, Industrial-Age business. The rising complexity of software- and AI-driven product transitions seems to require early vertical integration in order to succeed. Tesla is the obvious example of vertical integration leading to an edge in EVs. One of the most successful examples of vertical integration leading to a large platform lead is obviously Apple, which is heavily involved in phone development (including its own processors) but outsources manufacturing (which has created some supply chain human rights violations). Apple may take the same tactic in cars by using an existing car maker as an outsourced facility for a rumored Apple Car. The FT reports Hyundai may be looking at that outsourced role. The rumor raises a broader question of who is best positioned to bring analog businesses into the AI Age? Will more software leaders of the 21st century increasingly use the 1900’s roster of industrial companies as contract manufacturers across a wider range of products? It’s a basic idea of abstraction, and the heart of the issue is where the intelligence lies and where platform value is created.

Technology Taxes to Restore Local Coiffures?
As industries go from analog to digital, one unexpected result is a loss in local taxes. This issue was identified and partially resolved in ecommerce, but it’s starting to impact other areas. Gasoline taxes are one interesting area to watch as cars shift from combustion to EV – those taxes often pay for road maintenance, and the EVs are using the same roads. (Unrelated, checkout these commercialized EV batteries with nanoparticle electrodes that have a 5-minute charging time!). Another interesting example is local cable franchise fees paid by the cable operators to towns across America. As people shift to streaming, this revenue is lost, so three municipalities in Georgia are suing Netflix, Hulu, and others to get 5% of their gross revenues for people living in those districts. Overall, a “digital tax” is likely coming to many areas of the economy.

Intel: The Case Against Outsourcing
Jon Bathgate chimes in on Intel: We heard another call for Intel to spin out its manufacturing operations last week from Ben Thompson, echoing the message from Third Point's open letter to Intel last month. The fundamental problem with this argument is that, while a potential Intel foundry would be forced to survive on its own (which would certainly energize some much needed innovation), it would have no customers besides Intel's core operation, at least for several years and possibly much longer. The context is important here – Intel has dipped its toe into the foundry market several times over the last decade but never made any noteworthy progress, and its flagship foundry relationship with Altera was a disaster up until it was swept under the rug when Intel acquired Altera outright in 2016. The foundry business is all about manufacturing leadership, ecosystem, and customer centricity, and Intel is unlikely to have any of these attributes for at least the next five years.

Why would any customer choose to work with Intel's inferior technology with no surrounding ecosystem? Obviously, manufacturing on US soil is an advantage for sensitive applications like aerospace and defense, but there is very clear momentum for US chip manufacturing starting with TSMC's Chandler, Arizona fab – which they deemed "Phase 1" on their earnings call last week (hinting at plans for a much larger US footprint) – and Samsung is likely building out more leading-edge capacity on US soil as well. The US government could offer to be a demand backstop for Intel, but it would need to offer up billions of dollars in support (in the form of guaranteed demand); we feel US government support would be much better put to use supporting the broader semiconductor ecosystem – including Samsung, TSMC, and their customers – as they move towards propping up leading-edge manufacturing in the US.

In the meantime, based on Pat Gelsinger's early observations at Intel, the company is committed to the IDM (integrated device manufacturer) model. He also indicated they will work more closely with the broader ecosystem (which Intel has not done well historically) and has not shied away from the idea of licensing process technology from Samsung or TSMC (as mentioned by Bob Swan in this interview), which could be a fairly elegant, non-zero-sum outcome for all parties. The computing part of the cloud is increasingly a commodity as the data center itself replaces the server as a broader unit of compute. In that context, Intel needs to remain cost and power competitive per unit of performance against Arm, ASICs, GPUs etc. See Intel’s Paths Forward in #277 for more context.

Fake News Engine Stalls
WaPo reports that online misinformation about election fraud dropped 73% after Trump was banned across social media sites. “The research by Zignal and other groups suggests that a powerful, integrated disinformation ecosystem — composed of high-profile influencers, rank-and-file followers and Trump himself — was central to pushing millions of Americans to reject the election results and may have trouble surviving without his social media accounts.”

Miscellaneous Stuff
Stratospheric Lightning
Blue jet lightning is a type of lightning that shoots straight up from the tops of clouds up to around 50 km above the surface of the earth and lasts for several hundred milliseconds. The phenomenon may happen as a result of turbulent mixing within a cloud that causes oppositely charged regions to come within close range (~1 km) of each other. A cool looking blue jet was recorded by the space station in 2019 and recently analyzed in a Nature paper.

Stuff about Geopolitics, Economics, and the Finance Industry
Analyzing SaaS Growth Stocks
The largest forty companies that went public in the last two years have current market caps totaling $1.1T. In our year-end letter, we mentioned that some growth stocks imply a single-digit expected return over the next decade. By way of example, we created a simple matrix that yields implied expected returns of high-growth SaaS stocks. (If you aren't interested in a speculative dive into high-growth stock valuations and hurdle rates, now is the time to jump to the end for conclusions or better yet, go do something more productive with your day!) Longtime readers know how I feel about DCFs (see “Non-Ergodic Systems Bury the DCF”), and this matrix is guilty of being a type of implied DCF. The nice thing about using SaaS for this example is it has several knowns with reasonably predictable ranges of outcomes. The TAMs (total addressable markets) for various parts of enterprise software are well analyzed by the industry, and we are far enough into cloud adoption that we know the degree to which it can be TAM expanding. There are also reasonable estimates that put total cloud spending at ~$1-1.5T in ten years. We have a decent idea about mature FCF margins because it’s fairly easy to look at a category leader in on-premises software and subtract 10-20 points of margin for the cloud infrastructure tax (depending on the data intensity of the application). That leaves us with two macro assumptions in the matrix, which are the same two that make all DCFs a challenge: 1) predicting interest rates and 2) knowing the path through time (see prior link for more). On interest rates, what I did for the purposes of this table was to assume that, in ten years, a quality compounding-growth company would have a similar multiple of FCF as it has today, which is around 30x forward FCF (I realize that’s a completely debatable statement on multiple levels!). I would also advise that, if you do an exercise like this one, you use the real, fully-diluted share count and adjust ten-year growth rates by several percent in dilution per year for stock option grants.

With those assumptions and warnings aside, I’ll give one final cautionary note: this is not meant to be stock advice. I am not talking about any specific stocks, and this is just a tool to think about implied expected returns for high-growth stocks. For this matrix, I assume an FCF multiple of 30x in ten years, an FCF margin of 35%, and I start every company at $1B in revenues today because many SaaS stocks are around that level (or will be approaching it soon). Let’s pick a couple of examples of hypothetical stocks out of the table based on those parameters. All multiples below are EV to forward-1-year revenues.*
-A company that can compound revenues at 10% per year from today will give you a 5% expected return if it’s trading at 15x, or a 12% expected return if it’s trading at 8x. In ten years, it would have $2.3B in revenues.
-A 20% topline compounder over ten years trading at 21x today will give you a 9% expected return and have $5.2B in sales in ten years. At 15x, it would yield 13% expected return.
-A 30% topline compounder over ten years trading at 39x today will give you a 9% expected return and have $10.6B in sales in ten years. At 24x, it would yield a 15% expected return.
-Just for fun, a 50% compounder over ten years trading at ~100x sales would give you an expected return of 11-12% and end with $38B in revenues and $13B in FCF.

What did I learn from this? It comes down to your expectations around the range of outcomes for a specific high-growth software business and what your expected rate of return is for the risk associated with that range of outcomes. The faster a company is growing, in general, the wider the range of outcomes and the higher sensitivity to the growth rates. So, is a 5-6% expected return worth an investment given the range of outcomes for the early-stage company and its ability to win or expand its TAM? It’s important to look at the revenue number in ten years and think about the TAM the company operates in, as well as adjacent TAMs. It is very difficult to find examples of software companies that compounded well above 20-30% once they achieved over $1B in revenues (but, it’s also still a relatively small sample size for cloud software companies of this size, and cloud is obviously different than legacy on prem). This comes down to your own hurdle rates and your portfolio construction process. At NZS Capital, we account for these variables in position sizing. If the range of outcomes is wide and the expected return is not as high as one would like to accommodate that range of outcomes, then it is an Optionality position if the asymmetry still warrants owning it. The last point I would make is that the rare SaaS company with real FCF today has the potential to create higher returns over time by redeploying that capital into M&A or stock buybacks and dividends. That scenario is hard to capture in an analysis like this, but it’s important to think about how the management team will allocate capital, whether it’s from FCF or dilution from acquisitions and stock options.
✌

*The equation for each ‘EV / 1-Year-Forward Revenues’ cell is:
[(Starting Revenue) x (1+Rev CAGR)^9 x (FCF Margin) x (FCF Terminal Multiple)] / [(1+Hurdle Rate)^9] / [(Starting Revenue x (1+CAGR)]

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.