Transcription
Take a moment to look at this chart. This company went public in June of 2012 with a market cap of roughly $2 billion. Within 5 years, it had 10xed worth $20 billion. By January of 2025, it had peaked at $220 billion. That's a 100x return from IPO in 13 years. For a brief moment, it was one of the largest pureplay enterprise software companies on the planet by market value. And then this happened.
Wednesday, April 23rd, Q1 2026 earnings are reported. The stock dropped 18% in a single session. The biggest single day drop in its entire history. As of Friday's close, the market cap sits at $93 billion, down 43% year to date and down 57% from the peak. More than $125 billion in market value now in the hands of those who played their cards.
Right now, if I told you the same company just delivered 22% subscription revenue growth, that they beat guidance, raised the full year, that they're sitting on $4 billion in cash, renewal rate at 97%, and the fastest enterprise software company in history to reach $15 billion in revenue, and that 19 months into their AI strategy, they're already booking $1.5 billion a year in pure AI revenue. You'd probably think that I was reading off the wrong stock, but I'm not. This company is Service Now.
So, in this episode of the Liam Highland portfolio, we're going to do a complete deep dive. We're going to figure out what this company is actually worth and whether the market has correctly identified Service Now as a casualty of the AI transition or whether the fear has run ahead of the math. By the end, we'll have the bare case, the base case, and the bull case with the specific share prices. We'll have the exact KPIs to watch, and we'll have data checkpoints, the precise earnings dates over the next 12 months that tell us whether the thesis is winning or losing. Let's begin.
Before we get into the valuation, let me explain what Service Now actually does. In 2022, the average large enterprise ran 367 different software applications. Every department has bought its own tool over the last 20 years and more recent studies show them sitting at around 897 applications and even higher at 928 applications when you include things like the shadow IT environment. So what's the problem? None of those tools talk to each other. When something complicated needs to happen, such as onboarding a new employee, resolving a complaint that touches billing and tech support, investigating a security incident, the work doesn't live inside just one app. It threads through multiple. And historically, the way that thread happened was through humans copying data, sending emails, opening tickets, chasing for the approvals.
Service Now addresses this problem by positioning themselves as the layer that orchestrates all of that. It sits above the 897 applications and runs the workflows that connect them all. Now, I want to show you something because this is how I think about where Service Now actually sits in the broader enterprise data stack. Think of enterprise data architecture as a seven-layer stack. At the bottom, you have data origination, the raw inputs, your ERP, your CRM, your HR information system, and your IoT devices. Every enterprise has this. That's a given. Above that is storage, your data lakes, your warehouses, Snowflake and Data Bricks live here. Above that is compute, the cloud infrastructure, AWS, Azure, GCP. Then you get to the semantic layer. This is where raw data gets transformed into structured meaning where you take a million rows of transactions and turn them into a knowledge graph that understands relationships, dependencies, and context. This is where Palantir lives. And if you watch my Palantir video, this is the layer their ontology occupies. Above that is reasoning, the AI models, Anthropic, OpenAI, Google. This is the layer where the costs have dropped 99% in 2 and a half years. This layer is commoditizing fast. Above that, you then have action and orchestration. This is the layer where decisions get executed, where bots and agents actually interact with systems, click buttons, move data, trigger processes. This is where UiPath lives. And if you watched my UiPath video, this is the maestro layer. And then wrapping around everything is governance, identity, permissions, audit trails, compliance, the layer that knows who is allowed to do what across every other layer in the stack.
Service Now has positioned themselves at the top. They operate across both layer six, which is the workflow orchestration, and layer seven, which is the governance wrapper. And here's the insight that matters. The lower you go in the stack, the more commoditized the resource becomes. Data is effectively free. Storage is cheap. Compute is cheap. Intelligence is getting cheap. But governance, the institutional knowledge of how a Fortune 500 company actually runs, who approves what, which compliance rules apply in which jurisdiction, how incidents get resolved across 12 systems. That compounds over 20 years. You cannot replicate it with compute. You cannot download it from an API. Service Now has 80 billion workflows running across its entire customer base every year, processing 6.5 trillion transactions. They've got 85% of the Fortune 500 as customers. The renewal rate has been at or above 97% for five straight quarters. And this is the key point. Customers don't pay Service Now because they love the software. In fact, I see a lot of complaints about the UI, but not from the execs who actually decide what software to procure. So, those complaints don't really matter. But those customers pay because replacing it would cost more than tolerating it. And that's not feature differentiation. That's a switching cost. And remember this because when we get to Signal 4, which is the displacement debate, I'm going to come back to this exact point. Okay? So, that's what they do. That's where they sit.
Now let's talk about why the stock is down 43%. The simplest answer is that the market has stopped looking at Service Now as Service Now. It started looking at it as a category. Let me show you something. This is the software repricing matrix. You can split every major software company into four quadrants based on two questions. How much coordination does the software manage? And how directly does AI drive the revenue? We have AI native infrastructure things like Snowflake, DataDog, Cloudflare, roughly flat year to date. We have domain defenders, CrowdStrike, Palo Alto, Zscaler, down about 18%. CrowdStrike down 10, Palo Alto down 8. Painful but not existential. The disruption zone, Asana, Monday, Atlassian, Figma, GitLab, down 51% on average. Atlassian's down 58, Figma down 53. The market is saying these business models might genuinely break. The platform transformers, Salesforce, HubSpot, Adobe, Service Now, down 48%. Salesforce down 26, Adobe down 23, and Service Now down 43%. Here's what I want you to notice. Service Now is in platform transformers, but it's being priced like disruption zone. The market is not differentiating.
Makesha Aurora, CEO of Palo Alto Networks, posted this on X the other day after Service Now's earnings. Quote, "I would have expected the market to start discerning between SaaS that is impacted by AI, SaaS that needs to evolve, and SaaS that benefits from AI. The constant paranoid reaction of the market will continue to create buying opportunities for the discerning." Service Now's CEO Bill McDermott made the same point on Bloomberg.
"How much more do you think you have to buy to really shrug off, push off this bare case that's got into the investor mindset?"
"I think the bare case is really just the terminal value of software as a service companies. It has nothing to do with just Service Now."
"It has nothing to do with just Service Now. I want to pause on that concept, terminal value."
"The terminal value of software as a service companies."
Because most people scrolled right past it, and it is the single most important concept for understanding what is happening to the stock. So for those who don't know, when institutional investors value a company like Service Now, the value comes from two pieces. The first is the discrete forecast. These are the cash flows over the next 5 to 7 years, the part you can actually forecast things like revenue, growth, margins, free cash flow. For Service Now, those near-term numbers are fine. Q1 was fine. The guide raise was fine. Nobody is arguing Service Now stops generating billions in cash flow over the next 3 years. The second piece is terminal value. The present value of all the cash flows from the end of the discrete forecast to infinity. And for a company growing 20%, which is where Service Now is, terminal value typically represents 60 to 70% of the entire enterprise value. So the market is not repricing the next four quarters. It is repricing the next 40. Will Service Now still be growing 15% in year 8, 10% in year 12? Will the workflow layer still generate premium margins when AI agents can do the coordination? If the answer shifts even slightly from "probably yes" to "genuinely uncertain," the terminal value collapses. 60 to 70% of the value comes down with it. That takes 40 to 50% off the stock even though the near-term cash flows haven't changed. That is why Service Now can beat earnings, raise guidance, and drop 18% in one session. The near-term print was fine. The terminal value got repriced.
So the four signals I'm about to walk through are not just quarterly health checks. They are instruments for testing whether the terminal value repricing is justified. If the signals hold, the market panicked and the math is mispriced. If, however, they break, the repricing is rational. Signal one: we look at if the demand is real or financially engineered. Signal two: is the profit real. Signal three: can Service Now charge for agents the way they charged for seats? And Signal four: is the moat starting to crack? Let's take them one at a time.
So, first, let's look at the bare case on demand. We know we had 22% reported growth. But the argument goes like this. They had 300 basis points of FX tailwind, 125 basis points from the Armis acquisition, and 75 basis points were the Middle East delays. Strip it all out and you're at 18 or 19% organic constant currency growth. Growth is therefore decelerating and the company is dressing it up. It's an interesting claim. So, let's dig deeper. The single best test is constant currency current remaining performance obligations. It's the dollar value of contracts customers have already signed for the next 12 months but that haven't been recognized as revenue yet. If revenue is the past, CRPO is the near future. And CRPO can't be inflated by acquisitions the same way revenue would. Armis' contracts have termination for convenience provisions limiting what shows up in the reported CRPO figure. So what was it? Well, Service Now's constant currency CRPO growth in Q1 was 21%. Higher than the bears' 18% adjusted revenue figure. And here's the part that doesn't really fit the bare narrative at all. CRPO has been almost perfectly stable for the last five quarters. 22%, 21.5, 20, 21, and 21. That's not deceleration. That's stability at 21% for five consecutive quarters through every AI fear cycle the market has thrown at this stock.
The bull case looks a little further. McDermott told the market AI annual contract value or ACV is now at a run rate of $1.5 billion a year, up from a billion dollar target set at the start of the financial year. And on Bloomberg the next morning he said, "I think we'll run through that."
"And that's another thing we said we would do. 20 in 2026, a billion of net new uh annual contract value on pure AI additive to the platform. Yesterday we upped it to a billion and a half and I think we'll run through that."
Someone on the value investing subreddit actually captured it well. "In 30 months, Service Now built software that generates $1.5 billion in revenue. That's amazing." So, it's a simple test looking at the Q2 CRPO. Below 19%, the bears are winning. 19% to 21%, the base is intact. Above 21%, the bulls are winning. We've got one signal, one number, one date closes that argument.
Signal 2 is looking at whether the profit is real. GAAP net income in Q1 was $469 million. The prior year was $460 million. It grew less than 2% on a company growing revenue at 22%. The bears are not making this up. We'll come back to what's driving the GAAP. Non-GAAP operating margin was 32%, up 100 basis points. Non-GAAP EPS at 97 cents, beating the 96. On a non-GAAP basis, the leverage is exactly where you'd expect. The GAAP, unsurprisingly, with these SaaS companies is stock-based compensation, $558 million in Q1, up $90 million from prior year, inflated by Moveworks, Visa, and Armis retention equity. Is share-based compensation a real cost? Yes, absolutely. Anyone who tells you otherwise is selling something. It completely dilutes shareholders. But Service Now is buying back stock to offset it. Share count is up only 1.5% year-over-year, and the board just authorized $5 billion in additional buybacks. McDermott personally bought stock. The C-suite have halted automated selling programs. That's the strongest insider signal you can get. Then we have the Armis acquisition, 75 basis points headwind to FY26 operating margin, 200 to free cash flow margin. The full-year guide of 30% to 31.5% operating margin only expands 50 basis points because of Armis, but the company said explicitly, quote, "strong AI efficiencies from Now and Now are expected to normalize margin expansion trajectories in FY27." The dilution is guided, transparent, and one-year. And for the GAAP question, the $469 million number is real. It's why the stock trades at a forward PE of 21 and not 50. The GAAP closes if they execute. The buybacks, the share-based compensation trajectory improving, and the Armis integration. If they don't, the bears win on Signal 2. But for now, Signal 2 is intact. And our flag. There's a historical parallel becoming relevant. Hold that thought.
Signal three, we have the most forward-looking and thinnest data. Can Service Now actually monetize AI agents? The traditional SaaS model is per seat. You pay for every user. If an AI agent does the work of five seats, do you still pay for five? That's the question breaking the disruption zone companies. Asana, Monday, their revenue model is structurally exposed. Now, there's an essay from Lionus Capital called "The AI Threshold Effect" that captured this underlying dynamic. You see, durable value in AI cycles doesn't accrue to the most powerful unconstrained tool. It accrues to the constrained version, the one that wraps raw capability with permissions, audit trails, and predictable cost. Quote, "For enterprises, control is not friction. It is the product." That sentence is the entire bull case. Service Now isn't selling AI capability. Intelligence is commoditizing. Inference costs have dropped 99% in 2 and a half years. What Service Now is selling is the infrastructure that makes AI safe to run in production at a Fortune 500 company. Things like identity resolution, permissions, audit trails, cross-system orchestration, the stuff that takes a brilliant agent and makes it not a brilliant liability. Now, McDermott's line from the executive briefing, quote, "AI agents need the platform more than humans do." And why is this the case? Well, you see, humans have intuition about boundaries. Agents don't. The more capable agents become, the more they need the orchestration layer. And the data points suggesting that this monetization is working well. Creator and other share of trailing 12-month new annual contract value has gone from 17% to 21% in 12 months. The non-traditional agent-aligned product mix is growing the fastest. And the test, well, if that share of revenue stays flat or declines, that supports the bare case. If, however, it continues growing, we're in the base. And if it crosses over 25%, the bull case is clearly winning the monetization signal.
And then there's Signal 4, the one with the noisiest narrative yet the cleanest data. We have the bare case. Marc Benioff has been claiming customers are switching from Service Now to Salesforce Agent Force. AI-native startups are rebuilding the category. The displacement is real, just early. Multi-year contracts haven't come up for renewal yet. That's the case. Here's the data. Renewal rate for the last five quarters: 98%, 98%, 97% (but that one was a US federal agency closing, the underlying was still 98%), 98%, and then 97%. Through every AI fear cycle, through Benioff's claims, through the Anthropic mythos panic, if displacement were happening at scale, this number would break first. It hasn't budged. And then there's the number of customers spending over $5 million in ACV: 516 a year ago, 630 today. 22% growth in large customer count. Average contract value among those 630 grew from $14.2 million to $14.9 million. Customers are not leaving, they're expanding. And if we look at the 2011 cohort, 15 years in is at 228% of initial ACV. Through every single previous tech disruption story, the cohort data just kept on compounding. And to go back to my earlier point, customers don't pay because they love the software. They pay because replacing it costs more than tolerating it. Ripping out Service Now takes years and tens of millions of dollars. The configuration management database or CMDB, the brain of Service Now alone, was built over a decade. The alternative has to be 10 times better, not just slightly cheaper.
The shift in enterprise AI is moving the industry's focus away from function and towards interface. So you see, traditionally software was valued for its specific function, what it does. However, as AI agents begin to perform these tasks autonomously, the human-facing interface becomes secondary. The real dividing line is whether a platform serves as the governed environment where these agents act. In this transition, Service Now's value is redefined. It stops being a tool humans log into and becomes the critical secure infrastructure that AI agents use to execute work. If agents do more enterprise work in the future, they need somewhere governed to act. Service Now becomes more valuable, not less. And the test: renewal rate falls below 96%, the bears have it. 97% to 98%, we're in the base case. If it continues remaining stable, plus we have named enterprise wins against competitors, we're in a bull case.
Now, earlier I mentioned a similar pattern in the past happening. You see, late 2022, Salesforce was at $125, down 47% from the peak. Word for word, identical narrative. Death of SaaS, can't grow, activists piling in. 18 months later, it had nearly tripled. I'm not predicting Service Now does the same thing. The setups aren't identical. Service Now has been growing 20% through this panic. Salesforce was barely at 12%. Renewal data is stronger. AI monetization story is clearer. The fundamentals coming into this drawdown are actually better. But the pattern is the same. Large enterprise platforms get derated on category fears that are simultaneously real and overblown. Multi-quarter window where fundamentals are intact and price is depressed. Data catches up, market rerates, window closes. The question is whether we're in that window.
The math tells us how much it's worth. Now, let's look at the valuation. We have a market cap of $93 billion, cash at $4 billion, enterprise value of about $89 billion, trailing free cash flow of $4.6 billion, forward free cash flow of around $5.5 billion. At $90, you're paying 16 times forward free cash flow on a company growing 20%. The median enterprise software company growing at 20% trades at around 30 to 40 times free cash flow. Service Now is trading at the multiple of a slow-growing mature company.
So let's look at the cases. For the bare case, we have growth decelerating to 15%. Margin stalls, 14 times free cash flow. The share price range in this instance is around $58 to $76 per share. 20% to 30% downside from here. For the base case, growth is assumed between 17% to 19%. Margins normalize, and we have 20 times free cash flow, giving around $79 to $96 range, essentially flat to modest upside. And in the bull case, growth reaccelerates, margins are above 33%, market rerates to 35 times. You get $101 to $124, 12% to 38% upside. At $90, you're at the low end of the base range. The bull case is not paid for. The bare case is partially priced in. The asymmetry has flipped from 12 months ago when the stock was at $211 per share.
Now, let's run this analysis through a hypothetical trade. With an entry zone of $85 to $95, at this price, the math doesn't require optimism. If the stock drops to $75 to $85 and it's based on a broad software sell-off, not a Service Now specific signal, that's potentially an add, not an exit.
So, let's look at what we need to watch in order to monitor which case this company is tracking. For the bare case, there are five potential exit signals. Two or more of them in the same quarter is enough. First, we have current remaining performance obligation below 19% for two quarters. Renewal below 96%. A named Fortune 500 displacement. Discretionary insider selling at depressed prices. An FY27 guide below 16%. For the bull case, the potential signals to scale in the position are a current remaining performance obligation above 21.5% for two quarters, an operating margin above 33% post-Armis, creator share of revenue above 25%, and continued named enterprise wins against competitors. And finally, AI ACV target raised above $2 billion.
So when do we get the answers? Well, checkpoint one is July 21st. This is Q2 earnings. The most important single data point is current RPO. Below 19%, the bears are winning. Above 21%, the bulls are winning. Operating margin at 26.5%. Checkpoint 2 is in October, Q3 earnings. Then we have two consecutive quarters, and we start getting a trend for 2026. Checkpoint 3 is January of 2027 with Q4 plus FY27 guidance. Above 18% organic guide, we have the bull confirmation. Below 16%, the bare case is present. And then checkpoint 4 is April of 2027, full thesis test.
Now, I don't believe in probability-weighted expected returns. I think assigning a 30% probability to the bull case and a 25% to the bear is just thumb-sucking dressed up as precision. Nobody knows what those probabilities are. But for the sake of illustration, let's run it. Say you give a one in four shot for the bear, 45% for the base, and 30% bull. The expected return from $90 at the midpoint of each range comes out at roughly flat. At the upper end of the band, it's roughly about plus 10%. Add the buyback yield, $5 billion in authorization, and a $93 billion market cap, roughly 1.5% annualized share count reduction, and you're somewhere around 12% over the next 12 to 18 months. Even giving the bare case a 1 in 4 probability, the expected value is positive. That's not because the upside is huge. It's because the downside is already partially priced in at $90 a share. But that's not how I use my frameworks. My preferred approach is simpler. Run all three cases, know the ranges, then track which case the data is confirming quarter by quarter using the signals and checkpoints I just laid out. You don't need to assign probabilities. You need to watch current RPO on July 21st and know whether you're tracking bare, base, or bull. The data tells you, not the model. And at each point, this builds conviction either to exit or to scale the position.
And for Service Now, you don't need heroic execution here. You need the base case not to break. And you know exactly how to test whether it's breaking. So let's zoom out. The cost of intelligence dropped 99% in two and a half years. When a resource becomes a commodity, economic value migrates to whatever is scarce and complimentary. Steel became cheap, value accrued to the architects. Compute became cheap, value accrued to the software companies. Intelligence is becoming cheap, value accrues to the execution layer that applies it safely at scale with context. So you see, raw power goes viral as we see almost every day online. However, constrained products capture durable value. AutoGPT went viral. Perplexity captured the revenue. OpenClaw went viral. Enterprise governance platforms will capture the revenue every time.
And this is where the terminal value question gets its answer. If AI commoditizes the workflow layer, if agents can orchestrate themselves across hundreds of applications without governance, then yes, the terminal value is broken. Service Now is a 15-year compounder, not a 30-year one. But if agents need more governance, more identity, more resolution, more audit trails than humans ever did, then the terminal value hasn't changed. It's gotten stronger. Every agent deployed in a Fortune 500 company needs some way governed to act. Jensen Huang said it like this. "But one of the things that's for sure, Service Now is in such a great position because this is where our employees is the platform of our employees. This is the operating system of our company." That's the CEO of the company building the hardware that powers every frontier AI model telling you the workflow layer goes up, not down.
And in terms of the data we have around Service Now's terminal value being repriced, here's what we know. Current RPO has been stable at 21% for five straight quarters. That's not a company whose forward demand is structurally impaired. The 2011 customer cohort is at 228% of initial ACV after 15 years. That is terminal value compounding in real time. The creator and other share of revenue has shifted from 17% to 21% in 12 months. Meaning the revenue base is diversifying away from legacy channels into agentic products that didn't exist three years ago. And the renewal rate is 97% through every AI panic cycle. If sophisticated enterprise buyers genuinely believe Service Now was being displaced, they'd be signaling intent to leave. They're not. In fact, they're expanding.
Now, you can agree or disagree, but you have a framework for testing it. Four KPIs, four dates, two quarters. That's not hype. That is a plan. Now, do I personally think Service Now is a buy at $90? I think the framework supports a position at this price, not a concentrated bet, a position size to the probability that the base case holds and the bull case has a non-trivial chance. The near-term cash flows are intact. The current RPO hasn't broken. The renewal rate hasn't broken. The AI monetization is scaling faster than any comparable platform. The insider behavior, McDermott buying stock, C-suite halting selling, $5 billion in buyback authorizations, is consistent with management that believes the stock is mispriced. What gives me pause is the terminal value question. I don't know whether agents commoditize the workflow layer or make it more valuable. I believe the governance bull case, but I hold it with the appropriate uncertainty. Because the honest answer is nobody knows. We are all watching the same experiment unfold in real time. What I do know is at $90, you're not paying for certainty. You're paying 16 times forward free cash flow for a company growing 20% with a 97% renewal rate. And you have data checkpoints July 21st, you have October, Jan, and that will tell you if signals one and four hold through Q3. The base case is confirmed and the bare case becomes harder to sustain. If either breaks, it's a different story.
And full disclosure, at the time of this recording and posting, I do not have a position in Service Now. This is not financial advice. Always do your own research. And if you found this kind of analysis useful, subscribe. I want to do more of these if the demand is there. I'll see you in the next.