328: Snowflake Saturation, Amazon DTC Push vs Shopify, Danaher Slice, Adobe + Figma, US vs China, Coal & Battery Prices, and Streaming Emmys

"Deriving meaning and purpose from within."

What Paul says about Peter tells us more about Paul than about Peter.

―Baruch Spinoza

🛀 A less cheery shower thought today: We’re pretty lucky that enriching uranium to bomb-grade or making plutonium, and getting everything else right to get a nuclear explosion, is as hard and costly as it is.

Imagine if all that stuff was much easier than it turned out to be?

What are some other technologies that we’re going to master in the future that could be similarly destructive, but they’re not nearly as hard and expensive to execute (or won’t stay hard and expensive long, because they’re largely digital)? There’s a reason many very smart people are worried about engineered bioweapons and AI 🤔

🥳 🥂🍾 Big news from my friend MBI (💎🐕) who joined Contrary Capital as Director of Research.

I’m very happy for him on the big Life Decision™️ and wishing him the very best — I’m sure he’ll kick ass and grow quickly in this new role (but don’t worry, MBI Deep Dives continues, because MBI is literally so good that he can do two jobs).

📒💡 TIL: Autotelic:

1. (of an entity or event) Containing its own meaning or purpose.

2. (of a person) Deriving meaning and purpose from within.

3. (art, of a work of art or literature) Not motivated by anything beyond itself; thematically self-contained.

I feel like that’s what I’m largely looking for in life.

Things that provide meaning and satisfaction in themselves, that aren’t a mean to get somewhere else, to some other thing.

I’m not reading all day and going clickety-clack on the keyboard to write this stuff because I want to get a job somewhere else, or to start a VC fund or whatever. I like the thing itself and want to keep doing it as an infinite game where the goal is just to keep playing.

There’s no destination, all of the meaning is contained in the journey itself.

h/t My friend David Senra (🎙📚) for teaching me that word on our call yesterday.

🖼 🎨🤖 Reader Transitory (👋) had a good question on the whole AI Art stuff:

there are many artists who are really good, but they never make it because they are not in the circle of influence. I wonder how that circle will be defined in the new AI world. Who or what will hold the power of influence?

That's a good question.

It feels like in the internet age, there are fewer gatekeepers and — given enough time — the internet is good at finding good stuff. So maybe there will be fewer obscure artists that only make it long after they're dead. Nowadays, someone on Reddit would’ve found it and made it go viral earlier.

But with a lot more supply, it's certainly possible that the 'churn' will also be higher, and that it'll be harder to remain on top for long, and the ‘notability’ bar may also be higher because of higher competition.

Though at the same time, there are so many more viable niches now than before, it's probably *easier* to stay at the top of some niche than back when there were just a handful of big 'center of the bell curve' verticals and everything else never reached critical mass 🤔

💚 🥃 Lots of things in life are complex, but some are simple:

If you feel like you’re getting value from this newsletter (and the semi-regular podcast), I simply ask that you consider becoming a paid supporter — it’s the only way I can feed my kids and justify spending my days doing this. Thank you!

---

🧩 A Word From Our Sponsor: Heyday 🧠

Do you have 100+ browser tabs open right now? 😬

Give your memory a boost with Heyday, the research tool that automatically saves the content you view, and resurfaces it within your existing workflows. 👩‍💻

It’s like cheat codes for your memory. 😲💡

🧩 Give your memory a boost today 🧠

---

🏦 💰 Liberty Capital 💳 💴

🎣 Amazon’s DTC Merchant Push 🚚 📦📦📦📦

Interesting move:

Amazon announced new marketing solutions that enable direct-to-consumer (DTC) ecommerce sellers to reach engaged Amazon shoppers and help drive sales on their own sites. Amazon is piloting the solutions as part of Buy with Prime, a new program that enables sellers to offer Prime shopping benefits to shoppers off Amazon. [...]

Participating sellers can showcase their DTC products on Amazon with a new Buy with Prime page within their brand’s store, a customizable multipage storefront on Amazon. Sellers can then direct Amazon shoppers to those DTC products by using Sponsored Brands, which are customizable ads for brands.

But there’s more:

Co-branded Buy with Prime social media ads, which are funded and managed by Amazon, help participating Buy with Prime sellers reach shoppers on Facebook and Instagram and drive traffic to products on their DTC sites.

Potential customers see ads from Amazon’s Buy with Prime page on Facebook and Instagram that feature specific sellers. When the shopper clicks on that ad, they’re taken to a product detail page on the seller’s DTC site where they can purchase directly from the seller using Buy with Prime.

This sounds a lot like what Ben Thompson (💚 🥃 🎩) suggested Shopify should do — by buying ads as proxy for merchants, they can use their aggregated 1st party targetting data in a way that no single merchant can match post-ATT. Ben called it ‘Shopify Advertising Services’.

❄️ Snowflake on what a ‘saturated’ customer is ☃️

An interesting dynamic with Snowflake is that they aren’t SaaS, they don’t charge $X per user per month or year.

They charge for consumption of compute and storage (as far as a I know they don’t really make margin on storage, they mostly pass through what the infra layer charges them). It takes a long time for customers to migrate all their data and workloads, and once they’re done, they tend to keep finding more use cases, and data naturally keeps growing year after year…

Historically, it took an average of around 8+ months for new customers to ramp up. But when do they stop? When do they reach saturation and level off? Let’s see what Snowflake CFO Michael Scarpelli thinks about this:

Q: can you maybe offer any sense of what a saturated account looks like if there is such a thing, which I have to imagine, there's always going to be data growth. But what is saturated?

A: Well, I would have thought our largest customer was a saturated account 2 years ago, and it's kind of gone from $29 million to $49 million run rate. And I know there's more opportunity there. So I don't really know what a saturated account looks like. I haven't seen it.

I’m pretty sure that account is Capital One.

But what kind of gives me the confidence, if you look at what Global 2000 companies are spending a year on hardware and people to manage that hardware and all the different security around that hardware, [our average of] $1.2 million is nothing.

I used to be the CFO of ServiceNow. We ran all of our infrastructure internally for like running SAP and HANA. And when you saw what you would spend a year for your hardware for HANA, this is nothing. And that's what gives me the confidence that they know it's a very small ticket relative to where they can go.

And when I look at our 246 customers that are $1 million plus a year, on average they’re $3.5 million… I'm not going to say this is going to happen, [but] within 6, 7 years, but they could easily spend $10 million a year on average easily.

The good news for Snowflake is that this ramping-up period for new customers seems to be getting shorter:

It was taking 240 days to get them up to that contracted rate that they were consuming within the first year. Now we've accelerated that with the help of partners and tools, that is now about 210 days to get them fully ramped into production.

Danaher: It slices, it dices! 🔪 🥩

After spinning out the industrial businesses into Fortive, Danaher is now spinning out its ‘Environmental & Applied Solutions’ division into another public company to focus almost entirely on life sciences:

The new company will be comprised of Danaher's Water Quality and Product Identification businesses and will be referred to as "EAS" until it is named at a later date. The transaction is intended to be tax-free to Danaher shareholders and expected to be completed in the fourth quarter of 2023.

Press release.

🏭 Global Coal Prices 🤔

Source.

🖌 Adobe to acquire Figma for $20bn 🎨👩‍🎨 💰💰💰💰💰

A half-cash-and-half-stock deal (plus another 6 million shares as RSUs to Figma’s CEO!). It makes a lot of sense on the product side, but it remains to be seen how much financial sense it will make.

From the release:

[Figma] is expected to add approximately $200 million in net new ARR this year, surpassing $400 million in total ARR exiting 2022, with best-in-class net dollar retention of greater than 150 percent. With gross margins of approximately 90 percent and positive operating cash flows

As for the strategic importance to Adobe, my friend MBI (💎🐕) puts it well:

For a "monopoly", nothing is as important as maintaining their monopoly, especially when FTC is sleeping or busy dealing with "fake monopolies".

ADBE may have ~50x ARR for Figma, but it was indeed a potential threat to ADBE's monopoly in Creative Cloud.

🇺🇸 U.S. Considering More Sanctions on China 🇨🇳 to protect Taiwan 🇹🇼

In Washington, officials are considering options for a possible package of sanctions against China to deter Xi from attempting to invade Taiwan, said a U.S. official and an official from a country in close coordination with Washington.

U.S. talks over sanctions began after Russia invaded Ukraine in February, but took on fresh urgency after the Chinese reaction to Pelosi's visit, the two sources said.

Taiwan is also asking Europe for help:

Taiwan has not asked for anything specific, only for Europe to plan what actions it may take if China attacked, one source briefed on discussions said, and has asked Europe to warn China privately that it would face consequences.

EU officials have so far shied away from imposing tough sanctions on China over human rights issues, as the country plays a far bigger role for the bloc's economy than Russia, said another person familiar with the matter. (Source)

You think global supply chains have been a mess in the past couple years? This would be something else…

‘The optimal amount of fraud is non-zero’ 💳 💸

Great stuff by Patrick McKenzie (aka Patio11):

Fraud is a unique subset of crime which occurs, to a major degree, subject to the enforcement efforts of non-state actors. A commanding majority of all fraud which is stopped, detected, adjudicated, and even punished (!) gets those done to it by one or more private sector actors. And the private sector has, in this case, policy decisions to make, which, like the public sector’s decisions, balance the undesirability of fraud against the desirability of social goods such as an open society, easy access to services, and (not least!) making money. [...]

The marginal return of permitting fraud against you is plausibly greater than zero, and therefore, you should welcome greater than zero fraud. You can think of it as a necessary expense, just like rent or salary or advertising is. You can even write it off on your taxes. (Ask your accountant; businesses frequently misunderstand the rules here.) [...]

All fraud is a) an abuse of trust causing b) monetary losses for the defrauded and c) monetary gain for the fraudster. You could zero fraud by never trusting anyone in any circumstance.

Trust, though, is an immensely socially useful technology. Human civilization has a fundamental limitation in that all humans can be trivially killed while sleeping. Huge portions of society’s efforts go toward establishing conditions where this trivial vulnerability virtually never gets exploited.

You can read the whole thing here.

---

🧪🔬 Liberty Labs 🧬 🔭

The price of lithium-ion batteries has declined by 97% since 1991 🔋🔋🔋🔋🪫 📉

Yes, this is a log graph! 🤯

battery with a capacity of one kilowatt-hour that cost $7500 in 1991, was just $181 in 2018. That’s 41 times less.

There’s going to be a bump in the descent around 2022 because of inflation and supply issues, but it’s still quite an accomplishment for humanity, and it has already been world-changing (esp. for mobile electronics, but increasingly EVs).

Without this, the electrification of transportation would be just a dream. 💭

For example:

Tesla Model S 75D which has a 75 kWh battery. In 2018 the battery costs around $13,600; in 1991 it would have been $564,000. More than half a million dollars for a car battery.

This should be rebalanced, though:

Source.

🔌 Power Usage of large AI training clusters ⚡️

Friend-of-the-show Doug (🐴 — that’s supposed to be a Mule, which is an Asimov reference… yes, we’re geeky 🤓) over at Fabricated Knowledge (Embarassing, I somehow mixed up the two! I’ll blame it on too many tabs open) Friend-of-the-show Dylan Patel wrote about Meta’s AI infrastructure and this bit stood out to me:

They commented that a large training cluster could be as much as 6 Megawatts today. They said that these training clusters would be 64 Megawatts in the future. The largest public supercomputers in the world are currently 20 Megawatts to 30 Megawatts. An incredible amount of power will be sucked down for training AI models. The costs to train these models will continue to soar.

And 64MW on the way to what? Where does it flatten out?

Jensen wasn’t kidding when he said that the new unit of computing was the data center…

---

🎨 🎭 Liberty Studio 👩‍🎨 🎥

‘The Emmys since Netflix launched House of Cards’

Via friend-of-the-show Matt Ball (👨‍🚒)

If you’ve read this far, please hit the ‘like’ button! 💙