14x and 28x. Even though they are both enterprise software, Salesforce and ServiceNow are valued so differently in the market. Even though the stock price has dropped by 301 TP3T over the past year, ServiceNow is still trading at double the industry average. My conclusion is Hold, with a 12-month baseline target of $118; let me break down why these figures were arrived at.
The analysis reference date is September 14, 2026. The KRW amounts in the text were calculated at 1,333 KRW per dollar (September 13), and the reference stock price is $132.53.
- Let me start with the conclusion about ServiceNow.
- What on earth does ServiceNow sell?
- ServiceNow (NOW) Q2 Earnings: Why Revenue Was Strong but Margins Collapsed
- ServiceNow Valuation: Is a forward P/E of 28x expensive?
- ServiceNow (NOW) Growth Outlook and 12-Month Target Price Scenarios
- A Day Created by GPT-6 Astra
- ServiceNow (NOW) Risks and Bear Triggers
- Stories outside the data
- ServiceNow (NOW) Investment Opinion Conclusion
- Frequently Asked Questions
- source
Let me start with the conclusion about ServiceNow.
First, the business itself is still running very well. Revenue in the second quarter of 2026 increased by 241% compared to a year ago, and the company even raised its annual forecast.
Second, however, the quality of earnings has deteriorated noticeably. The gross profit margin fell from 77.51 TP3T to 70.71 TP3T in one year, and the generally accepted (GAAP) operating profit margin plummeted from 11.11 TP3T to 4.11 TP3T.
Third, the fair multiple I calculated is approximately 24 times the expected earnings for the next year, but the current stock price is 28 times. Therefore, at this price, I consider it a Hold, and specifically, the lowest possible Hold level.
What on earth does ServiceNow sell?
If I were to explain ServiceNow in one word The central post office through which all requests exchanged within the company pass.It is. Tasks such as reporting a broken laptop to the IT team, assigning an account and workspace to a new employee, and receiving customer inquiries—these were originally handled separately by each department using Excel and email. ServiceNow places these tasks on a single platform (Now Platform) and makes them flow automatically.
It was founded in 2004 and is headquartered in Santa Clara, California. It has 29,187 employees, and the CEO is Bill McDermott, who led SAP. Almost all of the revenue is Subscription feeIt comes from [this], meaning it is not a one-time sale that ends, but money that is automatically renewed every year. This was the biggest reason why this company has been receiving a high price over the years.
The market capitalization is $137 billion (approximately 182.6 trillion won). For reference, a 5-to-1 stock split was conducted in December 2025, so you should divide by 5 when comparing with the previous share price to avoid confusion.
ServiceNow (NOW) Q2 Earnings: Why Revenue Was Strong but Margins Collapsed
I personally opened up the report card for the second quarter of 2026 (ending late June), which was released on July 22nd. Let's start with the good parts.
| item | Q2 2026 | 1 year ago | change |
|---|---|---|---|
| Total sales | $3.987 billion | $3.215 billion | +24.0% |
| Subscription revenue | $3.877 billion | — | +24.5% |
| Gross profit margin | 70.7% | 77.5% | -6.81 TP3T points |
| GAAP operating profit | $162 million | $358 million | -54.7% |
| GAAP Net Income | $298 million | $385 million | -22.6% |
| Free cash flow | $473 million | $526 million | -10.1% |
Subscription revenue of $3.877 billion (approximately 5.17 trillion won) exceeded the upper end of the company's guidance by 1.51 points. cRPOIt increased by 211 TP3T to $13.2 billion (approximately 17.60 trillion won); cRPO refers to “a contract that has already been signed and will be recognized as revenue within the next 12 months.” Simply put, since this means a significant portion of next year’s revenue is already booked, I consider this figure more important than quarterly revenue.
Achievements in the AI sector have also been released in actual numbers. The annual contract value (ACV) for AI has exceeded $1 billion (approximately 1.33 trillion won). ACV represents the annual recurring contract value, and achieving this amount solely through AI products is effectively a numerical answer to the question, "Are you making money with AI?" The number of agent-based AI deployments increased ninefold in nine months, and the AI control tower surpassed 500 clients in six months. New contracts exceeding $1 million also increased to 123 deals, reaching nearly 401 TP3T. Earnings announcement original textYou can check these figures directly at.
But I stopped here.
I, too, almost let it slide at first, thinking, "It's a good quarter," but I stopped at the cost of goods sold. The cost of goods sold rose from $724 million to $1.169 billion (approximately 1.56 trillion won). 61.5%I ran. If sales increase by 241 TP3T but costs increase by 621 TP3T, the margin won't be able to hold out.
First is the accounting cost resulting from the acquisition. ServiceNow, an Israeli security company in the second quarter Acquired Armis for $7.75 billion (approx. 10.33 trillion won) in all cashThat is what was done. When you acquire a company, a significant portion of the acquisition price is recorded as intangible assets, which are expensed annually through amortization. In fact, depreciation expenses increased by $235 million, from $172 million to $407 million (approximately 542.5 billion won), and this explains more than half of the increase in the cost of goods sold. It is not an actual cash outflow expense.
Second, it is the actual cost of running AI. The remaining half appears to be the server and inference costs incurred when actually running the AI model. This is what I am paying closer attention to. In the past, once the code was written, the cost barely increased regardless of whether there were 100 or 1,000 customers, but now, as customers use AI more, the cloud bill grows along with it. A gross profit margin of 70.71 TP3T means that for every 100 won sold, 70.7 won remains after deducting direct costs; since 77.5 won remained a year ago, it amounts to a loss of 6.8 won.
It changed from a net cash company to a net debt company.
Our financial structure changed significantly to finance the acquisition of Amis. We raised $4 billion in secured loans in April and repaid those loans by issuing $4 billion in senior unsecured bonds in May. For more details, Q2 10-Q DisclosureIt is listed in.
- Total borrowings: $2.431 billion → $8.453 billion (approximately 11.27 trillion won)
- Cash and short-term investments: $4.664 billion (approx. 6.22 trillion won)
- Net debt: $3.789 billion (approximately 5.05 trillion won)
- Goodwill: $4.541 billion → $9.837 billion (approx. 13.11 trillion won)
- Net value of tangible assets (net assets excluding goodwill and intangible assets): minus $1.102 billion
Dividing net debt by EBITDA yields a ratio of approximately 1.08. This means that the debt can be fully repaid by accumulating just over a year's worth of cash generated from operations, so it is not a dangerous level. However, The nature of the company changed from a net cash company to a net debt company.It is an undeniable fact that quarterly interest expenses increased elevenfold from $6 million to $66 million (approximately 88 billion won). It is structured so that an annual interest payment of around $400 million eats away at earnings per share (net profit divided by the number of shares) by about $0.3.
For reference, since the current ratio is 0.70, you might think, "Isn't there a shortage of funds?" but that is a misunderstanding. Of the $12.153 billion in current liabilities, $8.057 billion is Player earnings, In other words, it is money that the customer has paid in advance as a subscription fee. Since it is not a debt to be repaid but an obligation to be repaid through future services, it is perfectly normal for subscription companies.
ServiceNow Valuation: Is a forward P/E of 28x expensive?
This is the main point of today's post. I created a table and placed it side-by-side with competitors in the same industry.
| company | Preceding P/E | Sales growth rate | Operating profit margin | Stock prices over the past year |
|---|---|---|---|---|
| ServiceNow | 26.5 times | +24.0% | 4.1%(GAAP) | -30.1% |
| Salesforce (CRM) | 15.5 times | +10.8% | 21.4% | +2.1% |
| Workday (WDAY) | 14.0 times | +12.8% | 11.8% | -16.6% |
| Adobe (ADBE) | 9.1 times | +12.9% | 34.8% | -27.3% |
You can read this table like this. The median forward P/E of the three peer companies is 14.0x, but ServiceNow is trading at 26–28x. That is almost double. However, ServiceNow’s growth rate is 241 TP 3 T, which is double that of the others. So... “Since growth is double, the price is also double.”That is essentially the price tag the market has set right now.
The question is whether this logic holds true. Indicators that view valuation relative to growth rate PEGIt is referred to as the forward P/E ratio divided by the growth rate. If you calculate it, Salesforce is 1.43, Workday is 1.10, and Adobe is 0.71, resulting in a median of 1.10. If you multiply ServiceNow's weighted growth rate of 19.81 TP3T by this 1.10, the appropriate multiple is 21.8 timesIt comes out.
I added a bit of a premium to this. ServiceNow has a free cash flow margin of 311 TP3T, placing it among the top tier in the industry. This means that out of every 100 won in revenue, 31 won remains in cash after deducting capital expenditures. Furthermore, it is the number one player in the workflow platform market. With a growth rate of 241 TP3T and a cash margin of 311 TP3T, its "Rule of 40" score is 55 points; this is more than enough to warrant a premium. Therefore. The appropriate multiple I determined is 24 times (range 22 to 27 times).It is.
- Estimated earnings per share (NTM EPS) for the next year: $4.73 (approx. 6,305 KRW)
- Current Multiple: 28.0 times
- The appropriate multiple I saw: 24.0 times
- premium: +16.7%
I also calculated it using DCF.
I'm uneasy relying on just one multiple. DCFI also tried running it. DCF is a method where a company converts its future cash earnings into today's value and adds them all up. Since 1 million won in the future is less than 1 million won today, it is rounded down using the discount rate (WACC) and added.
| home | 값 | reason |
|---|---|---|
| Discount rate (WACC) | 9.0~10.5% | Beta 0.97, reflecting risk-free interest rate and new borrowing rate |
| End-of-life growth rate | 2.5~3.2% | Long-term GDP + Prices |
| forecast period | 10 years | Long visibility of the subscription model |
| Standard free cash flow | $3.8 billion to $4.9 billion | Partial deduction of stock compensation expenses from $4.571 billion over the past year |
I was stubborn about one thing here. For the past year, ServiceNow $2.185 billion (approximately 2.91 trillion won) in stock compensation expensesI wrote about it. It amounts to 14.81 TP3T of revenue; since this is salary paid to employees in stock instead of cash, it does not represent an outflow of cash on paper, so it is added back into the free cash flow calculation. However, issuing new stock dilutes the equity of existing shareholders. In fact, to prevent this, the company bought back $3.77 billion worth of its own stock over the course of a year. Therefore, this money is effectively an outflow. That is why I Deduct half of the stock compensation expense as an expenseI ran both a conservative version and an unsubtracted version.
- Base scenario without deductions: Per week $132.7
- Base scenario with half deducted: Per week $102.1
- Midpoint of two values: approx. $117
If you mix the multiple method (24x × $4.73 = $113.6) and the DCF median ($117) in equal halves About $115–$118It appears. The valuation determination is The boundary between fair value and overvaluationIt is. It isn't quite at the level of being called clearly overvalued, but it is true that there is a premium on the current price. Similar concerns Palantir (PLTR) Valuation Recalculation PostI tried it there as well, but the premium there was much more extreme, so the conclusion was different.
ServiceNow (NOW) Growth Outlook and 12-Month Target Price Scenarios
The company presented annual subscription revenue guidance for 2026 of $15.76 billion to $15.78 billion (approximately 21.02 trillion won), representing an increase of 22.51 TP3T compared to a year ago. The subscription revenue guidance for the third quarter is $3.975 billion to $3.98 billion. 20.5% growthIt is.
There is a point worth noting here. Q2 actual growth rate 24.51 TP3T → Q3 guidance 20.51 TP3T. Growth is dropping one level every quarter. The cRPO growth guidance was also lowered from 211 TP3T to 201 TP3T. While this is a natural phenomenon as a company grows in size, a multiple of 28x is based on the premise that growth is sustained. The weighted growth rate I used is calculated by multiplying historical earnings of 24.01 TP3T by 0.4, the average market forecast (consensus) of 19.01 TP3T by 0.4, and the enterprise software industry growth rate of 13.01 TP3T by 0.2, and adding them together. 19.8%no see.
| scenario | Target price | Compared to the current price | Apply multiple | Probability |
|---|---|---|---|---|
| stress | $155(Approx. 206,615 won) | +17.0% | 32.7 times | 27% |
| basic | $118(Approx. 157,294 won) | -11.0% | 24.9 times | 47% |
| Weakness | $88(Approx. 117,304 won) | -33.6% | 18.6 times | 26% |
Basic ($118)It is a compromise between an appropriate multiple of 24x and the midpoint of the DCF. The scenario envisions the multiple remaining at 24 to 25x as subscription revenue grows to around 201 TP 3T as per guidance, while the burden of Amis amortization persists.
Bullish ($155)This is a case where the fear that “AI agents will replace ServiceNow” is subsiding as the annual AI contract volume heads from $1 billion toward $2 billion and the cRPO growth rate rises back above 221 TP3T. Since the stock reached $149.60 just in August, this is a path it has already traveled.
Weak (88 dollars)This is the scenario where general-purpose AI agents actually start eating into seat-based contracts, and cRPO growth drops to the mid-10% range. It will be around this level if multiples converge to industry standards (14–18x). Since it actually hit $81.24 in April of this year, this is not a figment of imagination.
Let me also explain why I divided the probabilities this way. I set the base at 471 TP 3 T, but allocated a thicker weight to both the bullish and bearish sides at 271 TP 3 T and 261 TP 3 T. Currently, for this stock, it is not about earnings, but “The narrative of whether ”AI agents will replace SaaS”Since [the stock price] is shaking up the price, I predicted that the outcome was more likely to diverge to the extremes rather than converge in the middle. In fact, Wall Street target prices are split by a factor of 3.4, ranging from CLSA’s $72 to Bernstein’s $248. For stocks where opinions are divided like this, you should not concentrate your odds on a single point.
A Day Created by GPT-6 Astra
On September 3, OpenAI GPT-6 AstraIt released [it]. It is an agent-type model that directly manipulates the browser, writes and executes code, and handles software on behalf of others to complete multi-stage tasks. On September 8th alone, ServiceNow dropped approximately 51 TP3T, Salesforce approximately 41 TP3T, and Intuit also approximately 41 TP3T. An article summarizing the market reaction at the timeIt is also worth referring to.
The scenario the market is concerned about is this: if a single general-purpose agent can complete tasks across multiple apps, companies can purchase fewer expensive seats (licenses). However, if purchasing teams start comparing “how much it costs to hire an agent versus how much a ServiceNow license costs” side-by-side, their bargaining power will be shaken.
I don't think this fear is exaggerated or unfounded. The agent Screen and controlsEven though it can be replaced, the real reason companies pay ServiceNow is Data, Access, Audit Trail, and ComplianceThere is still a long way to go to replace even that. For banks to simply tell AI to “handle it automatically,” records of who approved what and when must remain, but ServiceNow still holds those records. In fact, CFO Gina Mastantuono stated at the Citi conference on September 11 that customer conversations are focused on “how to run AI at scale while maintaining security, governance, and cost controls.” I believe this perfectly describes where ServiceNow currently stands.
Wall Street is not intimidated yet, either. On September 8, BTIG raised its target price from $150 to $170, and on September 11, Needham raised it from $115 to $155. For reference, a case where the price actually went up within the same AI narrative is... Oracle (ORCL) Q1 Earnings ArticleI've organized it here.
ServiceNow (NOW) Risks and Bear Triggers
Risk 1 — Case where the agent eats away at the number of seats. If a general-purpose agent handles the workflow, license demand decreases, and prices are lowered during renewal negotiations. The metrics to check are cRPO growth rate과 Net Retention Rate (NRR)You can view this every quarter in the quarterly earnings reports and on the IR page. You should consider it a signal if cRPO growth drops below 201 TP3 T for two consecutive quarters.
Risk 2 — Case where AI costs continue to eat into the margin. As AI usage increases, inference and cloud costs are added, so even if revenue grows, the profit margin does not rise. The metrics you need to check are Gross profit marginIt is. You just need to check every quarter whether it drops further from 70.7% or recovers.
Risk 3 — If the acquisition of Amis falls short of expectations. If the $9.837 billion in goodwill is written off, it will result in a significant accounting loss, and the impact appears even greater since the net value of tangible assets is already negative. The indicators to check are those at the time of the earnings announcement. Security Sector Revenue Growth RateThis is the review of goodwill impairment in the annual report.
Risk 4 — Valuation itself. While peers in the same industry are trading at 9 to 15 times, it is trading at 28 times alone, so even the slightest fluctuation in earnings leads to a large correction. I am not sure how high valuation turns into a burden. Cadence (CDNS) Analysis ArticleIt was handled similarly in that as well.
So, what makes this stock a Sell?. I look at three things: ① the moment the cRPO growth rate falls below 181 TP3T, ② the moment the gross profit margin falls below 681 TP3T once more, and ③ the moment the 2027 subscription revenue guidance is presented as less than 181 TP3T. If two of these three coincide, I will lower the target price below $100.
Stories outside the data
While organizing the numbers for this quarter, what I lingered on the longest was the divergence in the directions of revenue and cash. Revenue increased by 241 TP3T, but free cash flow decreased by 101 TP3T. Of course, this could be a one-quarter phenomenon involving cash outflows related to acquisitions. However, when these two lines start diverging in a software company, I make a note of it. If they remain diverged in the next quarter, it is no longer a one-off occurrence. This is the first line I look for in the Q3 earnings at the end of October (scheduled for October 28 according to Yahoo Finance). A similar signal Synopsis (SNPS) Q3 Earnings ArticleI remember pointing that out as well.
One more thing. Seeing the stocks drop by 51 TP3 T in a single day on September 8 reminded me of what happened during the early days of Generative AI in 2023. Back then, people were asking, "Why buy SaaS when ChatGPT does it all?" and as a result, companies ended up buying both. However, there is one difference this time. At that time, SaaS stocks were trading at 40 to 50 times their value, whereas now Adobe is at 9 times and Workday at 14 times. The market has already largely surrendered.That is what it means, but the fact that ServiceNow is the only one holding its 28x return in the midst of this is the reason I cannot readily buy this stock. If the defense line collapses, the only place it can go is down. Honestly, I am only half-convinced about this point myself.
ServiceNow (NOW) Investment Opinion Conclusion
Hold (Medium Confidence) · 12-month baseline target price $118 (approx. 157,294 KRW)
Let me summarize the three reasons why I finished with "Hold." First, The valuation is in the premium range. Compared to the fair multiple of 24x that I observed, the current rate is 28x, making it approximately 171 TP3T more expensive. Calculating the expected return using probability weighting yields about -9.31 TP3T.
Second, However, the business did not collapse enough to warrant a sell. Revenue growth of 241 TP3T, an upward revision of annual guidance, 211 TP3T growth in cRPO, and annual AI contract volume surpassing $1 billion are evidence that the business is healthy. The decline in margins is mostly due to accounting for the Amis acquisition and intentional AI investments, not because the core business has collapsed. While it hits the Sell boundary based on mechanical expected return bands, I finished with a Hold rating, placing greater importance on the fact that the fundamentals have not broken. However, I must make it clear that this is at the very bottom of the Hold range.
Third, There is a specific price range that I am interested in. If it falls below $105 (approx. 139,965 KRW)—that is, below 22 times the projected earnings for the next year—I believe that would be a sufficiently attractive level relative to the growth rate. At the current level of $132.53 (approx. 176,662 KRW), the downward gap appears wider than the upward gap.
The Wall Street consensus, based on 46 analysts, is a Strong Buy (1.43), with an average target price of $143.84 (approximately 191,739 KRW). This differs from my conclusion. The only difference is how many times the appropriate multiple is considered. Wall Street still places weight on ServiceNow's average multiple over the past five years, while I have placed more weight on the current reality where peers have fallen to 9 to 15 times. Ultimately, which side is right will be determined by how much AI agents cut the price of enterprise software.
Frequently Asked Questions
What is ServiceNow's target price?
My calculated 12-month baseline target price is $118 (approximately 157,294 KRW). I set the target at $155 for a bullish scenario and $88 for a bearish scenario. For reference, the average target price of 46 Wall Street analysts is $143.84.
The stock price has dropped by 301 TP3T in the past year, so isn't it cheap right now?
It is true that the price has dropped significantly, but it is difficult to say that it has become cheap. Market earnings forecasts have been adjusted in tandem with the decline in the stock price. The current forward P/E of 28x is still about double that of Salesforce (15.5x), Workday (14.0x), and Adobe (9.1x).
Will ServiceNow go bankrupt if an AI agent like GPT-6 comes out?
It is too early to conclude that it will fail. While agents can handle screen manipulation, it takes time for them to take over the data, permissions, and audit records that are the real reasons companies pay. However, since there is a real possibility that the per-seat pricing model will be under pressure, I recommend checking the cRPO growth rate every quarter.
source
- ServiceNow Reports Second Quarter 2026 Financial Results (SEC 8-K)
- ServiceNow Form 10-Q FY2026 Q2 (Liability · Amis Acquisition Accounting)
- ServiceNow Investor Relations — Q2 2026 Earnings
- ServiceNow completes Armis acquisition (Newsroom)
- GPT-6 Astra Wiped 4%-5% Off Salesforce and ServiceNow in a Day (Yahoo Finance)
This article was drafted using AI research tools based on publicly available disclosures and market data, and I verified the numbers and logic step-by-step before publication. Stock price, exchange rate, and earnings estimates are based on the time of writing and are subject to change. The purpose is for informational purposes only and is not a recommendation to buy or sell any specific stock. Please remember that you are solely responsible for your investment decisions and their consequences.