Dear partners,
You should’ve received log-in credentials to your investor portal by now. If not, please let me know. Also note that you can find the quarterly performance at the end of this letter, in the performance appendix – though your actual returns may differ slightly depending on the timing of your subscription.
As this is the first quarterly letter, I want to provide the vision for the fund. I’ll also outline three aspects of our process and then end with an example of an investment that fits our criteria.
Vision
My goal is that Infuse would be the best-performing public equity fund over the next 50 years. I realize that sounds crazy but that’s the goal – long-term, world-class returns, not absolute AUM. To give us the best odds of accomplishing that goal, there are a few things that might be helpful.
To invert the premise, we can’t have the best-performing fund over 50 years if it doesn’t survive 50 years. One of the main ways that funds dissolve is through excessive leverage. In service of returns, leverage ends up destroying them. To me, the risk simply outweighs the reward, especially since world-class returns are typically derived from situations where there is a lack of agreement about a company. In these cases, stock prices can be wildly volatile. Adding leverage to that cocktail is a recipe for, in the best case, stress, and in the worst case, failure.
However, just making the claim that you want to be the best isn’t worth a whole lot without a strategy to make it a reality. To this end, our plan is to hold a concentrated portfolio of the fastest-growing, highest-quality companies in the world at the best valuations we can find.
Stock prices are a function of two things – earnings and other investors’ perceptions of those earnings. So in an ideal world, we’d buy stocks where earnings increase dramatically while the perception is that the company won’t be able to continue its success. However, perception can be quite fickle. Betting on how other people will view the nature of a company’s earnings in 5 or 10 years is not an easy task. That’s why we primarily focus on the velocity and durability of the earnings growth and then make sure we’re not overpaying.
While our goal is world-class returns, that’s really an output of good decisions. Focusing too much on explicit performance goals can create unintended consequences. As Theodore Roosevelt liked to say, “comparison is the thief of joy.” There’s nothing wrong with wanting to be the best, – however, there is something wrong when it takes your focus off the inputs to the process. Being consumed with the highest returns can also lead to decisions that make the process more fragile – adding leverage being an excellent example.
Therefore, with the long-term goal of world-class returns in mind, we seek to focus on the inputs to the process. After all, that’s what we have control over. Namely, we have control over our specific standards of growth, quality, and valuation. And when we focus, every day, on trying to raise those standards, the outcome will take care of itself.
Process Details
So that’s the vision. But, I also wanted to dive into a few more details on our process.
At the end of the day, we all just want high returns with a low probability of losing money. But there is tension. If you want to eliminate the potential of losing any money, then the upside will be quite low. Conversely, if you want the highest upside possible, there likely needs to be a non-zero chance of losing your entire investment. On one end of the spectrum is US treasuries. And on the other may be a seed investment in a biotech start-up. However, I do think there are some ways to increase the upside while decreasing the downside.
First off, we think concentration is important for growing your money. Extreme diversification can be good for wealth preservation but concentration is key for supernormal returns. You can look at nearly every example of the wealthiest people in the world – they pretty much had zero diversification, just owning a large stake in their business. Another way to look at it is if you were buying a collection of entire businesses – how many would you need to feel reasonably diversified? When you think of investing as buying actual businesses, I think the natural tendency is to concentrate. This seems likely because it takes quite a bit of time to deeply understand a single business, let alone 50 or 100.
We are bigger fans of what Stan Druckenmiller says about putting all of your eggs in one basket and watching it very closely. There is something very powerful about focus that sharpens decision-making. You can certainly be more intellectually lazy about a 1% position than a 20% position. Of course, humility is needed – I will be wrong. So there is certainly a middle ground – more than 4 or 5 but less than, in my opinion, 20.
Second, we as humans love putting things into categories. It helps our brain cope with complexities and makes our neocortex feel warm and fuzzy. As an example, labels like large cap and broad sector designations like tech are abstractions. What really matters is high upside and a low probability of losing money (high reward, low risk). So would you say a portfolio entirely consisting of software stocks is “risky”? At face value, it seems very risky. But why? Because sectors tend to trade similarly due to non-fundamental reasons like “fund flows” or technicals? Or because your portfolio’s Sharpe ratio will decrease from the added volatility? Or because it is accepted wisdom? It’s all an abstraction that misses the point – high reward, low risk. Now, of course, there are counter-arguments. In fact, there almost always are because reality is incredibly complex. Using the software example, there are real risks like the shortage of highly skilled programmers, missing a new paradigm, a worldwide cyber virus, low chip production which would affect overall cloud growth, or relatively low barriers to entry. Further, if your entire portfolio consisted of marketing software, that would be incredibly risky because competitors may compete on price, creating worse economics for the entire industry. And they may be competing for the same talent. But in general, labels like software and tech are typically lazy thinking. Just because the businesses have a code base doesn’t necessarily mean they should be lumped together.
Third, volatility is not the same as risk. We’ve defined risk as the probability of losing money. And volatility is the magnitude of price swings in a stock. If a stock crashes after a good earnings report, the volatility will have increased but the risk may be lower considering the shorter payback period. Instead of Wall Street’s love of a magic formula (risk = volatility), one legitimate reason investors equate the two is that high volatility usually means investors starkly disagree on the company’s future. If the business was highly predictable and barely grew, it would be much easier to value the company, and therefore, the stock price wouldn’t be as volatile. But a lack of agreement signals opportunity. So volatile stocks can actually be a ripe hunting ground because there is a greater probability of mispricing.
We’re talking a lot about the nature of risk because that’s at the core of investing. Put another way, upside is just good risk and downside is bad risk. Essentially, there is a wide spectrum of things that could happen, but gain is when good outcomes materialize and loss is when the bad ones do. Our goal is to tip the scales as much as is in our power to increase the probabilities of good and decrease the probabilities of bad. And once again, the outcomes will take care of themselves.
All in all, we tend to view these three aspects of investing – diversification, sectors, and volatility – a bit differently and we expect these views to contribute positively to our long-term returns.
Real Example
But enough theory, let’s put it all together and get to the good stuff – business analysis! One company that fits our criteria is Snowflake.
The company was founded in 2012 by two former Oracle database engineers – Benoit Dageville and Thierry Cruanes. The pair became friends shortly after Benoit interviewed Thierry. In fact, a colleague looped Benoit in because Thierry’s french accent was too thick to understand. The two hit it off since they had both done their PhDs at the same university in France. After years of working at Oracle (16 for Benoit and nearly 13 for Thierry), they became increasingly frustrated that Oracle was ignoring the cloud. So the pair jumped ship when Mike Spieser of Sutter Hill Ventures reached out to them about a new database idea.
Sutter Hill gave them $1 million to get started and the founders worked in stealth mode, with Spieser as interim CEO, for the first two years. In 2013, they hired Marcin Żukowski as a third co-founder since he created a way to make databases more efficient called vectorized query execution. All three technical co-founders had their PhDs and were the world’s leading database authorities. What’s more is that all three continue to work at Snowflake, rigorously improving the product (Benoit as President of Products, Thierry as CTO and Marcin as a VP of Engineering).
Once the company had its first product in 2014, Spieser stepped down and recruited a Microsoft executive named Bob Muglia, who ran the business until passing it off to current CEO, Frank Slootman. Slootman is arguably one of the most successful executives in enterprise software, having taken two companies public, Data Domain and ServiceNow (he grew Data Domain from barely any revenue to $600 million and helped ServiceNow go from $75 million to $1.5 billion in sales).
When Slootman joined, he brought in his long-time business partner and the CFO of ServiceNow, Mike Scarpelli. And to round off the core C-suite, Chief Revenue Officer, Chris Degnan, was employee #16 and the very first sales rep. So despite the fact that Slootman isn’t a founder, all three founders are still actively working on the product and Slootman definitely has a founder mentality.
Now let’s run the company through our three big criteria: growth, quality and valuation.
Growth
Snowflake, the product, is a consumption-based software that helps enterprises analyze their data. Customers typically buy credits for at least a year and then draw them down as they run various jobs. The big differentiator is that Snowflake was the first data warehouse born in the cloud (the name “Snowflake” even alludes to that fact). This allowed the founders to separate storage and compute, lowering costs dramatically. Traditionally, data warehouses like Teradata would charge customers for storage as well whereas Snowflake just stores data in the cloud and then only charges when computing resources are used. This wasn’t even possible before Snowflake.
Because of this innovation and solving a real customer pain point, the company has done over $1.6 billion in revenue over the past 12 months, growing nearly 90% year-over-year. By the end of this year, it will likely do well over $2 billion in sales which would represent a 4-year 114% CAGR. Over this period, the company’s average expansion rate was 175%, meaning that customers tended to spend 75% more on the product than in the previous year. Imagine for a moment, the power of a business where revenue grows 75% without adding a single new customer. That’s why the company has been singularly focused on scaling the business. Despite this rapid growth, the business still managed to throw off $300 million in free cash flow, though if you subtract stock-based comp, it is still negative. That doesn’t worry me though because the TTM incremental FCF margins after SBC (a mouthful to be sure!) are around 30%. I have no qualms about the underlying unit economics, especially when you recall the best-in-class expansion rate. It would actually be an irrational decision to not grow as fast when customers expand their usage that rapidly. It’s like refusing to pull a lever that spits out more money tomorrow.
Further, there is certainly no shortage of reinvestment opportunities. While management estimates the entire market opportunity at ~ $250 billion, up from $80 billion in the S-1, whatever the true numbers are, it’s clear the opportunity is huge. Data will become increasingly important for making business decisions and Snowflake is one of the primary engines powering this forward. And Snowflake isn’t just satisfied with analytics. The most recent product announcements from Snowflake Summit included Unistore which offers customers transactional capabilities.
In the world of databases, there are two main types – OLTP (online transactional processing) and OLAP (online analytical processing). Until recently, the company was solely used for OLAP workloads, like being the data source for powering a Tableau visualization. Now, the company is slowly moving into OLTP, where Oracle dominates. An OLTP database needs super low latency because it powers transactions – like a consumer buying on Amazon. Snowflake doesn’t need to be the fastest because the majority of transactions don’t need to happen in milliseconds but enabling customers to do analytics directly alongside their transactional data will provide quite a boost to efficiency. Moreover, to frame the market size, Oracle does more than $42 billion with 30% free cash flow margins. Snowflake doesn’t pretend like it’s directly competing with Oracle right now – because it’s not. But we are seeing the first green shoots of Snowflake’s grand vision.
As part of that vision, Snowflake’s next goal is to disrupt application development. This is still early but the company’s Native Application Framework will enable customers to develop applications right from Snowflake. Using the recent acquisition of Streamlit for Python development and Snowpark’s APIs, developers don’t need to keep a separate copy of customer data. Rather, they can just access data straight from the source through data sharing. This reduces costs and speeds up development. This is certainly something to keep an eye on and I suspect more and more applications to be built directly on top of Snowflake.
The company estimates it will bring in $10 billion in product revenue by fiscal year 2029 and will still be growing at least 30% at that point. I suspect they can actually reach the targets a little earlier but management’s estimates imply about a 37% revenue CAGR. If the company is still growing 30%, it’s possible that it would reach $18 billion in overall revenue in 8 years (assuming services still make up about 6% of revenue). Now, so much can change in 8 years and we’ll leave the assumptions for the valuation section, but it’s clear the growth opportunity is quite substantial.
Quality
The background of the company was important because the core group is absolutely top-notch. The three technical founders are still contributing mightily and Frank Slootman and team are maniacally focused on executing. It’s the perfect pairing of innovation and efficiency. Since the beginning, the founders wanted to stay out of the limelight because running a business wasn’t their strength and all they wanted to do was code. As evidence, Benoit has more than 100 patents to his name and Thierry has upwards of 50.
This team keeps pushing the boundaries. For example, the company launched a data-sharing marketplace in 2019 which has really strengthened the network effect. As customer data already lives in Snowflake, why not easily monetize it by sharing it with other Snowflake customers? Or beyond monetization, imagine how useful it would be to have a constantly updating dashboard of sales data from Walmart if you run marketing at Proctor and Gamble. Rather than sending gigantic CSV files and doing ETL (extract, transform, and load is a way of uploading data into a database that is often fairly cumbersome), you can access the data you need immediately. And in our example, Walmart can granularly control who has access so there are no data governance problems. As more customers use data sharing, the usefulness of the data marketplace increases which attracts more customers and the cycle reinforces – a classic network effect. And this also makes the entire ecosystem stickier. Data sharing is so important, in fact, that part of management’s comp plan is tied to stable edges. The company defines a stable edge as a customer that engages with the data sharing feature at least 20 times over a rolling 6 week period. And to give numbers around this, stable edges are growing more than 100% and about 20% of Snowflake’s customers maintain at least one. As more customers use Snowflake’s data sharing, the moat deepens, churn is reduced and expansion rates go up.
But if this doesn’t translate into the numbers, it’s not very useful – so here are the last three full, fiscal years of different bottom-line metrics.
GAAP operating income
2020: -135%
2021: -92%
2022: -59%
Non-GAAP operating income
2020: -106%
2021: -39%
2022: -3%
Free cash flow
2020: -73%
2021: -14%
2022: 8%
While the company is still aggressively reinvesting into the business, we are seeing clear signs of operating leverage. Over the same period, gross margins have also improved from 52% of revenue to 62% as larger scale improves the company’s negotiating power with the cloud providers, namely AWS.
You may notice the difference in GAAP and non-GAAP operating income and that is mostly a function of stock-based compensation. As part of recruiting Slootman and Scarpelli, along with options vesting for early employees, the company has had to work through quite a bit of dilution. The good news is that, according to management, dilution is running under 1% from here on out.
The company has also set targets for free cash flow, along with its $10 billion in product revenue estimate. Initially, the target was 15% but after a year, the goal posts were raised to 25%. Since Snowflake is already at 19%, 15% wasn’t very believable anymore. I imagine we could see some more improvement, especially as their negotiating power becomes stronger as they drive more volume to the cloud titans.
So there you have it, we have one of the most established teams in the history of enterprise software saying that they will grow at nearly a 40% CAGR for the next 7 years with 25% free cash flow margins, with an ever-growing network effect. That seems like quality to me.
Valuation
We’ve already done most of the work by looking at the growth opportunity and the quality of the company. Now we just have to put it all together.
If we use the $18 billion estimate and 25% free cash flow margins (FCF), that’s $4.5 billion in free cash flow. ServiceNow or Salesforce would probably be a good comparison at that scale. ServiceNow does $2 billion in TTM FCF and trades for 35x and Salesforce does $5.7 billion in FCF at 25x. If we use a simple midpoint, let’s assume Snowflake trades for ~30x FCF. If we tack on 1% dilution, that gets us a triple over 8 years, good for a 14% CAGR. Now, that’s not necessarily something to write home about but those are the assumptions that management has given. The fact that FCF margins were raised from 15% to 25% after one year indicates that there is room for moving up the estimates.
In a slightly more bullish scenario, I think the company could reach $11 billion in revenue in 5 years with 30% FCF margins and a 40x FCF multiple. That scenario, even after dilution, yields a 21% CAGR. I don’t think it’s particularly prudent to underwrite anything more aggressive than that, but I certainly don’t think it’s impossible, given Snowflake has a much larger TAM than ServiceNow and Slootman, himself, saw more potential. Frankly, I don’t see why this company can’t be valued in the hundreds of billions in the long run. But valuation is certainly important. A lower valuation speeds up our payback period and a higher one, lengthens it. There’s no way around that. At the same time, being too dogmatic with assumptions is a sure way to miss out on big winners. The very best companies tend to outperform high expectations. But high expectations require high levels of performance. In other words, if Snowflake doesn’t bust some base rates, our CAGR likely won’t be above 20%. On the other hand, based on the growth and quality assessments, Snowflake is not just a normal company.
When we truly think like business owners, prices going down only speeds up our payback period. For example, if Snowflake’s price dropped 20% from our buy price, the new expected CAGR from management’s targets would be 19%. So when you see prices go down, just remember that the forward-return expectations are now higher. You’d think the very same thing if you were interested in buying a local business. Let’s say you wanted to buy the laundromat down the street and finally put in debit card machines so people didn’t have to lug around quarters. If you saw that the owner kept lowering the price, you’d be more interested because you’d get paid back faster on your initial investment. Stocks are no different. In fact, they are usually an extreme example because the immediate liquidity can drive booms and busts. Ideally, we’d be able to take advantage of these extremes; however, by definition, extremes don’t happen very often so most of the time, we will just be out hunting for lower payback periods without compromising on our growth and quality standards. In times where we simply can’t make the payback scenarios work, we’re happy to continue to study companies but we may take a more conservative stance. To be clear, we don’t believe the current environment warrants conservatism. There are plenty of companies that fit our criteria with reasonable payback periods. We are actively deploying capital and we are quite excited by some of the prospective forward returns. At the same time, we think it is prudent to focus on companies with rock-solid balance sheets, huge customer value propositions, low employee attrition, and high profitability. But I might point out – it’s always a good time to focus on those characteristics.
Closing
I’m honored to have you as a partner. Thank you for your trust and support. It enables me to think long-term and will be our own competitive advantage.
The economy, like life, will have its ups and downs. All we can do is focus on what we can control and work hard to continually raise our standards. Our strategy is simple – hitch a ride to the world’s best entrepreneurs that are running the fastest-growing, highest-quality companies at the most attractive valuations we can find. Here’s to many more years of focusing on the inputs and letting the outputs take care of themselves.
Sincerely,
Ryan Reeves
Disclosures
Infuse Asset Management LP (“Infuse”) is an investment management company to a fund that is in the business of buying and selling securities and other financial instruments. This information is provided for informational purposes only and does not constitute investment advice or an offer or solicitation to buy or sell an interest in a private fund or any other security. An offer or solicitation of an investment in a private fund will only be made to accredited investors pursuant to a private placement memorandum and associated documents.
Infuse may change its views about or its investment positions in any of the securities mentioned in this document at any time, for any reason or no reason. Infuse may buy, sell, or otherwise change the form or substance of any of its investments. Infuse disclaims any obligation to notify the market of any such changes.
The S&P 500 is a U.S. equity index. It is included for informational purposes only and may not be representative of the type of investments made by the fund. References made to this index are for comparative purposes only. Reference to an index does not imply that the funds will achieve returns, volatility, or other results similar to the index. The fund’s portfolios are less diversified than this index. Returns for the index are total returns which includes dividends and do not reflect the deduction of any fees or expenses which would reduce returns.
An investment in the fund is speculative and involves a high degree of risk. The portfolio is under the sole trading authority of the general partner. An investor should not make an investment unless the investor is prepared to lose all or a substantial portion of its investment. The fees and expenses charged in connection with this investment may be higher than the fees and expenses of other investment alternatives and may offset profits.
The information in this material is only current as of the date indicated and may be superseded by subsequent market events or for other reasons. Statements concerning financial market trends are based on current market conditions, which will fluctuate. Any statements of opinion constitute only current opinions of Infuse which are subject to change and which Infuse does not undertake to update. Due to, among other things, the volatile nature of the markets, and an investment in the fund/partnership may only be suitable for certain investors. Parties should independently investigate any investment strategy or manager, and should consult with qualified investment, legal and tax professionals before making any investment.
The fund is not registered under the investment company act of 1940, as amended, in reliance on an exemption thereunder. Interests in the fund have not been registered under the securities act of 1933, as amended, or the securities laws of any state and are being offered and sold in reliance on exemptions from the registration requirements of said act and laws.
Performance Appendix
Note that Infuse Partners LP officially launched on August 8th, 2022 so Q3 was a partial quarter. The S&P 500 returns are calculated from August 8th to September 30th as well.