There is little dispute that AI has created the most significant technology platform shift since the advent of the cloud for software. OpenAI’s ChatGPT launch in 2022 was not unlike the seminal moment when Salesforce went public in 2004 (although to much less fanfare). Since then, there have been trillions of dollars of value creation in public market SaaS. Similarly, these market-leading companies are more valuable today than ever (more below), and we’re at a pivotal moment in SaaS. AI could potentially render the current SaaS model of selling vertical or horizontal, configurable workflow software at 30% free cash flow margins into perpetuity, a thing of the past. In an overly simplistic way, selling seats (or consumption) to knowledge workers to click, see data / dashboards, and process workflow on computer screens and/or mobile devices should and can eventually be disrupted massively by AI through agents. As a result, business models could be upended. The best SaaS companies make their money with high structural retention and upsell. Change management is a real issue; moving from one software vendor to another is hard work, particularly for large companies with embedded processes, existing data (data gravity), and built-out integrations. It’s even more challenging to move from system of record platforms, which comprise the most successful public SaaS companies. However, it could become much easier with AI since agents could theoretically manage the change for you. Taking this a step further, if AI can disrupt the delivery of the software i.e., no longer workflow on a screen and a database backend, as well as the output, i.e. agents are doing the work, pricing models will also be disrupted. Seat models could die out as the agent is the application, workflow, and, therefore, the atomic unit of value (and this is just when the software is sold!). The structural moat for a SaaS business has generally been – again in a very simplified fashion – a great, product-driven CEO with domain experience, engineers, and product managers to build technology that solves an acute business problem with a vast surface area in a large, greenfield market bound together with structural moats such as data gravity, integrations, and simple inertia. The go-to-market (GTM) team markets, sells, and upsells, and revenue growth compounds when more products are built and sold and usage increases. This model can produce significant operating profits and free cash flow over time. However, what if the moat to create great software became eradicated as AI agents build the software while simultaneously offering the agent workflow? Or hyperpersonalized software experiences? Or by abstracting away the UI entirely? If creating the software becomes commoditized, then pricing will compress (and so will gross margins and profits / free cash flow).
This is an extreme bear case, and the reality will likely differ significantly. Many savvy investors have written about these various points, and this potential wholesale disruption of SaaS has created much fear and anxiety. Will the 100+ public SaaS companies (as well as the hundreds of private unicorns) today go the route of some of the networking or broadband companies where business models were upended? One could argue that SaaS has had its heyday for 20+ years since the Salesforce IPO. Could the oversimplified fact of delivering workflow and analytics software over the internet to knowledge workers through subscriptions and creating trillions of dollars in value come to an end due to the massive AI platform shift?
From a financial perspective, the value of a SaaS company is the net present value of future cash flows. SaaS companies pay to acquire a customer upfront and recognize this annuity stream of revenue and cash flow over time. Moreover, the best companies have high retention rates and net dollar retention of well over 100%, which creates massively profitable cohorts over time. The outcome for market-leading companies can be durable growth at scale, with gross margins upward of 90% and free cash flow margins upward of 30%. It’s no wonder that when running SaaS companies through a DCF (discounted cash flow analysis), the net present value of those cash flows spits out numbers that value the best / market-leading businesses at 15-20x+ NTM (next-twelve-months) revenue today as a proxy for those future cash flows.
The timing of this platform shift, with the massive decrease in revenue multiples and slowdown in revenue for public SaaS in the post-ZIRP (zero-interest-rate-policy) market, compounded the idea that “SaaS or software is dying”. This is untrue.
AI and agent technology is still early, yet it has enormous potential to disrupt traditional SaaS workflow software business models and even core systems of record. While not widespread, at least one company, Klarna, has announced they have churned from core systems of record in Workday and Salesforce and are rebuilding the functionality using OpenAI as a backend. For the time being, this is likely to be a rare circumstance – very few companies have the resources, expertise, and willingness to take on the risk of rebuilding massive third-party applications and workflows offered by the likes of Salesforce and Workday. Agents, while they provide extraordinary promise, are in the early stages and are generally not ready for industrial production scale. With that being said, agents will only improve, and agents will undoubtedly become part of the offerings from current SaaS platforms and create new companies. Agents will also open up more market opportunities as they will fill the white space not offered by the status quo SaaS platforms. The surface area of software will eventually consume every workflow – both professional and personal – and agents could enable this to happen quicker than anyone ever thought possible. It’s important to note that this platform shift differs from others in the past; companies and technologists are heralding AI and agents as the next greatest thing, i.e., there are no “naysayers”. During the move to mobile and the cloud, there were skeptics. Those naysayers are fewer and farther between in the age of AI, and legacy SaaS vendors like Salesforce are pushing just as quickly toward AI as the next batch of YC companies are. In the end, this should be great for both buyers and users of software as there will be more and better options, and the software markets will grow and become more valuable. Massive companies will be created from both standalone and new companies, and existing ones can still take advantage. Many will not make it and will be disrupted, like with any new platform shift. Disruption is likely to happen unevenly across industries and time.
This transition could be similar to how the move to the cloud helped some incumbents (many didn’t innovate, too). Adobe, for example, launched their cloud product in 2009, grew run-rate subscription revenue (a proxy for ARR) from ~$50M → ~$21B over 15 years (!), and created almost $220B in market value*. For reference, in 2009, Adobe’s total revenue was around $2.9B, about 1/7th of its current run-rate subscription revenue. You can see how well Adobe took advantage of SaaS below, and we will likely see other SaaS incumbents have charts that look like this but with AI revenue.
Adobe Cloud ARR vs. License Run-rate ($M)
Source: Adobe public filings
The companies analyzed below in this post could also similarly take advantage if they innovate on their technology as many own large and durable distribution channels. While companies that don’t integrate or innovate with AI in their stack can be disrupted, the ones that take advantage could be even bigger as their TAMs (total addressable market) will be expanded dramatically. One could also argue that AI will be just a part of the evolution of software as even “SaaS” or “cloud” has evolved from software → on-demand → SaaS → Cloud → and now AI. For the next generation of market-leading companies, AI / agents will be infused into the fabric of every piece of their software and the lines will blur on what is SaaS or cloud or AI. The image below shows how a select group of public software businesses have described themselves over the years in the first sentence of their S-1’s. AI is just a part of this software evolution, albeit a huge one, and not just in nomenclature only.
Source: Public company filings
The reality is that SaaS isn’t dead, and AI could make it bigger. We analyzed the entire public SaaS universe over the past ten years and looked at the top 10 companies by year. We defined the yearly top 10 as the highest valued revenue multiples companies (EV / NTM revenue) for the most days in each calendar year. We then took each company’s average for all days in the calendar year and the total median of those metrics across the top 10. The following series of analyses will show that it’s never been a better time to build a market-leading SaaS company in the public markets. They are bigger, more valuable, and more efficient than ever, and AI, while it has the potential to disrupt those who don’t innovate, could massively accelerate value creation.
SaaS Top 10 Median Market Cap and ARR Multiples ($M)
The median market cap has risen ~eleven times from ~$4B 10 years ago to ~$43B today. Multiples went up to over 50x ARR in 2021 and are now back at 17x. Moreover, these companies have grown through massive multiple compression and are larger today than in 2021 on a market cap basis, when they had record-high multiples. While multiples are significantly below their 2021 highs, they are still much higher than 2014 → 2018 (~50% higher), while interest rates are up considerably. This implies that investors are assigning even more value to market winners today than ever before.
Source: CIQ as of 16-Sep-2024 and company filings. See footnote on final paragraph for calculation detail of annual metrics.
SaaS Top 10 Median Growth-adjusted Revenue Multiples and 10-Year Treasury
While EV / ARR multiples are important, the following looks at the median growth-adjusted multiple and the 10-year treasury yield. Growth-adjusted revenue multiples are a similar concept to the classic PEG ratio (Price/Earnings-to-Growth), which compares a stock’s price/earnings (P/E ratio) to its earnings growth rate. In our growth-adjusted multiples, we compare an NTM revenue multiple against forward revenue growth rates to understand what the market is willing to pay for a unit of growth. The higher the number, the more “expensive” a company is relative to its growth rate. The lower the number, the more “undervalued” it is. You can read more about growth-adjusted multiples here. Even as interest rates have increased, the market is willing to pay more today for a unit of growth today than from 2014 → 2019 (up over 100% from 2014!). This also has downstream effects on the private markets, as investors are much more likely to pay more for a unit of growth today than before 2020, creating a more aggressive fundraising environment. Even though multiples have been down massively since 2021, perceived market leaders are more highly valued than in 2014 → 2018.
Source: CIQ as of 16-Sep-2024 and company filings. Daily growth-adjusted revenue multiple calculated as enterprise value divided by NTM revenue divided by NTM revenue growth rate. See footnote on final paragraph for calculation detail of annual metrics. 10-year Treasury Yield is a median of all days.
SaaS Top 10 Median ARR, Rule of 40, and Year-over-year Growth % ($M)
As the median market cap has grown dramatically, so has ARR, up ~six times from the 2014 median. While growth rates rose and then came down, the absolute ARR scale is significantly larger. This year, the median top 10 company is almost ~$2.3B in ARR, growing 26% YoY, and is a Rule of ~50 company.
Source: CIQ as of 16-Sep-2024 and company filings. ARR represents annualized revenue run-rate and calculated as current quarter total revenue multiplied by four. NTM Rule of 40 calculated as NTM (last-twelve-months) revenue growth plus NTM (last-twelve-months) free cash flow margin. See footnote on final paragraph for calculation detail of annual metrics.
SaaS Top 10 Median LTM Gross Margin, OpEx (%) & ARR per FTE ($ in 000’s)
Here is a view of the financial profiles of these businesses by year. Gross margins have increased, Sales and marketing and general and administrative as a percentage of revenue have come down, showing operating leverage, and research and development as a percentage of revenue has increased slightly. Moreover, ARR per FTE has risen dramatically, particularly in the past three years, with the increased focus on efficiency. These companies are more efficient than ever before.
Source: CIQ as of 16-Sep-2024 and company filings. All financial figures are non-GAAP, if applicable. See footnote on final paragraph for calculation detail of annual metrics.
SaaS Top 10 Companies by Year
Which businesses are represented here? The following outputs each company (ranked 1→10) by year and are highlighted to show trends.
Source: CIQ as of 16-Sep-2024 and company filings. Companies are sorted by EV/NTM revenue multiple. See footnote on final paragraph for calculation detail of annual metrics.
Here is a view of the number of top 10 finishes. Atlassian (TEAM) has been the most consistent top 10 highest multiple company, although they have fallen out of the top 10 this year. Atlassian is predominantly a seat-based model and could come under pressure as AI could enhance developer productivity and they could have fewer seats to sell to.
Source: CIQ as of 16-Sep-2024 and company filings. See footnote on final paragraph for calculation detail of annual metrics.
Final Thoughts
Software isn’t dead; it’s never been a better time to be a market-leading public SaaS company, and AI could make it even bigger. These aforementioned companies are bigger and more efficient today than at any time over the past ten years, and YTD (year-to-date), the median top 10 company, is worth over $40B. In 2014, when this index started, there were no public SaaS companies worth more than $50B; today, there are 10, representing ~$1.3T in total market cap. These valuation levels have been reached without much, to any, AI revenue (yet) although Wall Street expects it as they’re ascribed high multiples. Over the next ten years, we expect this same index of companies to comprise incumbents who took advantage of AI and AI-native companies that would not have existed without this platform shift (and to be larger than today’s $43B median!).
----------
Note: We defined the yearly top 10 as the highest valued revenue multiples companies (EV / NTM revenue) for the most days in each calendar year. We average all metrics across all trading days in a calendar year, creating an average across each company by year, and then take the overall median. A company must be trading for at least two months to be included in a year’s top 10 set. *Adobe market cap as of March 31, 2009 and September 16, 2024. Market caps are based on issued and outstanding common stock.
Thank you to my colleagues Anthony, Tanner, Austin, and Dan for their help on this post.
To receive these posts by email, click here.
No investment advice. *Meritech Capital is a current or former shareholder in Alteryx, Coupa, Datadog, NetSuite, Okta, Salesforce, Snowflake, and Twilio.