ServiceNow began 2026 with $8.56 billion more cash than debt. It ended the first nine months of the year with $563 million more debt than cash, a swing of roughly $9.1 billion that the company produced not by losing money but by buying companies with the money it had. Calcalist worked the figures out of ServiceNow’s own filings and published them on Nov 8. The acquisition line for the period was $9.17 billion: about $7.6 billion for Armis, the asset intelligence and security firm; about $1.2 billion for Veza, an identity security company; and a run of smaller transactions making up the rest. By Sept 30, cash had fallen to $6.96 billion while total debt had climbed to $7.52 billion.
The direction matters more than the arithmetic. A software company that turns net cash into net debt inside nine months is telling shareholders it has decided the next phase of growth will be bought rather than built, and it is saying so in a year when its reported profit fell. Third-quarter GAAP net income was $277 million, down 44.8% from a year earlier. Non-GAAP net income in the same quarter was $1.04 billion. The gap between those two numbers is the strategy arriving in the accounts: amortization of acquired intangibles and stock-based compensation are excluded from the non-GAAP figure and included in the GAAP one, and a company that closes several deals in a year accumulates both. The acquisitions did not make ServiceNow less profitable in cash terms. They made it less profitable in the way auditors count.
What it bought is a coherent bet rather than a set of bargains. Armis keeps an inventory of every device attached to a corporate network and watches what those devices do. Veza maps who can reach what, and under which permissions. Together the two answer the questions a chief information officer cannot answer about a network now populated by laptops, factory sensors, contractors and, increasingly, software agents acting on an employee’s behalf. Those agents explain the timing. ServiceNow’s core product is a workflow engine billed largely by the seat, and the seat is the unit that enterprise AI is quietly dissolving. A customer that once licensed software for every employee who filed a ticket now buys access for a smaller group of administrators who supervise agents doing the filing, and owning the identity layer and the device layer lets the company sell into the part of the stack that does not shrink when headcount does.
That is a defensible reason to pay up. Security is the last budget line a technology chief cuts, and consolidation is the theme of the cycle: buyers who once ran forty vendors want eight. A platform that answers the device question and the permission question is easier to defend in a procurement review than two point products that each need their own integration project. ServiceNow also has the cash flow to carry the debt, because subscription revenue is contracted and recurring, which is why lenders price its paper the way they do.
The objection is about price, and nobody outside the negotiating room can settle it. Armis and Veza are private companies, and private companies do not publish the revenue that would let an analyst calculate what ServiceNow paid for a dollar of sales. What is public is the cash that left the balance sheet. What is not public is whether the assets bought will grow faster than the amortization they created. That question has a history in enterprise software. The 2021 vintage of acquirers bought growth at prices that assumed it would continue at the rate of the pandemic, and when it did not, the write-downs arrived two years later and the goodwill line on their balance sheets read like a confession. ServiceNow struck its deals in a different market, but the mechanism that punishes an overpayment has not changed.
Integration is where the risk turns concrete. Armis sells to security teams, which sit in a different budget from the IT service desk that buys ServiceNow. Veza sells to identity architects, who report somewhere else again. Neither buyer is the one the company’s sales force knows how to reach, and cross-selling across budget lines is the part of an acquisition that looks simple in the announcement and takes three years in practice. ServiceNow’s own record cuts both ways. It grew for most of its history by building a platform and absorbing small teams whose products folded into the core, a strategy that produced high-margin subscription revenue and very little goodwill. The 2026 campaign is an order of magnitude larger, and it commits the company to running businesses it did not start, with engineering cultures it did not shape and customers it has never billed.
What investors will watch is free cash flow rather than earnings. Amortization is a non-cash charge that fades as intangible assets are written off, so the GAAP number will improve on its own even if nothing operational changes. Cash flow will not. If the acquired businesses add revenue fast enough to cover interest on $7.52 billion of debt and replace the cash that was spent to obtain them, the balance sheet was a tool. If they do not, the debt becomes the constraint on the next deal, and a company that has just spent its flexibility has to earn growth the slower way.
There is a wider reading that extends past one company. The years in which enterprise software was rewarded for hoarding cash are over, and boards now ask management to put it into AI-adjacent categories before a competitor does. ServiceNow answered that question early and expensively, and it did so while its largest customers were still deciding how many humans they intend to employ. Whether it answered correctly will be settled by the next four quarters of disclosure, and the filings will say so before management does.
