127: Cloudflare Q1, Nvidia Investor Day 2021, S&P Global & Snowflake, Pipelines & Crowdstrike, mRNA Vaccine Patents, Age Limits on Boards?, and IBM's 2nm Chip
"a perfectly good bird, but it can't fly, so there's no way it's ever going to get to the cloud"
When someone tells you something is wrong, they’re usually right.
When someone tells you how to fix it, they’re usually wrong.
🌞 I almost can't believe it — Today is the first day in 5 weeks without any kids in the house, no remote-school, no constant interruptions... Some quiet to think and do deep work. *inhale* *exhale*
📖 Ok, this edition’s investing section is a bit long, with kinda massive sections on both Nvidia and Cloudflare. I hope it’s not too much… I almost shifted the Nvidia one to Wednesday because it’s less timely anyway, but I figured I’d go with the double-barrel. I’m sure that, like Col. Nathan R. Jessep, you can handle it.
🛀 Anyone else thinks the expression “it’s always in the last place you look” is kind of weird?
I mean, of course it’s in the last place you look. Who keeps looking after they found the thing they were looking for?
🤔 I’ve often heard “Money can’t buy taste”.
I don’t think that’s quite true, or rather, it’s not the real problem in practice.
The way I see it, taste decomposes into the appreciation side and the creation side.
Can you recognize quality/good design/beauty/do you have a good feel for aesthetics? vs. can you create things that fit that profile?
I think the latter is informed by the former.
ie. Some people have great natural ability, but it’s augmented/fueled by studying the great things that others have made.
It’s also possible to have the appreciation part without much horsepower on the creation side. That’s kind of where I fit. I think I have pretty decent taste in multiple fields, but I’m not much of an artist/designer/creator myself.
So that brings me back to the idea at the beginning of this: What gets people in trouble is when they don’t have taste and don’t have this “appreciation” ability to recognize that their taste is crap. That’s when you see true crimes (only magnified by money, because you can afford bigger crimes).
But if you have a solid feel for what’s good, even if you can’t quite create it yourself, then it’s very possible to buy taste — hire people who have good taste (architects, designers, artists, musicians, writers) to do it for you.
So semantics: You may not be able to buy the ability for yourself, but you can rent it from others so that it’s applied to the things you do. And you should.
💚 🥃 I’ve set the goal for this project to have 5% of paid supporters and 95% of free readers. That feels sustainable to me.
Right now we’re at 3.4% of you buying me a coffee or scotch per month and 96.6% reading for free.
If you enjoy this project and want to see it continue, it would mean a lot to me if you supported it. It takes 20 seconds here:
Investing & Business
Cloudflare Q1 Highlights
I still find this company very fun to learn about, so here’s a look at their recent Q1.
A few metrics that jumped out at me:
Revenue of $138 million, up 51% year-over-year
Dollar net retention increased to 123% (+400 bps sequentially — “Since very little of our revenue is usage based, our success with this metric is driven by our success selling our platform to our customers”)
Total customer count 4.1 million (+600,000 free & paying customers in Q1)
119,200 paying customers, +34% year-over-year (“notable acceleration in large customer additions… We ended the quarter with 945 large customers, representing an increase of 70% year-over-year.”)
88% of our contracted customers now use 4 or more Cloudflare products (“our usage data suggests once someone is using that many products, customers consider us a core platform that is very sticky and difficult for any competitor to match”)
RPO in the first quarter $439 million, representing an increase of 14% sequentially and 88% year-over-year
Q2 revenue guidance: +46% to 47%
Guidance FY21: +42% to 43%
Some highlights from the earnings call:
Our strategy with products has always been to get into the market quickly and then relentlessly improve until we are the category leader across each of the features that make up our platform.
They give their bot management product as an example of this:
We launched our advanced Bot Management feature 2 years ago. At the time, there were a number of companies that focused just on this one problem.
We started small with only a handful of initial early adopters, but quickly, using incredible data from our vast platform, we iterated to deliver a product, we believe, is now the leader in the space.
Revenue from this product has grown north of 400% compounding annually. And by the end of this year, we believe we will generate more revenue from just this product than most of the companies focused exclusively on BOT Management.
If this is a repeatable playbook, because of human factors, like the company development culture and talent, and of infrastructure factors, like their Workers platform which allows them to build and deploy new products more rapidly than the competition, it’ll be a powerful engine of growth.
today, products like Cloudflare Workers and Cloudflare Pages are following a similar path to Cloudflare advanced Bot Management. Potential categoric pillars on their own, but with all the benefits for our customers of everything else, our platform is capable of.
Another benefit of having their multi-product platform tightly integrated and built organically is that they avoid this issue:
A Fortune 500 retailer signed a 3-year $1.2 million contract to adopt Cloudflare's platform. They preferred our well-integrated approach to performance and security over cobbling together multiple point vendors or using the Frankenstein like solutions other companies have tried to create through M&A. [...]
one thing that's a little bit different about us than some others is we have been able to develop the broad set of platforms without engaging in very significant M&A over that period of time. And that has a benefit that is not only more efficient, but it also means that our platform is very tightly integrated
A lot of the time as a customer, either you have to cobble together a system from multiple vendors, or you get one vendor, but they themselves have cobbled it together for you via badly integrated M&A, so the problems can be similar.
This isn’t always so bad: It’s possible to pick a bunch of best-of-breed companies that all play well together (and often have partnerships, like Crowdstrike with ZScaler and Okta, for example), but if you can find a vendor that owns fully a best-of-breed integrated platform, that can simplify your life further.
Cloudflare clearly understands this, as they are increasingly bundling everything into packages like Cloudflare One to increase the value proposition through simplicity.
When a quarter goes as well as this one, the natural temptation is to wonder what could go wrong. So I thought I'd share some of the things that I worry about: First, I worry about security. There's been a dramatic uptick in the volume and sophistication of cyber attacks targeting our customers and our company directly. While our team security posture and awareness is world-class, we are vigilant in monitoring and addressing the rise in these sophisticated attacks.
Second, I worry about our team and our culture. At Cloudflare, I believe we did an incredible job transitioning to a remote work environment. But in the months to come, as we transition to whatever is next, I worry there will be an upheaval across industries as employees reassess not just future of work but their future outlook. Finally, I worry about regulatory risk. It seems that tech companies have swung from being able to do-no-wrong to now being able to do-no-right.
Above is Matthew Prince (CEO) with an attitude that I quite like and wish more CEOs would emulate. Celebrate success, but be ever-vigilant.
The U.S. represented 52% of revenue and increased 61% year-over-year.
EMEA represented 26% of revenue and increased 54% year-over-year.
APAC represented 16% of revenue and increased 30% year-over-year.
Looks like APAC is holding them back right now (Baidu issue). But I know they recently opened a new Singapore office (James Allworth is over there now!), so they’re no doubt working on increasing their presence in the region.
The company was built on a very low-touch, self-serve sales approach, very devops-friendly.
They’re not losing that, but as they get ever bigger deals with large enterprise customers, they’re adding the salesforce to service that. Thankfully for them, they are a very attractive place to work:
we disclosed that in 2020, nearly 200,000 applicants applied to work at Cloudflare. About half of those were for sales and marketing roles and the caliber that those people are coming from is really just extraordinary. [...]
We invest behind demand. We invest in a way that ensures that we have a success culture across our sales team. And that has allowed us to continue to move upmarket and sell more to new logos and deliver that.
This last part about “investing behind demand” I’ve already covered in edition #115. It reduces uncertainty and risk a lot.
the way that we see customers is that what they want is a network they can plug into and not have to worry about anything else. So they want a network which is fast, a network which is reliable and a network which is secure. And so there are a lot of different products that map to some of the legacy hardware boxes that are out there. So load balancing or firewall or VPN, those sorts of things. [...]
If you look at our products, it's a little bit like a Mastermind game where we really don't care where a customer starts. [...]
our vision for Cloudflare One, which incorporates the Cloudflare for Teams products and the Cloudflare infrastructure products, that's the direction we think we're headed. And again, what it foundationally is providing is that network that you can plug into and get performance, security and reliability, all in one unified package
On the recent partnership with Nvidia (more on them below):
we want to make Cloudflare Workers the most powerful development platform in the world, not the most powerful edge computing platform in the world, we want it to be the most powerful development platform in the world. And we think of ourselves as competing not with niche edge computing vendors, but with the core computing vendors that are out there
This is ambitious stuff. They’re looking at AWS and Azure, not Fastly. Not necessarily in going head-to-head with similar products, but in ambition, they sure sound like they want to be a massive platform like these hyperscalers.
the regulatory environment around the world is getting much more complicated. We're seeing as different countries are requiring data locality and data residency requirements. [...]
my parents will be happy that the 3 years I spent in law school weren't completely for waste that, that allows us to think about how we can build products to not only help ourselves meet that increasingly complicated environment, but also help our customers.
Analyst: Matthew, I'm sure your parents are proud of you either way, law career or not.
one of the things that we invested in starting back in 2015, which is starting to pay off now, is the flexibility across our platform to not be dependent on any one particular chip vendor. So we can easily migrate not only between Intel and AMD chips, but also to ARM-based chips. And so right now, we have ARM-based chips in production running across our network and processing user requests in a way that's incredibly efficient. That flexibility allows us to, I think, have more options as the world is dealing with some of the shortages of chips and other semiconductors.
Very cool. Looks like they’ve architected their stuff to cross-compile to ARM and run on both instruction sets, probably originally to optimize for cost and future-proof, but it’s now paying dividends because the system is less brittle to supply disruptions.
Back to bundling:
If you look at the companies that get to the next stage, they are taking their broad feature set and combining it into what are effectively site licenses for large customers to deliver a broad platform that solves a large set of needs. [...]
as we compete with companies that just do one thing well, we are good at doing multiple things well, and we're able to use the overall platform to solve not just one need, but many and give our customers an enormous amount of flexibility as we do that. [...]
I think that in the past, there used to be companies focused entirely on something like DDoS mitigation, we've turned that into a feature [...]
so over time, that platform becomes richer and richer, and we're making sure that we have the sales sophistication as we move up the organizational chart, selling multimillion-dollar deals, to be selling what is essentially the entire platform to the CIO or CTO.
Here’s a vision of the future, now let’s build a bridge from here to there, one step at a time.
a lot of companies that in the past were sort of clinging to what we refer to as kind of their Emu boxes. Emu, the flightless bird, which is a perfectly good bird, but it can't fly, so there's no way it's ever going to get to the cloud.
we have a ton of respect for Fastly and Optima and Limelight. But we're in a very different business than they are. We're not usage based, by and large, which last year actually felt kind of hard because use was going up, but we weren't able to automatically bill more revenue. [...]
I actually looked up the data. And Fastly is a -- we just don't see them in deals because, again, I think that they are very much going after the media space and doing media delivery
we have 90% of our R&D organization that's really focused on how do we go deep on any of the products and features that we've launched [...] 10% of our team is really dedicated to an organization we call ETI, which is the emerging technology and incubations team. And their job is to invent the future. [...]
They take lots of shots on goal, not all of them are successful, but they are the team that came up with Cloudflare for Team. They're the team that came up with Cloudflare Workers. And so they are often when we do expand the TAM dramatically, they're the team that is driving that.
Ok, I think that’s a good place to end this one. Hope you enjoyed it 👋
S&P Global + Snowflake ❄️
In edition #124, I wrote about ‘Unfreezing the Data Lake 🥶’ and gave an example I made up about how S&P Global may be able to get incremental value for its quality data by selling it through the emerging ‘data cloud two-sided data marketplaces’.
Funny timing, because 2 days later, this came out:
Financial giant S&P taps Snowflake for better cloud data distribution
What we’re focused on is productizing data — creating new data-driven products, linking all of that together and combining it so that clients can get incremental value. And then also making it available to clients in the way they want to consume it. And that’s what we’ve really done with Snowflake, which is make all our data on the S&P Global Marketplace available through the Snowflake distribution and couple it with Snowflake compute power, so that clients can take advantage of bigger data queries, and all the advantages of compute power, so that they can study and research and analyze and evaluate, not just our data, but our data in combination with their own data. [...]
What Snowflake provides us is a modern addition to our array of distribution, and has additional advantages such as the ability to utilize the compute power as data gets bigger and bigger. Clients want to do new and interesting things by bringing different data sets together, so the ability to access compute power is so important. That has opened up all sorts of new opportunities for us and for our clients in the way we deliver new capabilities, new content, new products, and additional value.
Nvidia Investor Day Highlights
Another one that is always fun to learn about. I read the transcript their investor day and thought I’d share a few highlights with you:
Omniverse, a platform to create and simulate shared virtual 3D worlds. It connects to other worlds using USD, Universal Scene Description, if you will, the HTML of 3D worlds. Omniverse is the foundation platform of our AR and VR strategies.
The HTML of 3D worlds. The PDF of the metaverse. The .doc for CAD monkeys…
We announced DOCA 1.0, the software stack for data center infrastructure computing. DOCA will be for BlueField, what CUDA has been for our GPUs. [...]
So now DOCA 1.0 and BlueField 2 are ready for production. BlueField 3 is right behind its heels. And you could tell how excited I am about this area because very simplistically, a data center is becoming software-defined.
what that means, it takes networking, storage, virtualization and now really, really importantly, cybersecurity, and it puts it in software. [...]
It runs on the CPU. The software-defined data center stack is now overloading the CPU, not to mention it's weird to have the control plane, the security plane agents, all running co-mingled with the application, which could be the intruder.
So the right answer really is to isolate it [...] accelerate the workload so that you could take it off of the CPUs
CUDA for data-center traffic routing/inspecting… Tensorflow for CISCO CCNA nerds?
Man, Jensen’s good at these analogies.
The data center is the new unit of computing. The software-as-a-service trend has caused software to be refactored into disaggregated microservices that can easily scale out and run across the entire data center like it's 1 computer.
I’ve shared this before, but it’s worth re-iterating.
The the next waves of AI are big, the enterprise industrial edge and robotics [...]
Autonomous driving is one of the first mass market robotics applications. It is also one of the most intense machine learning applications that requires decade-long investment. [...]
Effectively, a massive AI computer and a fleet of AV computers that are running continuously in a loop improving continuously. This is the machine learning loop.
Automakers now realize that the car is more than the vehicle, but an installed base, an installed base platform that they own. And if the computer is powerful, it will host valued software services they invent for years to come.
Jeff Fisher on Gaming:
Every person born today will be a gamer. That's 140 million more gamers this year alone. [...]
Why watch a movie, when you can play a movie that's as cinematic and rich as anything produced in Hollywood.
I guess big companies aren’t above checking Google Trends:
Google Search for NVIDIA RTX was up 6x compared to that around the turning launch
The two biggest segments are not doing too badly:
Our gaming business grew 41% year-on-year to record $7.8 billion [...] Gaming revenue has seen a 4-year compounded annual growth rate of 18% [...]
Data center had a record year with revenue increasing 124% year-on-year to $6.7 billion, including almost 70% growth for data center compute [...] 4-year compounded annual growth rate of 69%
our non-GAAP gross margins have increased 640 basis points over the past 4 years, and our non-GAAP operating margins have also increased at even a faster rate, growing from 32.1% in fiscal year 2017 to 40.8% in fiscal year '21 [...]
cash flow from operations growing from $1.7 billion in fiscal year 2017 to $5.8 billion in fiscal year 2021, a 37% compounded annual growth rate
They also mentioned that they’re still supply constrained, but that supply should ramp up over the year.
Jensen: our primary goal is to pioneer accelerated computing. Well, before you could do that, you really have to build a whole stack because the entire way of computing is refactored. We refactored computing from the application to the algorithms, to the solvers, to the system software, all the way down to the silicon, as you guys know well. [...]
our strategy, is to develop it in 3 layers, which is the 3 layers of computing. It's the hardware layer, the system software layer [...] and the third layer, of course, is applications
This is clearly the best — though not the easiest — way to do things, if you can swing it. That’s why Nvidia has more software than hardware engineers even though most people think of it as a semiconductor, hardware company.
The best way to sell the most chips is if you can create the conditions in which those chips are most useful.
Sure, you can take the modular approach and hope that after you make great chips available, others will develop the other components in the stack, but that leaves a lot to others (who have their own — different — ideas on how things should go), and you have the chicken & egg problem.
If you can develop the whole thing in parallel and release a fully functioning, immediately useful stack of stuff to customers, it’s a lot easier to get adoption and get the snowball rolling downhill (I feel like this metaphor needs to come back — it was everywhere back when you saw the name Alice Schroeder mentioned weekly).
They may give away some parts of the stack because that’s what leads to the best overall results, but the value comes from all of it, not just the part with a price tag on it.
(kind of like how Apple is also largely a software company, but they mostly monetize it through selling hardware — just ask Apple customers if they’d rather run Apple software on a Microsoft Surface laptop or run Windows on a Macbook — there’s a lot of that going on at Nvidia too)
each one of the higher layers increases NVIDIA's opportunity by an order of magnitude. Let me give you an example. There are 100 million cars sold a year. And that's the entire opportunity annually for chips, 100 million cars. However, those cars are driven 10 trillion miles. [...]
you sell a chip once, that's the economics of selling chips. [...]
then on top of that, it's really skills and tasks, skills -- AI skills that perform tasks. And they tend to be per end user or per task or per instance. And I could imagine a day where an AI is paid by the hour, just like a particular skill, someone who performs a particular task is paid by the hour, the AI, of course, will be paid by the hour. And so that's kind of the economic funnel of our company. And my expectation is that over time, the layer that's on top will be the largest of all.
But it's built on top of the next 2 layers. It's enabled by the next 2 layers. [...]
our economics are just so much broader -- so much richer and so much larger when we have more people around the world in the ecosystem supporting our architecture.
Ok, this is important stuff, and I think not broadly understood.
Nvidia isn’t making computer chips. That’s too small an ambition for them. They just care about solving problems through computing, and they’ll monetize various parts of the stack that they own to varying degrees depending on the use case and maturity of the vertical.
At first you gotta kickstart things, so you give away a lot of tools so people can build on your platform. But the dynamic can change over time…
all of our stacks are cloud native. Everything from Omniverse to GeForce Now to Clara, to Jarvis, to Metropolis, and they're used by the cloud service providers.
Jensen on an aspect of 5G that I don’t see mentioned enough, “private 5G”:
consumer 5G is really fantastic, and it will incrementally and over time, increase the capacity of broadband and very importantly, continue to drive down cost of broadband because what goes along with a lot of capacity, is the decrease in cost. And so I think consumer commercial 5G is really fantastic.
But what's brand-new is private 5G, all these spectrums that companies could license to operate a wireless network that is secure and private in farms and entire factories. And as you know, these factories are gigantic.
The vast majority of the Earth’s livable space is covered by farms and infrastructure. And finally, we have the networking available to deliver computation services out to the edge. And the computation service that I think is most exciting is sensors connected to AI applications, AI skills, they're monitoring and predicting, making sense of the world. And 5G makes that possible.
That’s it folks. Good night, and good luck!
My Thoughts on Age Limits for Company Boards
In my opinion, age limits are way too blunt an instrument, and assume that everybody is the same (when they clearly are not). It’s simply not a good proxy for the thing you want to measure.
Some people are sharp as ever to very advanced ages, and others sadly see their health and capacities diminish rapidly at relatively young ages. Even the averages there have changed over time, and are likely to change even faster in the coming years as we make progress on curing the diseases of aging (ie. the average 70yo board member in 1960 clearly wasn’t the average 70yo board member in 2020, and you can expect big changes to the average health of a 70yo board member in 2050).
If the goal is to avoid senile or incapacitated board members, then that should be stated directly, and some mechanisms to protect against those conditions should be put in place (if it's too delicate to do directly, offload it to some neutral third party with the relevant medical expertise).
And if the goal is to have some rotation of board members, and not have them stick around forever, then age is also a bad proxy (and as I wrote in edition #125, there’s likely some ageism discrimination going on too).
You could have someone who joined the board when they were 30 and still be there at 55 (25-year tenure) and nobody raises an eyebrow, while someone who joined at 70 could be pushed out at 75 because “sorry, we need fresh blood”. ¯\_(ツ)_/¯
mRNA Vaccine Supply not Constrained by IP
“There is no idle mRNA manufacturing capacity in the world. This is a new technology, you cannot go hire people who know how to make mRNA — those people don’t exist,” [said Stéphane Bancel, Moderna CEO].
He added that the patent waiver “will not help supply more mRNA vaccines to the world any faster in 2021 and 2022”, the most crucial period of the pandemic.
“If you were to start today, you’re going to have to start by hiring people. Those vaccines don’t fall from the sky,” Bancel told the FT US Pharma and Biotech Summit later on Thursday. “There is no mRNA industry . . . When we hire people that come from traditional pharma, we have to train them in the art of mRNA.” (Source)
h/t Byrne Hobart
Science & Technology
IBM’s Shrinkage (and not the share count…)
Ok, that’s a good intro line to an article:
Every decade is the decade that tests the limits of Moore’s Law, and this decade is no different.
This is of course about IBM announcing that they created “the world’s first 2nm chip”.
Seems like IBM is pretty good at bragging rights and PR stunts, but not exactly the best at shipping top-notch products that use these technologies, but that’s a different story…
Just to clarify here, while the process node is being called ‘2 nanometer’, nothing about transistor dimensions resembles a traditional expectation of what 2nm might be. In the past, the dimension used to be an equivalent metric for 2D feature size on the chip, such as 90nm, 65nm, and 40nm. However with the advent of 3D transistor design with FinFETs and others, the process node name is now an interpretation of an ‘equivalent 2D transistor’ design. [...]
Today’s announcement states that IBM’s 2nm development will improve performance by 45% at the same power, or 75% energy at the same performance, compared to modern 7nm processors. IBM is keen to point out that it was the first research institution to demonstrate 7nm in 2015 and 5nm in 2017
Gotta love Anandtech for actually asking IBM what they meant with their marketing-speak:
IBM states that the technology can fit ’50 billion transistors onto a chip the size of a fingernail’. We reached out to IBM to ask for clarification on what the size of a fingernail was, given that internally we were coming up with numbers from 50 square millimeters to 250 square millimeters. IBM’s press relations stated that a fingernail in this context is 150 square millimeters. That puts IBM’s transistor density at 333 million transistors per square millimeter (MTr/mm2).
Shoulda used Crowdstrike…
Colonial Pipeline fell victim to a cybersecurity attack on Friday that involved ransomware, forcing it to temporarily shut down all pipeline operations [...]
Colonial Pipeline transports nearly half of the East Coast’s fuel supply [...] Colonial’s system spans over 5,500 miles between Texas and New Jersey, connecting refineries on the Gulf Coast to more than 50 million people in the southern and eastern U.S.
The pipeline transports gasoline, diesel, home heating oil and jet fuel. It also supplies the military. (Source)
The Arts & History
Still slowly making my way through a rewatch of 'Mad Men'.
Got to season 4's 'The Suitcase'. What a great, great episode, nested in a run of other great episodes. What a show.
Is anything of this quality airing today? Can we even know in real-time, or do we need the benefit of some time and perspective to know when something is truly special?
There’s plenty of shows that had 1 or 2 very good seasons in recent years, but keeping quality extremely high for long runs like Mad Men, Sopranos, Breaking Bad, Six Feet Under, The Wire, etc, is a very difficult filter to get through for greatness.
Some examples of shows where I really liked 1-2 seasons, but stopped watching because I felt the quality dropped off:
Killing Eve (S1 is amazing, but then Phoebe Waller-Bridge left)
Sneaky Pete (really enjoyed S1, dropped S2 early, just didn’t feel right)
Patriot (S1 is amazing — I still need to really try S2)
Justified (S1-2 were great)
Masters of Sex (S1 was great — I think the title kept it from being as popular as it should’ve been — but S2 lost the magic)
Marvellous Miss Maisel (S1 was great — S2 lost me a few eps in)
The Americans (S1-2 were lots of fun)
Fargo (S1-3 were really good, I think S1 was the most evenly good. I haven’t tried S4 but I hear bad things. No first-hand knowledge, though)
Homeland (S1 was extremely good, S2 very un-even but still had some really good eps, S3 was more mediocre and I dropped after it)
The Expanse (S1 was amazing, I thought it was the next Battlestar, then somewhere in S2 the writing changed and many characters just weren’t the same people anymore and the writing was weaker and quality more uneven, dropped in S3)
Dexter (S1-2 were great, then it was still ok for a while but more uneven, and I felt the quality progressively dropped — I bailed early in S5, I think)
I know many will disagree with me on many of these, and that’s fine. Different people have different tastes, are looking for different things in their entertainment/art, are sensitive to different aspects of filmmaking, etc.
I may even disagree with myself if I was to re-watch some of those — it’s certainly happened to me before that I saw something when I wasn’t in the right mood to enjoy it, and then came back later to it and liked it a lot more.
That being said, these are my current views on these ¯\_(ツ)_/¯