Defending Chamath Palihapitiya and Debunking Revisionist History
That Was The Week 2023 #32
A reminder for new readers. That Was The Week collects the best writing on critical issues in tech, startups, and venture capital. I selected the articles because they are of interest. The selections often include things I disagree with. The articles are only snippets. Click on the headline to go to the original. I express my point of view in the editorial and the weekly video below.
This Week’s Video and Podcast:
Content this week from @kteare, @ajkeen, @chamath, @geneteare, @daveclark85, @mattturck, @scottbelsky, @krishnanrohit, @abhim_eth, @altitude_vc, @eladgil, @JaapVriesendorp, @Tanayj, @EricNewcomer
Editorial: Hit Job
VC Returns - The Power Law
Rest - The case for sabbaticals
The Scam in the Arena - Is Chamath a Scammer?
Editorial: Hit Job
I never believed that a SPAC was a good idea for a company.
InFarm, where I was head of corporate development during its exploration of a SPAC filing, spent over 12 months assessing the viability of a SPAC and decided not to proceed. That was in 2019-21.
And only yesterday, Better.com listed its SPAC merger, and it fell 95% within the first day, wiping out all remaining investors.
From Fast Company
Shares of the Softbank-backed company plunged 93% as it began trading as BETR on the Nasdaq Thursday, falling more than $16 per share to $1.19 by mid-day. (Better went public via a merger with special purpose acquisition company [SPAC] Aurora Acquisition Corp.)
The SPAC is still the vehicle taking Better public, but it’s not the source of financing it once was. The money raised in the deal will come in the form of a second convertible note from a SoftBank affiliate.
But I am also not prepared to say SPACs were or are a bad idea in 2019 or today. And I do not believe sponsors form SPACs to defraud investors and enrich themselves. More on that below.
A class of company typically needs significant capital expenditure and a long-term view of success, for whom, in 2019 and 2020, a SPAC seemed a great path. Virgin Galactic comes to mind, and OpenDoor. And many others.
In 2019 and 2020, as capital markets were paying enormous multiples for business plans, it seemed an even better idea. Since mid-2020 that potential dried up as the likelihood of such valuations receded and then disappeared.
Chamath Palihapitiya was one of the key architects of those and many other SPACs.
Eric Newcomer published what I consider a hit job on Chamath this week and led with a chart showing the dire performance of the now-public SPACs since listing.
His article - The Scam in the Arena - is clearly driven by some anger or passion. I have not spoken to Eric, and I do subscribe to Newcomer and admire and value him in general - but this quote kind of sums up his mood for me:
And the reality is that Palihapitiya got away with it. He’s boastfully summering in Italy and reveling in his luxurious wedding on the All-In Podcast that he hosts. Elon Musk and Grimes were apparently in attendance.
To me, the SPAC mania was such an obvious cash grab and ego play that I couldn’t get all that upset about it when it crumbled. The writing had been on the wall from the beginning.
In 2019, nobody saw the writing on the wall, except perhaps in an abstract sense that valuations were super high and possibly would not last forever. “The End is Nigh” kind of analysis.
What Eric is doing in his piece is known as Reading History Backwards. It attributes to the past an intent to create the outcomes we see in the present.
But when we take a current fact - SPACs have not worked out - and then seek to imply that those engaged in them both knew that would happen and worse, plotted to take money off investors for personal gain, despite that knowledge. Well, that is a hit job.
Eric is nothing less than full-throated in his purpose. It’s a long read, but I recommend you see it for yourself.
Eric's broader conclusion is that Silicon Valley abandoned its principles in 2019-2022 and that SPACs were part of it.
From my vantage point, Silicon Valley’s pivot to SPAC mania was part of a broader abandonment of the venture capital industry’s principles by money-hungry pockets of the industry.
This is also wrong.
Silicon Valley really has no principles, and so never abandoned them.
The Valley is money seeking to make more money. SPACs were a way to do so.
Of course, the companies themselves had a broader purpose. However, investors should not be confused with operators and founders.
Investors can and must have a singular focus on money. That is the promise they make to their investors - Limited Partners.
It was not irrational to believe that SPACs could deliver value to both companies and investors at that time, and in the future, they may serve that purpose again.
Indeed, despite the stock price Better.com now has over half a billion dollars to execute from. Its market capitalization is under $20m at the current price. Less than 4% of its cash balance. The management has the opportunity to build a business with capital that would have been hard to raise in private markets.
Eric speaks about a SPAC as a shortcut for a loss-making company. The implication is that the purpose of the SPAC is the shortcut, and making a loss is a red flag. His narrative implies deceit, if not fraud.
But there is nothing wrong with a shortcut or making a loss. The actual purpose is to build a balance sheet to execute from (for the operator) and to get liquidity in the stock (for the investors).
Of course, if you always make a loss, then Eric is right.
But most big ideas need funding, and before they mature, they make investments to build future success. You could call that a “loss”. But it is an investment in an outcome that requires time and effort.
The idea that public markets are not a good place for growth companies (another word for loss-making) is spurious and elitist. It implies that only rich people who qualify to be private investors should be able to invest in them.
A SPAC is a means of a team with a big idea that needs capital to grow to seek that capital from the public. The growth that the capital brings can then be shared by public investors. And that is why I think we will see SPACs again.
Eric ends with a comment:
Chamath! That was the trade that you were engaged in. You swapped your reputation for SPAC sponsor fees. You should have done some diligence on what you were getting yourself into. How were you underwriting the damage to your reputation?
I think it deserves a response on Chamath’s behalf. Not that Chamath needs help.
Eric seems to have had a bad week. His claims are no better than a conspiracy theory. His wedding and celebrity references are really just a cheap shot and come across like jealousy.
Eric’s attack implies forethought and planning intending to defraud investors. It really does not hold up.
Some diligence on the broader applicability of the SPAC as a mechanism to enable growth companies to execute long-term capital-intensive runways would be welcome. A lot of innovation requires this kind of financing.
This week’s newsletter is full of IPO stories, Instacart is one, and many of them are loss-making. They will, and should, go public. Not as an exercise in self-serving money-making but because the capital available in public markets is sufficient to fund big ideas.
A call-out to the piece about IRRs being boosted by artificial early exits and to David Clark’s piece on the Power Law of venture capital and how it works. My colleague Rob Hodginson’s Series B Ecosystem report is good work.
Essays of the Week
The Power Law and its Consequences
Whilst a select few are rich enough to invest their own money, 99% of VCs raise money from rich people, large institutions and other corporations.
In particular, the backbone of the VC ecosystem is institutional investors such as college endowments, pension funds and large foundations.
VCs have a fiduciary duty to deploy their capital responsibly. However, in recent times, the incentives guiding institutional investors, VCs and the founders that receive investment misaligned.
In this article, we’ll unpack how institutional investors construct their portfolios, make investment decisions and the impact that has on the VC ecosystem.
The Endowment Model
Institutional investors all have different return expectations, risk limits, assets under management and portfolio construction preferences. Some investors might purely invest in funds and others might supplement fund investing with direct investments in startups.
Amongst the largest institutional investors are grant-making endowments and foundations. This means that every year, they must allocate a certain percentage of their portfolio (typically 5%) to grants for their community. These investors also look to make grants in perpetuity, meaning they have low liquidity requirements.
As a result, most of these institutions follow a popular portfolio construction methodology called the Endowment Model developed by the late David Swensen. Swensen spent 36 years managing Yale's Endowment fund and pioneered a focus towards investing in illiquid assets such as Real Estate, Infrastructure, Private Equity and Venture Capital. By doing so, Swensen theorised that the illiquidity of these asset classes should be embraced in line with the Endowment’s low liquidity requirements.
If you’re enjoying this article so far, I’d love to have you as part of our growing community! Make sure to subscribe to get next week’s article :)
As such, most institutions that adopt the Endowment Model might have around 10-30% of their portfolio allocated towards private, illiquid investments. Out of this, 2-10% could be allocated to Venture Capital.
Over the last 13 years, assets under management grew rapidly for all of these endowments. Either through stellar returns, or large donations, many of these institutions had more money than they knew how to deploy. Obviously, the measly 2-10% allocation to VC also grew in absolute dollar terms resulting in a requirement to deploy large chunks of cash at a quicker pace.
However, deploying capital into private investments is tricky. Fund managers only raise new funds periodically, and aren't capable of taking on an unlimited supply of funds. Moreover, capital is drawn down over time through capital calls, rather than deployed upfront.
This means that institutional investors plan out their prospective commitments usually at the start of the year, with a mid-year review. The goal of this process is to ensure that the portfolio is either at its target allocation or has a line of sight towards getting there.
The planning exercise requires a bit of modelling by taking into account new commitments over the year, capital calls from previous commitments and any distributions/returns that might come back from fund managers. It is in the LPs’ best interest for fund managers to call capital (e.g., quarterly) and raise new funds (e.g., every 2-3 years) at a regular cadence.
Investing in a VC fund is basically Power Law^2.
VC Funds run on a power law basis, where a small number of investments contribute to the majority of returns.
On top of this, VC fund returns also exhibit a power law distribution. High-performing funds are rare and can easily make up for a bunch of duds.
From the table, above, we can look at the distribution of returns from funds in each given vintage. In 2013, the top decile threshold is 3.58x, meaning for every $1 invested, the best 2013 vintage funds, returned at least $3.58 back, if not more. As you can see, the disparity in TVPI between the top decile and the median is quite substantial between 2013 - 2016.
To exploit the power law, VC funds usually invest in 20 - 30+ companies with the expectation that a couple will return the fund. This doesn’t translate to fund investing due to a few incentive misalignments and nuances that exist for institutional investors.
Most employees at an institutional investment fund are not incentivised by outperformance. Many receive a standard salary package and perhaps an annual bonus, but no form of carry or share in the institution’s annual profits. As a result, there is no direct correlation between an employee's ability to pick strong fund managers to their remuneration.
These funds look to maintain their assets under management in line with their grant-making activities and inflation. Thus, their return expectations are usually between 3.5% - 5.5% + CPI. This isn't a hard hurdle to hit, meaning funds don't need to take on extra risk in order to achieve their return target.
As with VCs needing to fight for access to the best deals, institutional investors also need to fight for access to the best fund managers. In many cases, top VCs such as Sequoia and Benchmark already have a full stable of LPs that will continue to invest in future funds. That means access is only available when one of these LPs doesn't invest in the next fund, which is a rare occurrence.
Institutional investors have their own fund concentration limits to abide by. This means that for an institutional investor to invest in a VC fund, their commitment, shouldn't exceed X% of the total VC Fund size. Usually, this limit is 20-30% of a fund. E.g., if VC Fund A is raising a $100M fund, Institutional Investor B can only invest $20-30M in that fund.
Once an institutional investor chooses a fund manager, they will look to invest in multiple funds over a period of time. By investing in funds that start investing in different years, investors hope to get vintage-year diversification. The vintage year for a fund is the year in which it makes its first investment. This is important because fund performance might vary wildly depending on the broader macro climate during its investment period.
As a result, institutional investment portfolios are typically optimised for easy ongoing management and 'lower' risk-taking. That means that it is incredibly hard for new fund managers to receive capital from an institutional investor with an existing private investment portfolio. As the saying goes, "Nobody gets fired for buying IBM"…. more
𝗦𝗼𝘂𝗿𝗰𝗶𝗻𝗴, 𝗦𝗲𝗹𝗲𝗰𝘁𝗶𝗻𝗴, 𝗪𝗶𝗻𝗻𝗶𝗻𝗴 & 𝗘𝘅𝗶𝘁𝗶𝗻𝗴 – 𝘁𝗵𝗲𝘀𝗲 𝗮𝗿𝗲 𝘁𝗵𝗲 𝗰𝗼𝗿𝗲 𝗰𝗼𝗺𝗽𝗼𝗻𝗲𝗻𝘁𝘀 𝗼𝗳 𝗩𝗖.
Exiting portfolio companies stands as one of the most overlooked skills in Venture Capital. 𝗧𝗵𝗲 𝘀𝗶𝗴𝗻𝗶𝗳𝗶𝗰𝗮𝗻𝗰𝗲 𝗼𝗳 𝗿𝗲𝗰𝗼𝗴𝗻𝗶𝘇𝗶𝗻𝗴 𝗻𝗼𝗻-𝘄𝗶𝗻𝗻𝗲𝗿𝘀 𝗲𝗮𝗿𝗹𝘆 𝗰𝗮𝗻 𝗰𝗿𝗲𝗮𝘁𝗲 𝘁𝗵𝗲 𝗱𝗶𝘀𝘁𝗶𝗻𝗰𝘁𝗶𝗼𝗻 𝗯𝗲𝘁𝘄𝗲𝗲𝗻 𝗮 𝗴𝗼𝗼𝗱 𝗮𝗻𝗱 𝗮 𝘁𝗼𝗽-𝘁𝗶𝗲𝗿 𝗩𝗖 𝗳𝘂𝗻𝗱. In this context, exiting doesn't solely pertain to selling portfolio winners; it's equally about divesting from the medium and non-performers as soon as feasible to capture maximum value for LPs.
Selling portfolio winners resembles offering candy to children, while selling mid- and non-performers is more akin to selling sand in the desert. Some GPs believe companies are bought (a sit-and-wait strategy), while others believe companies should be sold (a plan-and-push strategy). I firmly advocate for the latter - 𝗚𝗣𝘀 𝘀𝗵𝗼𝘂𝗹𝗱 𝗰𝗼𝗻𝘀𝗶𝗱𝗲𝗿 𝗵𝗶𝗿𝗶𝗻𝗴 𝗮 “𝗖𝗵𝗶𝗲𝗳 𝗼𝗳 𝗘𝘅𝗶𝘁” 𝗳𝗼𝗿 𝘁𝗵𝗲 𝘀𝗲𝗰𝗼𝗻𝗱 𝗵𝗮𝗹𝗳 𝗼𝗳 𝗮 𝗳𝘂𝗻𝗱 𝘁𝗲𝗿𝗺.
This insightful analysis by Vintage Investment Partners underscores the significance of early divestment from non-winners. 𝗧𝗵𝗶𝘀 𝗻𝗼𝘁 𝗼𝗻𝗹𝘆 𝗯𝗼𝗼𝘀𝘁𝘀 𝗜𝗥𝗥𝘀 𝗯𝘂𝘁 𝗮𝗹𝘀𝗼, 𝗶𝗺𝗽𝗼𝗿𝘁𝗮𝗻𝘁𝗹𝘆, 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗲𝘀 𝗲𝗮𝗿𝗹𝘆 𝗗𝗣𝗜𝘀, 𝘄𝗵𝗶𝗰𝗵 𝗵𝗼𝗹𝗱𝘀 𝗿𝗲𝗹𝗲𝘃𝗮𝗻𝗰𝗲 𝗳𝗼𝗿 𝘀𝘂𝗯𝘀𝗲𝗾𝘂𝗲𝗻𝘁 𝗳𝘂𝗻𝗱𝗿𝗮𝗶𝘀𝗶𝗻𝗴 𝗳𝗼𝗿 𝘀𝘂𝗰𝗰𝗲𝘀𝘀𝗼𝗿 𝗳𝘂𝗻𝗱𝘀. To be fair, the analysis assumes that a GP can identify non-performers starting from Year 4 and proactively divest from them - a task surely not as straightforward as it may seem.
Nevertheless, if successful, 𝘁𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁𝘀 𝗰𝗼𝘂𝗹𝗱 𝗹𝗲𝗮𝗱 𝘁𝗼 𝗮 𝟰𝟭% 𝗶𝗻𝗰𝗿𝗲𝗮𝘀𝗲 𝗶𝗻 𝗜𝗥𝗥 (𝗳𝗿𝗼𝗺 𝟮𝟭% 𝘁𝗼 𝟯𝟭%) 𝗮𝗻𝗱 𝗮 𝟯𝟱% 𝗶𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁 𝗶𝗻 𝗗𝗣𝗜 (𝗳𝗿𝗼𝗺 𝟮.𝟴𝘅 𝘁𝗼 𝟯.𝟴𝘅). Interestingly, the decline in future IRR would be less drastic, as the IRR becomes more "locked-in" due to the early exits.
Which strategy do you believe in: 𝘁𝗵𝗲 "𝘀𝗶𝘁 𝗮𝗻𝗱 𝘄𝗮𝗶𝘁 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆" 𝗼𝗿 𝘁𝗵𝗲 "𝗽𝗹𝗮𝗻 𝗮𝗻𝗱 𝗽𝘂𝘀𝗵 𝘀𝘁𝗿𝗮𝘁𝗲𝗴𝘆"?
In stark contrast to the private equity and venture capital business, fund investment business is not that sexy. In fact, I often wonder how I ended up here. At dinner parties, whenever I am asked to explain what I do, and I tell them that I invest into private equity and venture capital funds, I can see them glancing over my shoulder to find another person to talk to. And I get it. It seems abstract and boring at first. But I do believe it can be very interesting from a strategic, analytical and even philosophical point of view. For that and mostly other reasons, it seemed a good idea to explain some more about fund investments. Below the first piece on fund of funds. Our bread and butter at Welt and Marktlink Capital.
What’s in a name: fund of funds, umbrella funds or multi manager funds
What is a fund of fund? Whereas private equity and venture capital funds invest directly into companies, often referred to by a fancy name as ‘assets’, fund of funds invest into a blend of private equity and venture capital funds, often referred to by a fancy name as ‘managers’. Fund of funds are therefore also often referred to as umbrella funds (i.e., many funds under one umbrella) or multi manager funds (i.e., one fund with multi underlying managers). If you would compare it to stock market investing, it’s similar to holding a relatively portfolio of individual stocks, compared to stock picking (investing in one company or fund) or buying an index (investing in all companies or funds).
Needless to say, this alters the risk-return profile considerably of the investment. Whereas investing in a single company can be considered ‘high risk’ and investing in a single fund ‘medium risk’, investing in a fund of fund can bring the risk profile further down. This is also logical, given that the number of portfolio companies increases significantly from a single asset, to a fund (on average 10–20 assets) to a fund of fund (on average 250+ assets). Because of the risk limiting nature (and other factors such as access and diversification), fund of funds have become popular in higher risk asset classes such as private equity and venture capital. In a study conducted by Cambridge Associates on the performance of different asset classes between 2006 and 2021, private equity and venture capital ranked both as the highest performing asset classes (both in terms of median and top quartile performance) and as the most risky ones (where risk is defined as the dispersion between the best and the worst performance).
Because predicting future performance of a fund is a tricky thing (I will write a separate article on the art and science of it), people tend to invest either in multiple funds or in a fund of fund. This statistically dramatically reduces the risks. Lets take venture capital as an example. If you would have invested randomly into one fund between 1991–2005* (i.e., monkey taking a marble out a jar of 3077 marbles) this would have led to you losing money in 23.6% of cases. Depending on your risk appetite, you may think this is low or high. What is more interesting however, is that investing in 9 funds historically however, reduces the chance of you losing money to less than 1%. So, by spreading your bets over 9 funds, you reduce the risk of losing your money by over 25x.
August 18, 2023
As the venture market has significantly slowed, so has the pace of minting young startups as unicorns.
With valuations going down across the board, investors seem more wary than they have in the past few years about anointing a $1 billion-plus valuation to a company before its Series C.
Of the 47 unicorns minted through the first seven months of the year, only 18 of them joined The Crunchbase Unicorn Board after an early-stage funding round — defined as seed, Series A or Series B — an analysis of Crunchbase data shows.
That is a far cry from 2021, when 107 early-round funding startups were given a $1 billion valuation. Even last year, 77 such companies hit that valuation after an early funding.
At the current rate, only 31 early-stage funding startups will hit unicorn status this year, putting it on pace for the lowest count since 2019.
Of course, the numbers are not shocking, as venture capitalists have started to follow the lead of public investors and are putting more emphasis on profitability and positive cash flow. That is something often difficult for early-stage startups to reach, as many are barely in their go-to-market phase.
The number of young startups to hit unicorn status quickly also would be a lot lower if it weren’t for two letters — AI.
The U.S. led all regions in early-round unicorns with nine, with China just behind with six. Half of those U.S.-based, early-round unicorns minted this year come from the generative AI sector, including:
at a $1 billion valuation. The company’s AI platform helps with enterprise content creation such as product shots, blog posts, social media ads and job posts that meet brand specifications.
Other industries related to EV charging, solar, energy storage and mineral extraction also saw some young startups become minted unicorns in the year’s first seven months.
Perhaps the most unusual of all fast, early-round unicorns this year was the minting of Colossal Biosciences. The Dallas-based startup raised a $150 million Series B that gave it a valuation of more than $1 billion, per reports. The startup, which launched in 2021, is developing a de-extinction platform that could bring back species such as the Dodo bird and the woolly mammoth, and help with conservation and human health care.
Fast to rise, fast to fail
While early-round unicorn numbers are down, it also is important to remember unicorn numbers in general are down. While last year saw 311 total unicorns minted, there have been fewer than 50 so far this year through July.
Fast, early-round unicorns actually make up a greater percentage of all unicorns minted this year when compared to past years, as 38% of new unicorns in 2023 achieved that status in a seed through Series B funding round.
The case for sabbaticals
AUG 21, 2023
Picture a young Albert Einstein working as a patent clerk in 1905. He has a steady job, but his mind remains restless, filled with ideas that clash with the rigid conventions of physics. During the day, he mechanically processes patent applications. But away from the office, he plumbs the mysteries of the universe. This freedom allows bursts of imagination outside his daily routine. He asks, what would happen if you rode a beam of light? That year, he publishes four revolutionary papers that reshape science.
This is either the beginning of a tragedy or an incredible boon, depending on whom you ask. If you’re running the patent office, you might wonder why someone of such intellect would not be interested in improving the office itself, instead of imagining themselves riding alongside a photon. If you’re a scientist you might wonder how close we were to actually losing Einstein, if the patent office work was just a bit more time-consuming or difficult!
Today, it's nearly impossible for most professionals to enjoy such intellectual freedom. In the always-on economy, taking months or a year for unstructured exploration has become extinct. Two weeks of hurried vacation is the norm, if you're lucky. Sabbaticals are a quaint relic, reserved for tenured academics. Yet history reveals the immense value of this "long time.”
And we’re paying the price. Creativity has stagnated across industries. The few who escape enjoy epiphanies like Einstein in 1905. The always-on economy has robbed most professionals of this gift.
Isaac Newton did some of his most important work during the year Cambridge closed for the Great Plague in 1665-66, producing the principles of calculus, optics and gravitational theory. Albert Einstein worked at the patent office for 7 years after college before his "miracle year" of breakthroughs in 1905. Charles Darwin spent 5 years traveling the world on the Beagle voyage, crucially important for formulating his theories of evolution.
Marie Curie took a break from her studies to tutor and study independently, and this started her on research that led to Nobel prizes.
It seems to happen more in creative circles. Tolstoy took a break to recover from a spiritual crisis and immersed himself in studying religion, eventually ending in writing Anna Karenina. Joseph Heller took a year off after writing Catch-22 to travel and reflect.
Paul McCartney, not one to make lesser songs and known for his incredible work ethic, took a break to spend time on his farm. This is the man unburdened by creative roadblocks, who composed Hey Jude while driving to meet John Lennon’s ex-wife and son.
Not to mention Eat, Pray, Love. Or when James Cameron spent a decade becoming a significant oceanographical explorer and inventor in his own right, and then made Avatar. Or where JK Rowling moved to Portugal to teach English, and started writing the Harry Potter novels in that time.
It also happens in entrepreneurship, especially in tech, where people start companies after taking breaks from selling their previous ones. Brian Acton comes to mind, where he took a year off to travel after leaving Yahoo and then came up with WhatsApp. Drew Houston wrote the code for Dropbox during travels too. Travis Kalanick the same re Uber. Max Levchin apparently same for Slide and Affirm.
This is also the answer to one of Patrick Collison’s questions, on why people sometimes decide to make big changes to their lives.
Most days, people don't decide to change their lives in big ways. On a few days, they do. What's special about those days? How much is it about the stimulus versus their own inner state?
Because they almost never get a chance to.
To put it simply: higher LTV:CAC → higher margins → higher valuation.
Investors often use 3x LTV:CAC as a rough benchmark of a consumer company’s financial health. If your customer lifetime value (LTV) is 3 or more times your customer acquisition cost (CAC) within 5 years, that means your company has efficient returns on sales and marketing spend. But there’s little discussion of how a higher LTV:CAC actually translates to long-term profitability and, ultimately, valuation.
Solid unit economics has a cascading effect across your business: a higher LTV:CAC ratio means for every dollar of sales and marketing investment, your company has higher margins and so more profit to reinvest back into its business, which means that you can build better products and, hopefully, capture more market demand. Companies are ultimately valued on their future cash flow generation, so the higher your margins, the higher your valuation.
In fact, improving your LTV:CAC from 2x to 3x can nearly triple your valuation.
To illustrate, let’s walk through some calculations using the long-term margin projections across 60+ US public consumer internet companies with extensive sell-side equity analyst coverage.
A few notes before we jump in. For the sake of this post, we consider LTV as your gross profit (as a proxy for contribution profit) from a customer over time and CAC as the total sales and marketing (S&M) spend that goes into acquiring that customer. We generally prefer to calculate using historical monthly customer cohort data for LTV and the CAC associated with that specific month, but there are several ways to calculate LTV:CAC. (Our piece on 16 startup metrics offers more details on how we approach calculating it, while this piece is a great deeper dive into the various considerations that go into an LTV:CAC calculation.)
In our calculations for this piece, we’ve calculated both margins and valuations based on gross profit. Relative to enterprise companies, consumer companies have a wider range of business models (e.g., first-party e-commerce, third-party e-commerce, social networks, subscription-based content, etc.), and thus COGS (cost of goods sold) structures. Focusing on gross profit can help standardize across these differences. As such, S&M, R&D, G&A and operating margins are calculated here by dividing by gross profit, instead of by the more commonly used revenue.
Higher LTV:CAC → higher margins
Let’s start by looking at the inputs into a company’s OpEx (operating expense) lines. Typically, OpEx spend breaks down into 3 buckets: R&D (research and development), G&A (general and administrative), and S&M (sales and marketing). Subtracting OpEx from gross profit is a company’s operating income. In the chart below, we can see that companies with higher gross margins tend to spend more on R&D and G&A when calculated as a percentage of revenue. However, when we adjust those spends to be a percentage of gross profit, they’re much more consistent across company buckets—R&D and G&A margins average around 20% and 14%.
If, as this suggests, R&D and G&A spends as a percentage of gross profit are constant across companies, then your S&M spend as a percentage of gross profit is what drives your long-term operating income as a percentage of gross profit. Your S&M spend is your CAC—the higher your LTV:CAC ratio, the more efficient that spend. In other words, optimizing how you acquire customers has the most impact on how profitable and efficient your company is in the long haul. For companies that have the same gross margins, this relationship between S&M efficiency and a company’s margins can be extended to margins when calculated as a percentage of revenue as well.
We see this play out in the math below. Using the 20% R&D and 14% G&A of gross profit from above and a hypothetical S&M spend of $100, we can see that companies with higher LTV:CAC ratios have higher operating income and margins. Improving your LTV:CAC from 2x LTV:CAC to 3x, for instance, could improve that from 16% to 33%.
Like we said earlier, higher margins means you can invest more dollars back into your business. For that same $100 S&M spend, higher LTV:CAC ratios lead to more gross profit—which means that those same R&D and G&A as percentages of gross profit equate to more dollars invested in the business. As the graphic below shows, for a $100 of CAC spend, $68 is invested into R&D and G&A for a 2x LTV:CAC business, $102 for a 3x business, and $170 for a 5x business.
Higher margins → higher valuation multiple
A company’s value is the net present value of its future cash flows. More profitable companies are more valuable because they can generate more profit for each dollar earned. This means that the higher your margins, the higher your valuation. It’s worth noting that investors also consider growth when calculating a company’s valuation multiple. The 2023 market has thus far rewarded margins over growth, but both are important contributors.
When we look at the relationship between long-term margins and gross profit multiple, we find that consumer internet public companies with ~16% margins (the above derived margin in our 2x LTV:CAC example) trade at ~1.5x forward gross profit. Companies with 33% long-term margins (our derived margin for 3x LTV:CAC) trade for ~5.3x, and companies with 46% margins (5x LTV:CAC) trade for ~8.4x. So for a given amount of gross profit dollars, our 3x LTV:CAC business is worth more than 3 times our 2x LTV:CAC business, and our 5x LTV:CAC business is worth more than 5 times!
In short, there’s a good reason why LTV:CAC is a key metric of best-in-class consumer internet companies. For more on metrics and benchmarks for B2B companies, check out our Guide to Growth Metrics.
AI of the Week
Published on August 1, 2023
by Chandar Lal
In the summer of 1999, the Economist published a special report on “The Net Imperative”. It led with a bold prediction from Intel’s chairman:
In five years' time, all companies will be Internet companies, or they won't be companies at all.
This would have sounded like hubris during the ensuing dotcom collapse, but it ultimately came true. A massive platform shift took place – and now the default state is to be an “Internet company”, whether you’re selling books, handling payments, or running taxis.
Fast forward, and generative AI is emerging as a shift of similar scale.
Some of today’s tech executives see AI as “more profound than fire or electricity” – set to usher in an “age of abundance”, or perhaps to destroy human civilisation. There is a lot of hype – but it’s reasonable to expect that LLMs and generative AI will be a foundational layer for the next decade’s tech companies, and this opportunity is easily big enough to excite us.
The generative AI venture landscape has blossomed rapidly, from applications to infrastructure and tools. In this post, we focus on the application layer – where we have historically been most active. It seems inevitable that generative AI will be embedded in most applications we touch, enabling entirely new experiences as well as large-scale automation. Less certain is the pace of this change, and its second-order effects. It’s one thing to tinker with ChatGPT; it’s quite another to adopt a product that structurally reinvents a long-held human task.
And when this adoption ramps up, what happens to the nature of work and the structure of the workforce? We take an optimistic view: as our venture partner, Benedict Evans, puts it:
We should start by remembering that we’ve been automating work for 200 years. Every time we go through a wave of automation, whole classes of jobs go away, but new classes of jobs get created. There is frictional pain and dislocation in that process, and sometimes the new jobs go to different people in different places, but over time the total number of jobs doesn’t go down, and we have all become more prosperous.
This is a Schumpeterian moment, and a generational opportunity for startups. But we are no longer in an era of zero interest rate policy, and not all boats will float despite the abundance of capital being deployed. So how do we navigate this emerging landscape?
Mosaic: long-term investors in machine learning applications
At Mosaic, we’ve been investing in applications of machine learning since our firm’s inception. The common thesis, unsurprisingly, has been to invest in products that solve hard and economically valuable problems at superhuman scale – and are differentiated by their proprietary models and/or datasets.
Looking forward, generative AI will be an equally central component of our investment strategy. We expect that AI-native companies will dominate the next wave of product innovation – but as investors, the key questions and heuristics that matter are changing.
Evaluating generative AI applications: the key questions
New foundational models and tools make it easier for teams to build viable AI applications, with little training data or computing resources required. That sounds promising on paper. But if this is true, and proprietary datasets confer less of an advantage, what do the new moats look like? If it’s easier than ever to ship an AI-native product, but harder to build a truly differentiated one, is it incumbents who stand to win – benefiting from their head start with distribution? Could this a structural problem for new entrants?
There are many hard questions for an investor to explore. Generative AI applications are now proliferating: we’re currently tracking over 500 early-stage European startups offering LLM-powered applications. So we need to be judicious about picking the right problem to solve, a product approach that is truly novel, and the most effective route to market.
In the spirit of transparency, here are a few of the qualitative questions on which our investment decisions often hinge.
Market opportunity. For some of the most exciting AI products, there is no “current state” offering: they are greenfield opportunities, enabling a product that simply wasn’t possible before, or serving an entirely new audience. It’s notoriously difficult to estimate the market potential of these. As an example: what if we could make an AI-powered private tutor accessible to the entire population of high school students? If we could offer human-equivalent personalised tuition to everyone, what would that be worth (in terms of new demand and its price elasticity)? More on that question here...
Most opportunities, conversely, are brownfield – either augmenting or replacing an existing human activity. Here, we ask: which human tasks can be meaningfully assisted or even displaced by a novel AI application? What economic value is generated by this human work? How expensive or difficult is it; and how scarce are the people capable of doing it? Hence, what cost efficiencies or scale benefits are achievable through automation? This points us to an initial view of the market potential of automating a task that historically couldn’t be.
Product. Our first question is: what entirely new capabilities can be unlocked using a LLM? Crucially, what does a novel product need to look and feel like? The model alone is not a product, and text boxes for prompts will rarely be the form factor that eventually wins. How do you serve up a model in a way that users are intuitively able to interact with? Looking further into the future, how durable is the value being created? For example: if every sales outreach email you receive is highly personalised, what relative advantage will any one of those emails have?
Defensibility. One of the hardest questions for any emerging AI application is simply: where is the moat? There is a common set of questions we often ask. Is the product a 'thin layer' on top of a general-purpose LLM that delivers adequate performance – or is there substantial, domain-specific training or fine-tuning required? If there is a unique dataset needed to build a compelling product, does the team have access to it? Are there specific architectural features (prompt engineering / chaining; multi-model) that contribute to a noticeably better outcome? Does the company have a clearly defined and distinctive route-to-market to sell its solution? Is the application in a regulated industry, that has high barriers to entry which might deter competitors?
Go-to-market. Our hypothesis is that if generative AI products will become even easier and cheaper to build over time, then distribution / go-to-market execution is what will separate the truly great from the good. As a result, we lean heavily towards founding teams that can demonstrate strong commercial instincts, with a clear understanding of who their buyer is, and how to reach them most effectively.
The European AI opportunity
The generative AI wave is a global opportunity, but Europe is particularly well positioned to capitalise on it. The continent has produced its first crop of AI-native unicorns, and we think this will only accelerate, given:
its depth of research output and talent (especially with the UK's "Golden Triangle", France's Polytechnique, ENS, INRIA, and CentraleSupelec, and leading ML institutes in DACH);
its flywheel of company creation;
its leadership in setting global policy on safety, privacy, and alignment.
Conclusions and unknowns
Clearly, we’re very optimistic about the next generation of generative AI applications that will emerge in Europe. Yet at this point, we have far more questions than answers. This is little surprise, as we’re at the very beginning of a major platform shift, where the first major applications are just beginning to emerge. There are, of course, risks that we are observing carefully – privacy, safety, alignment – where Europe appears to be an emerging leader in setting the guardrails.
But if we allow ourselves to dream, it’s easy to imagine a new era of creativity and productivity, with every human assisted in their desire to create beautiful art, write production-ready code, or simply edit a blog post – as GPT-4 has done here!
If you’re building a generative AI application, we’d love to chat, and learn how you’ll reshape your market and capture enduring value. In our next posts, we’ll delve into specific industries that we think will be transformed by emerging AI applications, as well as the tools and infrastructure software that will underpin this wave.
Rather then view LLMs, Transformers, and diffusion models as part of a continuum with past "AI", it is worth thinking of this as an entirely new era and discontinuity from the past
AUG 21, 2023
I worked on early ML systems and products at Google and later at Twitter (after they bought my company, Mixer Labs). I then spent a decade working as a founder and executive & investing in machine learning companies. Until the rise of new AI architectures (in particular transformer-based and diffusion-model based approaches), roughly all machine learning startups failed. Value in prior AI waves went largely to incumbents over startups - as the capabilities were not advanced enough to create new market openings.
Here is a slide I used to use (borrowed from Brandon Ballinger) during 2017-2019 or so - this slide reflected the CNN/RNN/GAN world of the prior ML wave.
Thanks for reading Elad Blog! Subscribe for free to receive new posts and support my work.
Pledge your support
When many business people talk about “AI” today, they treat it as a continuum with past capabilities of the CNN/RNN/GAN world. In reality it is a step function in new capabilities and products enabled, and marks the dawn of a new era of tech.
It is almost like cars existed, and someone invented an airplane and said “an airplane is just another kind of car - but with wings” - instead of mentioning all the new use cases and impact to travel, logistics, defense, and other areas. The era of aviation would have kicked off, not the “era of even faster cars”.
(We should of course, fully recognize how important prior waves of ML and deep learning were and are to all this - however, to treat it as an ongoing continuum may miss the seismic nature of this technology shift).
Slide I now use.
The biggest inklings that something interesting was afoot came kicked with GPT-3 launching in June 2020. GPT-3 was a massive step up from GPT-2 and prior models. It was not quite good enough to do all the things we now view as hallmarks of “AI”, but it was highly suggestive of what was to come (I went on the A16Z podcast a few months later to talk about GPT-3, as it was so striking). For those in the know, the launch of GPT-3.5 in March 2022 solidified the perception of transformer-based models as the future. Internally at companies like Google, OpenAI, Microsoft, and Anthropic, early access to models gave a subset of people a glimpse of the future that was coming. This led to a Google engineer eventually proclaiming an internal AI chatbot named LaMDA as being “sentient” - this chatbot was a sort of predecessor to chatGPT and products like Character.AI.
The real starting guns for this AI wave in terms of a large number of founders jumping in was driven by two sets of launches. First were the launches of image-gen products like Midjourney and Stable Diffusion, followed a few months later by ChatGPT, which wowed the world, captured the public imagination, and was the AI startup big bang moment. ChatGPT truly highlighted the capabilities of these new forms of AI and the power of RLHF. OpenAI followed up with GPT-4, 4 months later.
True enterprise adoption is still many quarters/years away
ChatGPT’s launch was the starting gun for mainstreaming that AI is a big deal in terms of new capabilities and kicked off the large scale enthusiasm, hype cycle, and adoption for generative AI. This launch was only 8-9 months ago, and GPT-4 did not come out until 5 months ago. Given that large enterprise planning cycles often take 3-6 months, and then prototyping and building will take a year for a large company, we are still very far away from peak AI usage or peak AI hype. Most large enterprises are still trying to analytically sort what “AI” means for them, and are still many quarters from embracing this new technology.
4 Waves of AI Adoption
Indeed, there are likely at least 4 waves of AI to consider in these early days.
Wave 1: GenAI native companies. ChatGPT, Midjourney, Character.AI, Stable Diffusion, Github copilot, and other early launches that have now gained significant revenue and user traction. Obviously there are some great ML companies that pre-date GenAI that continue to participate in the current era (Hugging Face, Runway, Scale, WandB are a few that come to mind).
Wave 2 (current wave): Early startup adopters and fast mid-market incumbents.This is the first wave of startups to launch on top of GPT-3.5/4 like Perplexity, Langchain, Harvey or others. In parallel, a small number of founder led multi-billion companies like Navan, Notion, Quora, Replit, and Zapier launched AI-powered products quickly and are the early adopters of the wave. Microsoft, Adobe, and Google are notable outliers as very large enterprise moving fast to AI - Microsoft likely due to its inside track with OpenAI, and Adobe as diffusion models tend to be cheaper and simpler to train than the large scale LLMs.
Wave 3 (coming soon): Next wave of startups currently being founded. It will be exciting to see what is in this mix and may include new formats like voice and video in addition to using natural language in more verticals and more ways, as well as new types of infrastructure. Companies like Eleven Labs/LMNT/LFG Labs, Braintrust, and many more will provide incremental experiences. There is a big wave of new startups coming. The current YC batch alone appears to have a 100 or more AI startups….
Wave 4 (coming 2024/2025?): First big enterprise adopters. Since enterprise planning and build cycles are so long, anticipate the first really products (versus demos or prototypes) from larger companies other than MSFT, Adobe, Google, Meta to start to show up in a year or two. This is when revenue to AI infra companies will start to ramp significantly relative to today, when hype will peak, and we will see further accelerated investment in AI.
Published Aug. 24, 2023
By Andrew Hutchinson Content and Social Media Manager
Among the various use cases for the new slate of large language models (LLMs), and generative AI based on such inputs, code generation is probably one of the most valuable and viable considerations.
Code creation has definitive answers, and existing parameters that can be used to achieve what you want. And while coding knowledge is key to creating effective, functional systems, basic memory also plays a big part, or at least knowing where to look to find relevant code examples to merge into the mix.
Which is why this could be significant. Today, Meta’s launching “Code Llama”, its latest AI model which is designed to generate and analyze code snippets, in order to help find solutions.
As explained by Meta:
The tool effectively functions like a Google for code snippets specifically, pumping out full, active codesets in response to text prompts.
Which could save a lot of time. As noted, while code knowledge is required for debugging, most programmers still search for code examples for specific elements, then add them into the mix, albeit in customized format.
Code Llama won’t replace humans in this respect (because if there’s a problem, you’ll still need to be able to work out what it is), but Meta’s more refined, code-specific model could be a big step towards better-facilitating code creation via LLMs.
Meta’s releasing three versions of the Code Llama base, with 7 billion, 13 billion, and 34 billion parameters respectively.
“Each of these models is trained with 500 billion tokens of code and code-related data. The 7 billion and 13 billion base and instruct models have also been trained with fill-in-the-middle (FIM) capability, allowing them to insert code into existing code, meaning they can support tasks like code completion right out of the box.”
Meta’s also publishing two additional versions, one for Python specifically, and another aligned with instructional variations.
News of the Week
By Kate Clark
Aug. 24, 2023
The U.S. Securities and Exchange Commission approved new rules for private funds Wednesday meant to give limited partners—the individuals and institutions that invest in these funds—more transparency. The ramifications for the much larger private equity industry have drawn most of the attention, but venture capital funds will also need to bend to the new rules. VC firms will have to provide their LPs quarterly financial statements, which remarkably isn’t standard for all firms. They’ll also have to perform annual audits for each fund they manage.
Some venture capitalists, of course, aren’t happy, even though the new rules are less stringent than earlier proposals. “It’s completely wrong-headed,” said one venture capitalist, arguing that the SEC is trying to solve a problem that doesn’t exist. “It’s bad policy and it’s bad for the industry.”
The added cost and effort required to stay compliant under these new rules is fueling their frustration. Behemoths like Andreessen Horowitz already have large compliance departments who can handle these new requests easily. But small firms don’t have a single lawyer on staff.
These small fund managers are especially concerned with the SEC’s decision to require them to disclose side letters, a contract separate from the general LP agreement that can include special rights and privileges for certain LPs. For first-time fund managers, side letters are an important fundraising instrument. Using a side letter, they can offer an LP a lower management fee—the percentage of committed capital used to fund operations—if the LP agrees to serve as the fund’s anchor, the individual who provides initial funding, for example. (Such requests have become more common in the last year as LPs have demanded better terms.)
The SEC isn’t banning side letters, as it earlier considered doing. But the disclosure requirement means other LPs will likely demand the same treatment, making side letters far less useful. Without that negotiation tactic, first-time VC funds may not be able to raise capital. Emerging managers, whose ranks often include women and people of color, are already facing the most challenging fundraising environment in recent history. Additional challenges could quell diversity efforts in VC.
Still, the SEC wants to do away with preferential treatment and make sure all a fund’s investors are aware of any special terms. That’s understandable, especially when one considers just how long the VC industry has avoided most regulation. That was easier to justify when VC was a cottage industry of 1,000 firms back in 2007. As of 2021, there were more than 4,000 firms with $1 trillion under management. At that size, more transparency is necessary.
Of course, crafting regulation for the sprawling ranks of private funds, which range from tiny seed firms to private equity goliaths like Blackstone Group, which has $1 trillion in assets under management, is going to create problems.
Nonetheless, I think creating loopholes for newcomers might just make things more complicated and allow schemers to take advantage. You have to start somewhere.
This may be the beginning of a new era for VC firms, where they’re forced to provide frequent updates about their investments and keep their LPs more informed. That’s not a bad thing. As our reporting on startup implosions from Fast to FTX to IRL over the last year has shown, they’ll eventually find out anyway.
August 23, 2023
at a whopping $4 billion valuation.
The New York-based startup allows companies to store and use AI software. It hosts hundreds of thousands of open-source AI models that developers can use for AI applications.
According to the report, the new funding values Hugging Face at more than 100x its annualized revenue.
The new round doubles the startup’s valuation from May 2022, when the company raised a $100 million round at a reported $2 billion valuation.
It was just in March when Salesforce Ventures said it would launch a new $250 million generative AI fundthat will look at promising generative AI startups. The firm announced at TrailblazerDX it will initially invest in four AI companies — Anthropic, Cohere, Hearth.AI and you.com.
Three months later, it said it would double its fund size to $500 million, and based on the recent Hugging Face news it plans on deploying that capital.
While most associate Salesforce with sales, the CRM giant has been adding software development tools for years, so the Hugging Face investment seems to make sense.
Aug. 21, 2023 5:00 PM PDT
Here’s a declaration the tech press doesn’t get to make very often: It’s a good day to be SoftBank. Arm, a chip designer the SoftBank Group has owned since 2016, finally released its investor prospectus this afternoon, a key step before the initial public offering expected next month. If investors bite—no sure thing, as I explain below—Arm’s valuation could surpass $60 billion. That’s roughly double the price SoftBank paid seven years ago to take the semiconductor firm private.
But it’s not all glory for SoftBank—the striking caveat is that this Arm investment could have turned out much, much better for the Japanese conglomerate. Consider this mammoth missed opportunity: Its planned 2020 deal to sell Arm to Nvidia would have given SoftBank a roughly $25 billion stake in Nvidia that would now be worth $93 billion after the chipmaker’s incredible stock run-up this year helped it surpass $1 trillion in market capitalization. Instead, regulators sued to block the deal over antitrust concerns and Nvidia pulled out, nixing SoftBank’s potential stake in an even bigger semiconductor success story.
The eye-popping 275% return that wasn’t would have more than made up for notable disasters in SoftBank’s tech investing strategy, such as putting money on beleaguered WeWork, into which SoftBank has said it sunk more than $18 billion. (That fortune—more than the GDP of Nicaragua—might be worth next to nothing now, as WeWork warned this month it could go bankrupt.) Arm’s IPO, which is likely to be the biggest U.S. tech listing since Alibaba’s nine years ago, became SoftBank’s backup plan after the Nvidia sale fell through. “Plan B is actually not bad at all. It could be a better plan for us,” SoftBank CEO Masayoshi Son said, through an interpreter, on an earnings call last year.
The foiled Nvidia-Arm deal isn’t SoftBank’s fault, to be sure. But it was actually the second time SoftBank had missed a chance to get exposure to Nvidia’s more recent growth. In 2017, a year after SoftBank bought Arm, it also bought a 5% stake in Nvidia. It sold that stake a year and a half later. The company said at the time it had recorded a $3.3 billion gain on the investment, but it would have been worth tens of billions more had SoftBank held onto it.
What Could Trip Up Arm’s Bankers
Whether this history will haunt SoftBank largely depends on how well Arm does in the public markets. Arm’s financial story is mixed, its IPO filing revealed. The company reported declining revenues and narrowing profits amid broader struggles in the smartphone market. Still, it generates hundreds of millions of dollars in free cash flow a year.
The price at which SoftBank might be able to sell Arm shares could get a boost, fittingly, from Nvidia. That firm, in addition to businesses like Broadcom and AMD, is among the group of public semiconductor companies investors use as valuation comparisons for Arm. Nvidia, whose graphics processing units have become crucial AI infrastructure, trades at 45 times its last 12 month’s sales, a significant premium over other semiconductor firms. AMD, for instance, trades at 7.8 times that same multiple, while Broadcom is at 11.
Arm, which isn’t expected to grow as fast as Nvidia, is unlikely to get a multiple anywhere near Nvidia’s. Its AI-related story is more complicated than Nvidia’s. Yes, Arm argued in its filing that its own central processing units “will be central to this transition” toward AI, because it sells to firms like Nvidia, Meta Platforms and Cruise. But Arm also acknowledged that’s not a sure bet, a reality companies have to own up to in a legal filing. “New technologies, such as AI and ML, may use algorithms that are not suitable for a general purpose CPU, such as our processors,” the company noted.
At a $60 billion valuation, Arm would be getting a trailing revenue multiple of 23—between AMD and Broadcom on the one hand and Nvidia on the other. Given Arm’s growth rate, that multiple looks a bit steep.
Gené Teare @geneteare
August 23, 2023
After nearly two years, the tech IPO market may finally be reawakening, setting the stage for a broader rebound in 2024.
The performance of the two very different highly valued and well-branded companies — Arm, which benefits from the demand for processing power for AI, and Instacart, a grocery delivery service that experienced high growth during the pandemic — will be a test of investor appetite for the next wave of technology IPOs.
While those IPOs could rejuvenate a stagnant market, the bar is also much higher now for startups hoping to ascend to the public markets than it was a mere two years ago, when more than 350 venture-backed companies went public in the U.S.
Investors are asking “are these companies ones that could actually become profitable, or are they companies that continuously need additional capital injection?” according to Ran Ben-Tzur, a partner at law firm Fenwick & Westwho advises technology and life sciences companies planning to list on the public markets.
So far this year, there have been just 43 IPOs for U.S. venture-backed startups. More than 1,400 companies that have collectively raised close to $900 billion from private investors remain on The Crunchbase Unicorn Board, a global list of private startups valued at $1 billion or more.
IPO pipeline thaw?
Arm and Instacart — along with several other possible 2023 debuts including marketing email automation company Klaviyo — could prompt some of those startups to dust off their IPO plans too.
Arm’s Nasdaq listing is expected to be the largest IPO since Rivian’s public debut in late 2021. The SoftBank-owned chip designer filed its IPO prospectus on Monday and is reportedly aiming for a listing valuation of between $60 billion and $70 billion.
But perhaps more relevant for the startup world is Instacart’s expected IPO. The grocery delivery company, last valued by venture investors at $39 billion in 2021, is also predicted to file its IPO plans this week for a September debut.
Now more than a decade old, Instacart is one of the better known technology companies to face the public markets since 2021. It’s one of the 10 largest private unicorn companies based on its 2021 valuation, (though its more recent internal valuation, reportedly slashed to $13 billion, is a more accurate indicator of its current value as it prepares to file).
Instacart’s valuation reset is also indicative of the reality check facing many heavily funded startups in the aftermath of 2021’s bull run.
“Amongst our client base we are seeing companies doing down rounds, or liquidating or doing fire sales. So I think there will be a lot of fallout,” said Ben-Tzur, whose company also advises on private funding and M&A. “There’ll be a lot of companies that might have been viable IPO companies a couple of years ago that just aren’t in today’s market, and they’ll need to sort of figure out what their next step is,” he said.
While investors don’t necessarily expect profitability at IPO, companies should plan to show net income within six quarters of a listing, he said: “You’ll see companies even sacrificing some growth to show profitability.”
Mark to market
Ben-Tzur worked on a number of public listings during the 2021 IPO market peak, including Coinbase, GitLab and SentinelOne — all of which are currently trading below their IPO price amid a general market slump for newer public companies.
A Crunchbase News analysis shows that roughly 60% of 2021’s billion-dollar listings in the U.S. are now valued below $1 billion. Things look even worse for last year’s IPOs: 78% of 2022’s billion-dollar listings are currently valued below $1 billion.
The market would be more receptive toward new issuances if the 2021 cohort were valued closer to their debut valuations, according to Ben-Tzur. “They are trading well above their lows but not anywhere near their highs at the moment.”
Companies that will list first “have spent the last 24 months right-sizing their business and focusing on the right KPIs that public market investors would likely focus on,” said Larry Aschebrook, managing partner of Chicago-based G Squared, a growth investor with $5 billion under management.
Ultimately, each company that goes public is a test of current market values. Just as a blockbuster debut can breathe new life into the market, a highly anticipated stock that falters can torpedo it.
Back in 2012, Facebook “had a very rocky introduction to the markets, and that really actually impacted the entire market for a period of a good six to nine months,” said Ben-Tzur, who worked on the social media company’s IPO early in his career.
Startups as well as public and private investors are again unsure what the current climate is, Aschebrook said. “You need a few brave folks to test the market.”
Which companies follow Arm in going public?
AUG 22, 2023
IPOs have been few and far between in 2023, especially in the tech sector. Arm filed their F-1 prospectus yesterday in what will almost surely be the largest tech IPO in 2023.
Their filing and subsequent IPO could mark a gradual opening of the IPO window, although some companies might wait to see how their IPO fares.
This week, I discuss some of the companies that are expected to go public relatively early on in this next cycle later this year, or more likely late H1 next year to truly “reopen” the tech IPO window.
At a high level, some of the factors that determine which companies might be the “leaders” in reopening the IPO window include:
How much of a household name / category-defining the company is. I suspect the first few out are more likely to be ones that are well-known names or at $500M+ in revenue.
Their need for capital which depends on their current cash reserves and burn.
Their need to provide liquidity for employees and investors (especially if they’re facing issues with employee options/RSUs expiring, etc)
With that said, here are a few that are candidates to follow Arm in going public in the next 12 months.
Instacart, which met with over 50+ investors last year, ahead of a potential IPO, was forced to put its plans on hold back then.
“The markets are still extremely tumultuous making it highly unlikely that an IPO is possible for us in 2022,” - Fidji Simo, Instacart CEO
Now, they’re back at it. As The Information and others have reported, Instacart could file to go public as soon as next week, and the IPO could be as early as September this year.
Instacart was last valued at $39B in the private markets but has cut its internal valuation used to give grants to employees 4 times, most recently to $10B at the end of 2022.
But with $1.4B of revenue and 30% revenue growth, in large part to their advertising business in the first half of 2023, it will be hoping to go public closer to its original private mark.
Klaviyo is at a rumored 500M+ in ARR, and one of the leaders in the marketing automation category. It confidentially filed to go public in May of this year (one month after Arm), and so is a good candidate to be one of the earlier ones out of the block. Klaviyo tapped Goldman to take it public, and a date of later this year was even floated at the time of its filing.
Whether it goes out in 2023 will depend on the Arm IPO, but even if not, I expect it to be one of the first few out next year.
At $1B+ in revenue, Databricks would be a good candidate given its scale and notability.
Databricks does also likely need capital, given relatively high losses of 900M over the last two years, but on the flip side has not had any issue tapping the private markets for that capital and is rumored to be doing so again, per The Information.
Databricks also has no intention of being the first, as noted in the FT:
Databricks chief executive Ali Ghodsi has repeatedly stated his intention to take the company public but Databricks “won’t be the first out”, according to a person with knowledge of its plans. The company “will watch and see how [Arm’s IPO] goes,” they added.
However, given its scale and the current excitement around AI, I still expect it to go out next year, assuming a reasonable Arm IPO.
Reddit hired its first CFO in 2021, and filed confidentially to go public in early 2022. Markets derailed that plan, but Reddit would have likely been a good candidate to be early out of the blocks when an IPO window opened.
The recent changes to Reddit’s API and the backlash that followed which resulted in 30%+ of subreddits going dark might change that perhaps. It highlights both Reddit’s need to better monetize and also the fragile position it finds itself in relative to its community.
But when you consider some of the language below, its clear that the need for both profits and an IPO loom large.
We’ll continue to be profit-driven until profits arrive. Unlike some of the 3P apps, we are not profitable. - Reddit CEO, Steve Huffman
For that reason, I still expect Reddit to go out later next year if the markets look good, almost 20 years from its founding.
Stripe is in many ways potentially the most well-known private tech startup, and many have suggested that it missed a trick by not going public in 2020/2021.
Given their recent raise at a 50% discount to their previous valuation, done mostly to ensure early employees don’t get screwed over in terms of their RSUs expiring (I explained this in a previous post), they have a little less time pressure to go public immediately.
In addition, with Adyen, its closest comparable recently getting crushed after earnings and being down 50% in a week, Stripe is unlikely to be itching to go out yet.
So I don’t expect them to be early in the pack, but a Stripe IPO, if it were to happen in H2 2024, would likely signify that the market is in a good state, and given the “missed” opportunities, they are likely to want to go out the moment the time feels right.
Aug. 24, 2023 5:00 PM PDT
Bankers are busy. IPO lawyers are hard to reach. It’s go time again for Silicon Valley IPOs.
Arm, Instacart and Klaviyo are preparing their offerings, with filings from the last two expected tomorrow. Beauty company Oddity and restaurant brand Cava have gone out and are trading above their IPO prices. Companies are hiring chief financial officers who can take them public. And venture capitalists are excited that the market is open again.
But there’s one problem. Many founders still aren’t convinced it is a good time to go.
First, there is valuation. Several of the founders I know say they aren’t excited to go public at a potentially lower valuation. Public market investors may be interested in these businesses, but not at 2021 prices. That’s still a tough pill for founders to stomach.
Good tech companies also need the money less. They worked hard to improve their businesses and cut costs over the last year or so. So if they don’t need the money, why take it?
Personally, I think the second bucket of reasons offers good justifications for not going public, but the first doesn’t. Valuation today matters a lot less than valuation five years from now.
I have heard so many CEOs I admire, Reed Hastings among them, say that being a public company was immensely valuable to their businesses and how they operated. A lot of companies would benefit from taking the plunge sooner rather than later.
Since I won’t convince them all, be careful about predicting a windfall of new offerings. Venture capitalists will want to make it seem like there is a flood. They want to show their limited partners that they will eventually make them some money.
But you need supply as well as demand. And from what I’m hearing, that supply is still (pretty) stuck.
Chamath Palihapitiya took retail investors for a ride, got away with it, and just can't let himself take the win.
AUG 23, 2023
In any just world, Chamath Palihapitiya would be ashamed of himself.
He lent his reputation to a slew of companies going public via his special purpose acquisition companies. He was the ruinous SPAC king.
Now, almost a year after calling two SPACs quits, he’s still in denial that he was the pied piper who enticed retail investors into betting on speculative, money-losing companies.
He’s like a bully who stole someone’s lunch money and then says “stop crying about it” — except it’s for all the world to see.
Palihapitiya is dunking on people who say that they lost money because of him.
He has said he made roughly $750 million by throwing his reputation behind the stocks, taking them public via SPACs that he helmed, and then selling shares on the public markets. He hyped the stocks, he sold his shares, and he made a profit while the retail investors who trusted him lost money.
And the reality is that Palihapitiya got away with it. He’s boastfully summering in Italy and reveling in his luxurious wedding on the All-In Podcast that he hosts. Elon Muskand Grimes were apparently in attendance.
The media has actually been pretty gentle toward Palihapitiya given the epic stock depreciation of the companies he took public. Maybe it’s because outlets like CNBC played their own part in hyping up the SPAC fiasco. Or maybe it’s because SPACs were such a transparent grift from the beginning. Or maybe it’s just because Palihapitiya is fairly affable and inspirational and it would be nice in this cynical world to retain some heroes. (David Sacks is supposed to be the mean Republican and Palihapitiya the virtuous Democrat.)
To me, the SPAC mania was such an obvious cash grab and ego play that I couldn’t get all that upset about it when it crumbled. The writing had been on the wall from the beginning.
I profiled Palihapitiya back in November 2020 just as some of his SPACs were spinning up. I talked to him on the record for the story and titled the piece “The Man With Six SPACs.” It was obvious then that this was a game of regulatory arbitrage: SPAC sponsors believed that they could hype up companies, talk about future performance, and make money even if the target companies that they acquired fell in value. The companies that Palihapitiya was promoting were not top tier Silicon Valley startups. Companies like Opendoor and Clover Health were risky, money-losing companies. And I wrote in that 2020 story that sponsors could make money even if the companies they backed fell in value on the public markets.
Here was the billboard quote from investor Jeremy Levine: “The SPAC thing — that is going to be a crazy bubble. It will look like the 1999 internet bubble where basically all of those companies went to zero. There will be some good ones. Most of them will be true disasters.”
We knew. People should have known.
Palihapitiya was full of hype and bombast. He compared Metromile — one of his SPAC companies that ultimately saw its stock tank before being acquired in a fire sale — to Warren Buffett’s bet on Geico.
When Palihapitiya’s SPACs — which included Virgin Galactic, Opendoor, Clover, and SoFi — proved to be disasters and in September 2022 he gave up on two other SPACs he had planned, I kind of shrugged my shoulders. This was an obvious grift the whole time. It wasn’t worse than GameStop, initial coin offerings, or any number of crypto scams. If I spent all my time getting outraged about people who made money off empty hype, I would be in a constant state of fury. Plus, Palihapitiya was one of my favorite characters on my favorite business podcast.
Sure, I badgered Jason Calacanis that he and the other co-hosts should give Palihapitiya a harder time about the SPAC performance on the show. But I largely decided not to dance on the grave of Palihapitiya’s lucrative SPAC escapade.
From my vantage point, Silicon Valley’s pivot to SPAC mania was part of a broader abandonment of the venture capital industry’s principles by money-hungry pockets of the industry.
Venture capital was, I thought, supposed to be all about alignment of incentives — putting limited partners, venture capital firms, and founders on the same path to financial success. Everyone gets rich when a company actually works and can prove itself to the public markets.
Instead, after years of low interest rates, high private valuations, and weakening metrics, unscrupulous startup world players identified plenty of ways to get rich off momentum on the private markets:
Founders sold private shares, netting fortunes before their companies ever turned a profit or went public.
Investors flipped shares to later stage investors, locking in wins. They raised enormous funds, reaping management fees. They split off sidecars, insulating certain deals from the overall performance of their fund.
Limited partners dumped money into crossover funds who were playing a different game entirely, further distorting venture capital firms’ incentives.
All this money-making in the private markets was creating more and more pressure to figure out a path for speculative, money-losing startups to go public. And suddenly Silicon Valley seized on the special purpose acquisition company. SPACs could take startups public while projecting all sorts of wild financial performance into the future.
It was a financial vehicle to square the circle. The unicorn hordes needed to go public to justify massive private valuations— but many companies were in no shape to do so via a traditional IPO. Public market investors valued companies based off of actual financial metrics and while they were willing to give companies high multiples during the boom times there were limits.
But suddenly because of an apparent hole in the financial regulations, SPACs gave startups a window where they could claim that they would generate profits five or ten years from now — and so public market investors could credulously value companies relative to that forward-looking guidance even though those estimates were in many cases totally outlandish.
Palihapitiya simply stepped into a world that was desperate to be told these startups were good and that there was money to be made riding the momentum train. The private markets had been doing it for years. Why should retail investors be denied their time at the roulette wheel?
And so Palihapitiya gave the world what it wanted. He pitched the companies on CNBC, retail investors poured into the companies, Palihapitiya unloaded his shares, and slowly reality set in and the shares of the companies continued to fall.
So like I said, he got away with it. (Though he is facing lawsuits.)
But Palihapitiya can’t take his win at the regular person’s loss with grace.
Yesterday, he started dunking on people on social media, proclaiming himself (for not for the first time) the man in “the arena.”
Director of Product at LinkedIn, Angel Investor
August 23, 2023
One of our favorite tools we offer members on LinkedIn is newsletters — since launching in 2017, we’ve seen more than 365M total newsletter subscriptions on LinkedIn.
But the value of LinkedIn newsletters is really about sparking conversations and sharing knowledge with your community. We’re seeing people reading and engaging with newsletters more and more: to learn, explore a topic they’re interested in, and get better at their jobs. In fact, newsletter readership has tripled over the past year, now with more than 1.3M daily readers.
To continue to help you get more from newsletters, we’re excited to announce a handful of new features:
A new look and feel when publishing a Newsletter
Writing a newsletter takes time and effort, and we’ve heard that the article editor platform was a bit clunky, so we’re introducing a new, smoother editing and publishing experience.
The revamped article editor is purposefully designed to offer a seamless and dependable experience, allowing you to completely focus on sharing your insights and expertise. It's now easier than ever to format, layout, and add a mix of photos, videos, links and more.
You can save your articles in progress as drafts to seek a second opinion and gather feedback from others before it goes live. With customization options, you can schedule your article to publish when your community is most active and add a SEO title and description for even more reach.
Host multiple newsletters to expand your audience
Most experts have knowledge in a variety of topics, and the audiences for each might be different. That's why we're rolling out the ability to host multiple newsletters — you no longer need to pick just one interest area. Any member or Company via Pages can now host up to five newsletters in one place with different topics, design and frequency of posting for each newsletter to strengthen relevance and engagement.
Improve your discoverability with auto-follow
To broaden your reach and discoverability with relevant audiences now when a member subscribes to your newsletter it will trigger an auto-Follow from your Profile. This will help grow your community of engaged subscribers who have already shown interest in your content to stay up to date with all your updates including Feed posts, LinkedIn Live events, etc.
This builds on our recent updates to help boost your reach and grow your subscribers including one-click subscribe and newsletters appearing in search. With just a click, existing and potential readers can subscribe to your newsletter directly from LinkedIn or anywhere else you choose to share the link on social, email, or anywhere on the web. Additionally, when members search for you, they'll see your newsletter right in the results and can one-click subscribe.
If you’ve been thinking about authoring a newsletter, now’s the time. Pick a topic of interest and give the new editor a try to start sharing your insights! Learn more about starting a newsletter here.
These new features are all starting to roll out this month. As always, we look forward to your feedback and can’t wait to see the content you create.
By Martin Peers
Aug. 23, 2023 5:00 PM PDT
Nvidia CEO Jensen Huang should be taking notes about his daily life. This is sure to be a year he’ll want to remember in detail. Nvidia’s dominance of chips that are vital for generative AI is translating into growth that is extraordinary for a 30-year-old company. Not only did it report better-than-projected second-quarter revenue growth of 101% on Wednesday, Nvidia projected third-quarter revenue that would be up 170% on a year earlier. By the time the fiscal year ends next January, Nvidia should have brought in north of $50 billion in revenue, nearly double that of last fiscal year and nearly 5 times its annual revenue in fiscal 2020.
The surge is flowing through to Nvidia’s bottom line. Its net profit margin hit 46% in the quarter, compared with 10% in the year-earlier quarter. Just as a comparison, Intel hasn’t reported a net margin higher than 31% in the past 32 years, according to S&P Global Market Intelligence. (For more on Nvidia’s quarter, see here.) As my colleagues Jessica Lessin and Stephanie Palazzolo have written in the past couple of days, the luck might not last. Developments in AI technology may give rival chips more of a chance, for instance. And most of the big tech firms, including Amazon, Google and Microsoft—all currently Nvidia customers—are developing their own AI chips to replace Nvidia’s product, as we laid out in this story in May.
Not surprisingly, Huang is optimistic. He argued today that data centers are going through a major shift in computing that will underwrite long-term growth for Nvidia. He has proved skeptics wrong before. Back in 2017, we noted in this Huang profile that while investors saw Nvidia as well positioned to benefit from AI’s potential, competition from companies like Google building custom chips for AI posed a threat. As it turned out, though, Nvidia’s chips turned out to be exactly what the industry needed. Still, no company can maintain a 100% growth rate forever, as Snowflake, which also reported today, demonstrates. (On that, see below.)
Is Nvidia stock too richly valued now? That’s hard to say. The stock, which was going for well above $500 in after-hours trading, is up more than 240% year to date. It’s currently trading around 22 times forward sales, according to S&P, more than twice the multiple at which the next most highly valued chip stock, Broadcom, is trading. But Nvidia’s multiple is based on analyst revenue estimates for the next 12 months prepared before today’s earnings. Revised estimates will likely bring that multiple down as analysts ratchet up their estimates. But given all the variables of competition, and the potential constraints on Nvidia’s ability to meet demand, any estimates are guesses at best.
Back to Reality
Snowflake has returned to Planet Earth. Gone are the days of 100% revenue growth the data and cloud company routinely reported as recently as a year and a half ago. Snowflake reported today top-line expansion of 36% in the second quarter, although it forecast a slowdown in its product revenue—which accounts for most of its revenue—in the current quarter.
On a call tonight with investors, even the analysts were reminiscing about the good old days. One of them asked when software companies could expect a revenue bump from AI, pointing to the feverish impact it has had on Nvidia’s growth. CFO Mike Scarpelli suggested it won’t come until next year, given how long it is still taking companies to get their hands on the much-coveted GPU chips Nvidia makes and incorporate them into their data centers.
The good news for Snowflake investors was that, despite predicting a slowdown in the third quarter, the company didn’t lower its 34% product revenue growth rate forecast for the full fiscal year. Three months ago, a downward revision in fiscal year guidance spooked investors, sending Snowflake stock down more than 16% the day after its first quarter results.
“We're not seeing customers reduce their consumption right now,” Scarpelli said on the call, earlier noting “we are seeing encouraging signs of stabilization, but not recovery.” It seems stabilization is all investors were looking for, sending shares up roughly 4% after hours.—Akash Pasricha
Startup of the Week
I recently made an LP commitment to an all-women run fund called Pact VC; more info on them below. They've already done first close and have £20M committed -- if you invest in Seed Funds and want exposure to Europe. Reem will be in the Bay Area 28/29 September. Please let me know if I can make the connection.
Very experienced investors; all 3 have operational backgrounds:
The UK's first all female GP
LPs include Head of AI at Google, C-suite from Naspers, Alibaba, T.Rowe Price
Investing in early stage fintech, healthcare, climate in Europe
GPs have been founders, operators and bankers in tech for decades
Global network and experience in 10+ countries
X of the Week
SignalRank’s inaugural Series B ecosystem report