We're entering the Fourth Industrial Revolution: The Intelligence Revolution

The First Industrial Revolution (circa 1760-1840): Known for the rise of mechanized manufacturing, and the rise of new technologies like the steam engine and power loom

The Second Industrial Revolution (circa 1870-1914): The Technological Revolution. This period was characterized by the advent of electricity and mass production processes. Railways and the invention of the telegraph and telephone were key developments

The Third Industrial Revolution (circa 1960-early 2000's): The Digital Revolution. This period introduced the shift from analog and mechanical technologies to digital electronics. Computers, the internet and proliferation of digital communication were key developments

The Fourth Industrial Revolution (~2017 - Present): The Intelligence Revolution. Foundation models and AI are just starting to drive significant change, and we're only a couple years in. I expect huge developments in life sciences, robotics, software, and many other markets

But what needs to happen to enable the Intelligence Revolution? We need significant infrastructure buildout of data centers. Data centers are the enabling piece, the foundational layer. In every industrial revolution we had an infrastructure buildout followed by a dramatic increase in revenue / production. Much has been written on one component of the data center - the GPUs / compute. In particular, the spend on GPUs vs the relative revenue generated from AI Intelligence Systems. Right now we are in the early days - the infrastructure buildout phase of this industrial revolution. The revenue / production phase will follow

In my post below (end of thread) I explore the 11 core components of a data center

1. Servers
2. Networking
3. Storage
4. Cooling
5. Power Infrastructure + Power Utility
6. Security
7. Management and Monitoring
8. Racks and Cabinets
9. Cabling
10. Environmental Controls
11. Software 1. Servers. Servers are the backbone of data centers, providing the computational power necessary for processing and managing data. They house crucial components such as CPUs, GPUs, memory, and storage.

After Q3 '22 Snowflake guided to 47% growth in 2023. Yesterday they dropped that guide to 40%. What was the cause? 3 things primarily (in ascending order of impact)

1. Change in existing customer purchasing behavior
2. Weaker new logo bookings
3. Slower ramp from newer customers 1. Change in existing customer purchasing behavior

Snowflake sells annual / multi year deals that come with minimum consumption spend. Once customers reach those committed spend levels, they have the choice to re-up with another multi-year commitment. Usually, they do...

Some takeaways from the Datadog earnings call:

Positives
- Strongest new logo quarter
- Gross churn very low
- Healthy pipeline generation

Negatives
- More optimizations in Q4 than Q2 / Q3
- Things got worse in December
- Weak 2023 guide

$DDOG The best summary quote from the earnings call was "Overall, we observed slower usage growth with existing customers while continuing to scale our new logo acquisition and new product cross-sells."

Investing in infrastructure businesses, particularly data infrastructure, has been incredibly exciting over the last few years (and will continue to be!). A couple data infra stacks Im excited about

1) Analytics stack
2) Real time stack
3) Data Lake stack
4) AI stack

More below 1. The Analytics Stack (or Modern Data Stack). Typically consists of:
1) Data Source (transactional database, SaaS app, etc)
2) ELT tool (Fivetran, Airbyte, etc)
3) Data Warehouse (Snowflake, Redshift, etc)
4) Data Transformation tool (dbt)
5) BI (Looker, Sigma, etc)

Consumption models under fire this year given easier to optimize (reduce) spend quickly. This pain happens quick.

However, everything I'm hearing pointing to new customer signups posing the biggest challenges currently, and expansion / churn seeing relatively less pressure... ...Models more reliant on new customer signups for growth feeling more pain now, vs consumption models felt it earlier in the year.

What do I think of this? 2023 might be the year consumption models actually show more resiliency (less growth decay)...

Read through the Crowdstrike transcript. Some takeaways re macro. Let's start with the positives

1. Gross & net retention not deteriorating:

"Our dollar-based net retention rate was well above Q3 of last year and consistent with our Q2 performance, which was at the... ... highest level in seven quarters. Our best-in-class gross retention rates remained at record levels above 98%"

2. Enterprise sales cycles staying constant:

"In the enterprise, sales cycles or average days to close remain consistent with last quarter's modestly higher level."

Lots of interesting announcements from AWS at re:Invent today. Some new products, and some product enhancements. Data and Security are clearly top of mind for them. Some of my favorites below: AWS Security Lake: A central data lake purpose built to store and analyze security data (primarily log data)

Link: aws.amazon.com/blogs/aws/prev…

In early / mid 2012 Salesforce was growing ~35% with ~20% FCF margins.

What revenue multiple would it get today? At the time it traded ~7x NTM revenue. Atlassian is the closest comp today (growing rev 36% w/27% FCF margin), and trading at ~19x NTM rev. Atlassian is an outlier today in terms of their multiple solely relative to their growth, but of course they have the FCF support + a super sticky product like Salesforce.

There isn't another obvious comp today. ServiceNow maybe at 11x for 28% growth and 30% FCF margins

There's broadly two opinions out there on inflation:

1) Inflation is "transient-ish" and will wane the back half of this year and be back to 2% target early next year

2) De-globalization will drive sustained high inflation over the coming 3-5+ years

What do I think? As a growth investor it's imperative to have a point of view. Inflation impacts rates, and rates impact software multiples. Ive largely been on the transient-ish side, but it's important to see the other point of view

And by transient-ish I mean inflation lasting ~12 more months

Post GFC Venture Capital as an industry has massively "over earned." The data below shows pooled IRRs by vintage from Hamilton Lane. And this performance is despite the fact that private multiples have historically always been higher than public multiples

https://twitter.com/buccocapital/status/1505250453130461184

As a venture capitalist we take bets that companies will be worth more in the future, despite a "premium" multiple paid in the early days.

So yes, many private companies are "worth less" right after an investment if you applied a public multiple. But this phenomenon isn't new

What should a public software company growing 100%+ at $250M+ ARR be worth (multiple)? There are very few companies to hit this. Im excluding companies that hit that growth only bc of covid. I believe that list includes Snowflake, SentinelOne, Zoom, Crowdstrike, Shopify, Workday Expanding to 80%+ and I get a few more co's like Datadog, Monday, Elastic, UiPath, ZoomInfo, ServiceNow, Okta, Zendesk (not an exhaustive list but representative).

The big question - theres high growth software and then there's hypergrowth. What multiple should hypergrowth get?

Reflecting on digital transformations vs pull forward for cloud software over the long weekend - I think we've largely seen 3 different ways cloud software was affected by Covid:

1) Fake TAM creation
2) One time pull forward
3) Durable pull forward I don't think any business was truly unaffected. Therefore, it's important to have a perspective on which bucket each company falls into when predicting what future growth will look like. This applies to both public and private businesses

Bifurcation of software multiples as seen on a scatter plot (multiple vs growth). Historically all software has traded within red circle. Now a separate group has broken out. Is it warranted? After Q4 earnings, I believe we'll see more separation of contenders from pretenders Here's the scatter plot without the circles

Some takeaways from Morgan Stanley's Q4 CIO survey

- Software has the highest growth expectations in IT
- Strong demand in software persisting (not simply pull forward in 2021)
- Cloud computing remains CIO's top priorities
- Security software most defensible

More graphs below "Survey data suggests 25% of application workloads are running in the public cloud today, up from 23%... in 2Q21. The multi-year trend in the migration of applications to the cloud remains intact, with CIOs expecting 44% of workloads to reside in public cloud by 2024"

Trend in cloud software: Growth Durability 📈📈

Historically, many people have overestimated how quickly software companies would decelerate growth. below chart shows rough Datadog quarterly growth expectations at IPO (orange) and actual (blue) Datadog has grown much faster (by a wide margin) than they were expected to at IPO

The biggest effect of this? Compounding $$ growth

At IPO (Sept '19) Datadog was projected to do ~$610M of revenue in 2021. Current projections? $1.3B!

In just 2 years expectations have doubled🤯

Some confusion after hours in Snowflake. For some extra commentary on the Snowflake RPO figure (lots of talk about it this quarter), I'd point people to the pre-earnings research note Karl Keirstead at UBS published a few days ago. Screenshot below: $SNOW To try and explain this further - RPO measures the aggregate total unrecognized contract value from all customers. Let's say a company had one customer that signed a $100k one year deal. After 6 months, the RPO would be $50k. this represents the remainder of the deal, or 6 months

Asana and Monday are two fascinating businesses. I remember looking at private rounds many years ago and thinking "this is a commodity space, there won't be any big outcomes with a long tail of small / medium outcomes."

Both are ~$15B companies today. Software markets are huge! Both have executed extremely well and this shouldn't be understated. What's been incredibly impressive is the growth in the enterprise segments. Both define enterprise as customers with >$50k ACV. Asana grew that segment 92%. Monday grew it 226%. Way more than just a SMB business

The acceleration across cloud software companies so far in Q2 has been very impressive. As of now, most companies with a June quarter end have reported Q2. Most accelerated. Graph below shows (Q2 YoY growth - Q1 YoY growth). Positive numbers represent acceleration (faster growth) I removed 3 companies - Shopify, Olo and BigCommerce. All were major Covid beneficiaries. Q1 this year lapped a 2020 Q1 that did not see much Covid benefit, while Q2 lapped a 2020 Q2 which did see a Covid benefit. So the YoY compare between Q2'21 to Q2'21 isn't as relevant

Some great slides from the Crowdstrike investor briefing last week. A few call outs on what makes them such a great business, that others should aspire to:

1. TAM Expansion: The best businesses increase the size of the pie, not just their piece 2. Product Expansion Velocity: At IPO (in 2019) Crowdstrike had 10 modules. They now have 19. Amazing product development velocity

I always enjoy reading Bessemer's annual State of the Cloud report. One of my favorite slides below. The takeaway? Leading cloud companies don't decelerate growth nearly as quickly as they're expected to. Why? Cloud markets are almost always much bigger than anticipated Analysts constantly underestimate leading cloud companies ability to sustain higher growth rates for longer periods of time

A trend I'm excited for this year: DataOps & the Analytical Engineer

~10 years ago DevOps was born. The role of system admins and developers merged. Infrastructure became self-serve

Today the role of data engineers and business analysts are merging. Data is becoming self-serve Data infrastructure is becoming so powerful that the tools today allow non-technical folks to carry out the once complicated / custom code/ huge backlog jobs of data engineers.

Before getting into what this means, let's first discuss how we got here