NextFin News - Salesforce and ServiceNow are betting tens of billions of dollars that Wall Street is wrong about artificial intelligence. The two enterprise-software giants, whose shares were battered through much of 2026 by fears that autonomous AI agents will make their subscription businesses obsolete, have answered the "SaaSpocalypse" narrative with the two most forceful signals management can send: large-scale share buybacks and unapologetic public defiance. Salesforce, backed by a $50 billion repurchase authorization and a $25 billion debt-funded accelerated buyback executed in March, is retiring its own stock at a pace that would have been unthinkable in its acquisition-led past. ServiceNow, whose board added $5 billion of repurchase capacity in January and whose CEO committed to a $3 million stock purchase, is pairing capital returns with chief executive Bill McDermott's insistence that the company is "only just getting started." The question is whether the buybacks are a genuine statement of undervaluation or an expensive distraction from a structural threat that no amount of financial engineering can retire.
The Situation: A Sector Under Siege Fights Back With Cash
The setup is stark. Enterprise software stocks entered 2026 already nursing a valuation reset, and the arrival of agentic AI sharpened the anxiety into a sector-wide thesis: if autonomous agents can perform the work that Salesforce and ServiceNow seats were sold to automate, the seat-based subscription model that powered a decade of multiple expansion is structurally impaired. The term "SaaSpocalypse" moved from analyst note to market shorthand. Salesforce spent much of the year down roughly 30% from year-start levels; ServiceNow fell about 38% year-to-date by mid-May. Both names spent months as the consensus short setup in software.
The retaliation has been financial and rhetorical. Salesforce's capital-return program has shifted gears entirely. After repurchasing about $5.7 billion of stock in fiscal 2024 and roughly $10 billion in fiscal 2025, the company accelerated to $12.6 billion in fiscal 2026, then in February announced a $50 billion aggregate authorization. In March it issued $25 billion of senior notes at coupons between 4.5% and 6.7% and funneled the proceeds straight into an accelerated share repurchase, taking delivery of approximately 103 million shares on March 16. The balance-sheet consequence was immediate and verifiable in the company's first-quarter filing: noncurrent debt jumped from $10.4 billion to $39.3 billion in a single quarter, even as diluted share count fell 10% year over year. The repurchase alone lifted first-quarter non-GAAP earnings per share by 23 cents and GAAP earnings per share by 14 cents.
ServiceNow's playbook is smaller in scale but similar in message. On January 28, 2026, alongside fourth-quarter results that topped estimates, the company said its board had authorized an additional $5 billion for its repurchase program, supplementing about $1.4 billion of capacity left under prior authorizations. CEO Bill McDermott also entered into an agreement to purchase $3 million of ServiceNow stock, a rare insider commitment that the company highlighted as management conviction in the face of sector-wide skepticism.
The rhetoric has matched the capital. Marc Benioff has publicly dismissed Wall Street's AI-disruption fears as "dead wrong," framing the panic as a repeat of earlier technology transitions that incumbents ultimately absorbed. McDermott, for his part, has leaned into growth metrics: ServiceNow's AI products crossed $1 billion in annual contract value for the first time, agentic deployments grew ninefold in nine months, and the company closed 123 deals worth more than $1 million in net new annual contract value in the second quarter alone, up nearly 40% from a year earlier.
The market has begun to listen. After a brutal first half, software names staged a coordinated rebound in mid-August, with Salesforce, ServiceNow and Workday all climbing between 15% and 25% over a single week on no company-specific news — the classic signature of a sector rotation rather than an earnings reaction. As of August 19, Salesforce changed hands around $206, off its 52-week high of $269.11 but well above the $146.32 low. ServiceNow drew a $150 price target from Bank of America analysts, implying roughly 18% upside from recent levels. "We continue to see some very attractive investments in software for those investors who can afford to be somewhat patient," Gregg Moskowitz, a senior enterprise software analyst at Mizuho, said during the May rebound.
The Analysis: Buybacks Can Defend a Multiple, But Not a Business Model
What the Buybacks Actually Signal
A buyback is a claim of undervaluation backed by cash. Salesforce's is a claim backed by debt. That distinction matters. The company's pivot from organic, cash-flow-funded repurchases to a bond-financed accelerated program is less a calm assertion of value than a defensive maneuver: management is leveraging the balance sheet to shrink the equity base faster than the business is growing, compressing the share count to support earnings per share even as top-line growth decelerates into the low teens.
The mechanics are unforgiving. Salesforce raised $25 billion at 4.5% to 6.7% interest across maturities from 2028 to 2066. That is not free capital; it is a fixed claim on future cash flow that must be serviced regardless of whether Agentforce delivers its promised growth. The bet is simple: if the stock remains depressed, retiring shares at these levels is accretive and the market will re-rate the multiple. If AI disruption is real and revenue growth stalls, the company will have added roughly $29 billion of net new long-term debt on top of a slowing topline. The buyback amplifies outcomes in both directions.
ServiceNow's approach is more conservative. Its $5 billion authorization supplemented existing capacity and did not require a Salesforce-scale debt raise tied to the repurchase itself. That asymmetry is itself a signal: ServiceNow can afford to be patient; Salesforce appears to feel it cannot. The contrast is sharpened by the fact that ServiceNow has been the more aggressive acquirer — it agreed to buy cybersecurity firm Armis for $7.75 billion, its largest-ever deal, and funded that and other security purchases with roughly $7.9 billion of notes and term-loan proceeds in 2026. Salesforce, by comparison, used debt explicitly to buy its own equity.
"This was an outstanding quarter for Salesforce — record revenue, record deals, and cash flow," Marc Benioff said on the May 27 earnings call. "Agentic AI is the biggest growth opportunity for our customers, and for Salesforce."
The numbers behind the bravado are real but mixed. Salesforce's first quarter of fiscal 2027, reported May 27, delivered $11.13 billion in revenue and non-GAAP earnings per share of $3.88, up 50% year over year. Agentforce annual recurring revenue reached $1.2 billion, up 205% and above $1 billion for the first time; combined Agentforce and Data 360 ARR approached $3.4 billion, up more than 200%. Yet the forward guide told a more cautious story: second-quarter revenue guidance of $11.27 billion to $11.35 billion came in just below the $11.36 billion consensus, and full-year guidance of $45.9 billion to $46.2 billion implied roughly 11% growth — a reacceleration promised for the second half, not yet visible in the first.
The Cyclical Overshoot Versus the Structural Shift
Here is the judgment the market has not cleanly made: the 2026 selloff conflated two different phenomena, and the buyback rally is pricing the correction of only one.
The cyclical leg is clear and mean-reverting. Valuations were reset aggressively: the S&P Information Technology sector fell to roughly 20 times forward earnings in mid-2026, near the first percentile of its one-year range and below its 10-year average of 23 times, according to Citadel Securities' August market-structure review. Sentiment swung so far negative that software recorded the largest single-day retail liquidation the platform had ever observed, with average daily net selling more than twice the previous weekly record. Drawdowns of that size, driven by positioning and narrative rather than a collapse in fundamentals, tend to snap back when earnings prove resilient — which is exactly what ServiceNow's second quarter did.
"ServiceNow's exceptional Q2 results solidify our position as the fastest-growing major enterprise software and cybersecurity company," McDermott said in the July 22 release. "The company's sterling fundamentals have us operating to the Rule of 56, well on our way to the Rule of 60."
ServiceNow's second-quarter numbers, reported July 22, were the clearest evidence that the apocalypse was overstated in the near term. Subscription revenue grew 24.5% year over year to $3.877 billion. Current remaining performance obligations — contracted revenue due in the next 12 months — rose 21% to $13.20 billion. Total RPO reached $29 billion. Non-GAAP operating margin held at 29.5%, three points above the company's own guidance. AI net new annual contract value surpassed $1 billion. None of that reads like a business in terminal decline.
But the structural leg is not mean-reverting, and no buyback touches it. AI agents genuinely change how enterprises consume software, and the pressure is already visible in the companies that sit closest to the pressure point. Consumption-based platforms such as Datadog, Snowflake and MongoDB sold off sharply in early August after Datadog and HubSpot earnings raised questions about AI monetization and whether infrastructure spending is crowding out traditional software budgets. Salesforce's own model is migrating away from pure seat-based pricing toward consumption metrics tied to AI "Agentic Work Units" — 3.8 billion delivered to date, up 111% quarter over quarter — precisely because the old per-seat logic does not capture AI usage. That is adaptation, and it is necessary, but it also concedes that the pricing model investors learned to underwrite is changing.
So the correct framing is not "SaaSpocalypse or nothing." It is both: a structural shift in software economics playing out over years, overlaid with a cyclical sentiment overshoot that has now partially corrected. The buybacks address the cyclical half — they put a floor under the multiple and force short-covering. They do nothing for the structural half, where the test is whether AI becomes a revenue accelerant large enough to offset the pricing-model transition.
The Second-Order Problem Everyone Is Ignoring
The consensus read is that buybacks plus strong prints equal "the bears were wrong." The second-order question is different: what happens to the cost of capital for a company that funds shareholder returns with debt while its growth profile resets lower?
Consider the chain. A debt-funded buyback reduces share count and lifts reported EPS — mechanically bullish. But it also raises financial risk at the exact moment the business is navigating a model transition. If revenue reacceleration arrives as promised, the leverage is absorbed and the trade is vindicated. If it does not, the market will not simply leave the multiple where it is; it will apply a conglomerate-style discount for the added balance-sheet risk on top of the growth discount it already applied for AI. In other words, the buyback can work as a tactical multiple-support tool and still leave the stock cheaper on a risk-adjusted basis if the fundamental pivot disappoints. The two outcomes are not mutually exclusive, and the market is currently pricing only the first.
There is a second-order cross-asset implication as well. The software rebound has been funded partly by rotation out of AI hardware, on the logic that software is the counter-cyclical play in the AI cycle — seat-based and subscription models do not carry the AI capital-expenditure bill. That trade is coherent only if software demand proves genuinely insulated from hardware spending. The Datadog-HubSpot episode suggests the insulation is porous: when AI infrastructure budgets tighten, they can compress adjacent software spend rather than redirect toward it. A rotation trade built on assumed insulation is vulnerable to the very data it claims to be insulated from.
The Strongest Counter-Thesis
The bull case deserves its full weight, because it is not a strawman. Bank of America analysts, who rate ServiceNow a Buy with a $150 target, forecast 18% to 22% revenue growth from 2026 through 2028, driven by platform expansion and the monetization of agentic AI, with free-cash-flow margins of 35% to 37%. Morningstar argued in August that Salesforce, Workday and ServiceNow all look attractive relative to still-rising 2026 earnings estimates, framing AI as a tailwind for enterprise software rather than a death sentence. The core of the argument is empirical, not aspirational: AI ARR is already above $1 billion at both companies, agentic deployments are compounding, and the incumbents are winning the enterprise AI governance conversation rather than losing it.
The counter-thesis also has a plausible mechanism for being right. If AI expands the incumbents' platforms — more workflows, more data, more governance, more seats for the humans who supervise the agents — then the "SaaSpocalypse" was never a demand story at all. It was a pricing-transition story, and the incumbents are large enough, sticky enough and cash-generative enough to manage the transition while buying back stock at depressed prices. Under that scenario, today's rally is not a bear-market bounce but the first leg of a re-rating back toward historical multiples, and Salesforce's debt-funded buyback looks like inspired timing rather than desperation.
The signal that would falsify the skeptical view laid out here is specific and observable: ServiceNow's current remaining performance obligations growth and Salesforce's promised second-half organic reacceleration. If ServiceNow's cRPO growth holds above 20% — it guided third-quarter cRPO growth to roughly 20% on a constant-currency basis — and Salesforce delivers the second-half reacceleration it guided toward in fiscal 2027, while AI ARR continues to compound from the $1 billion-plus base at each company, then the structural-bear argument fails. It would prove that AI is expanding the pie faster than it is compressing the pricing model. Conversely, if ServiceNow's cRPO growth decelerates below 15% or Salesforce's promised second-half reacceleration does not materialize, the buyback will be revealed as a bridge to nowhere, and the debt load will compound the multiple compression rather than cure it.
Outlook: A Relief Rally, Not a Regime Change
The base case is a split verdict. Near term, the combination of buyback support, short-covering and resilient second-quarter prints favors a continued relief rally; the sector has room to run simply because positioning was so one-sided and valuations were reset to the first percentile. Medium term, the path diverges by company. ServiceNow, with 24% subscription growth, a self-funded buyback and a backlog that still compounds above 20%, is better positioned to grow through the transition than Salesforce, which has added roughly $29 billion of long-term debt to fund a repurchase while guiding to low-teens growth.
The upside case is the bull thesis realized: AI becomes a genuine platform expander, ARR compounds into the high single digits as a share of total revenue, and both companies re-rate toward 25 to 30 times forward earnings as the "SaaSpocalypse" is relegated to market folklore. The downside case is equally concrete: consumption-based pricing pressure spreads from the monitoring and SMB-software layer into the enterprise core, second-half reacceleration fails to arrive, and the market applies a dual discount — for slower growth and for the leverage taken on to fight it.
What to watch, in order: ServiceNow's third-quarter cRPO growth (guided to roughly 20% constant-currency); Salesforce's fiscal second-quarter results and any update on the promised second-half organic reacceleration; and the trajectory of AI ARR at both companies relative to the $1 billion threshold each has now crossed. Those three data points will separate the companies adapting to a structural shift from the companies merely financing a sentiment bounce.
The buyback is a statement about the stock. The business model is a statement about the future. Salesforce and ServiceNow have made the first loudly; the second is still being written, and no amount of bluster can pre-write it.
Explore more exclusive insights at nextfin.ai.
