Everything About AI
NextFin News - Microsoft reported fiscal first-quarter earnings per share of $4.89 on revenue of $92.28 billion after the closing bell on Wednesday, beating the $4.73-per-share and $90.63-billion consensus, while Azure sales climbed 48% to $33.17 billion and capital spending came in well below what investors feared. Hours earlier, ServiceNow delivered another beat-and-raise quarter, lifting subscription revenue to $3.988 billion and raising its 2026 AI contract-value target to $1.7 billion - yet its shares fell as the market weighed a $39.4-billion quarterly capex bill at Microsoft against a software peer that funds AI growth without building data centers.
The Tale of Two AI Models: Build It vs. Rent It
Wednesday's twin earnings reports laid bare the central tension of the artificial-intelligence buildout: the company that owns the infrastructure and the company that rents it both grew at roughly 20% or better, but the market rewarded them very differently. Microsoft shares rose more than 4% in after-hours trading, reaching $554.00, after the company showed that Azure growth accelerated to 48% while capital expenditures, including leases, topped $39.37 billion - below the $45.93 billion analysts had been bracing for. ServiceNow closed down 3% at $150.13 despite beating across its topline and profitability metrics and raising guidance, a reminder that in the current market a clean beat is not enough when the entire software sector is being repriced against the hyperscalers' spending plans.
The divergence is not just about one quarter. It is about two competing answers to the same question: in the agentic AI era, does the money accrue to the owner of the compute, or to the layer that controls the workflow? Microsoft is betting its balance sheet that the answer is the former. ServiceNow is betting that enterprises will pay a premium to govern the agents running on that compute - and that it can do so without the capital intensity that has made investors nervous about the rest of the software stack.
The scale gap between the two models is the first thing the numbers reveal. Azure's quarterly revenue alone - $33.17 billion - is more than eight times ServiceNow's entire quarterly subscription base of $3.988 billion. Yet ServiceNow's roughly 20% growth is coming off a much smaller denominator and without the capital intensity. Microsoft spent more than $39 billion in one quarter to keep the AI engine fed. ServiceNow's president and chief financial officer, Gina Mastantuono, framed the contrast directly on the company's results call: the company is "capturing this significant opportunity with a capital-efficient model, reducing capex and raising our full-year free cash flow margin target by 100 basis points to 36 percent."
What the Numbers Actually Say
Microsoft's headline beat masked a deeper shift in how the company now reports itself. The quarter was the first under a reorganized segment structure that collapses Azure, Microsoft 365 cloud services, productivity and server licensing, industry solutions, and frontier-model services into a single "Agents and Infra" division. That segment produced $76.8 billion, ahead of the $75.6 billion forecast. Azure alone contributed $33.17 billion, a 48% jump from $22.38 billion a year earlier - the fastest pace in more than a year and a full three points above the roughly 45% growth management had guided for in July. The Devices and Consumer group, now the smaller sibling, added $15.5 billion against a $15 billion estimate, even as Xbox hardware sales fell about 14% and Windows OEM and devices revenue dropped 22% on difficult comparisons to the Windows 10 end-of-support cycle.
ServiceNow's print was cleaner in a different way. Subscription revenue of $3.988 billion grew 21% year over year, or 20.5% in constant currency, and total revenue reached $4.106 billion, up 20.5%. Current remaining performance obligations - contracted revenue due within the next 12 months - rose 20.5% to $13.65 billion, while total RPO climbed 22% to $29.7 billion. The company closed 19 transactions larger than $5 million in net new annual contract value in the quarter, a 135% year-over-year increase, and ended with 680 customers above that threshold, up 23%. Perhaps the most telling figure was the one management chose to highlight twice: AI net new annual contract value is running ahead of plan, prompting the company to raise its 2026 AI ACV target to $1.7 billion, up 70% from the $1 billion target at the start of the year.
"This was more than another beat-and-raise quarter for ServiceNow, this was the inflection point for ServiceNow as the Super Intelligence Control Tower," ServiceNow Chairman and CEO Bill McDermott said in a prepared statement. "That's why AI adoption is accelerating, surging our hybrid business model to $30 billion in RPO, growing 23 percent. What other company at the forefront of the agentic enterprise has no capital or compute constraints? This is a one-of-one, rule of 571 growth company firing on all cylinders."
Put the two prints side by side and the strategic asymmetry becomes visible. Microsoft is converting AI demand into cloud revenue at a rate that now exceeds its own guided acceleration, while ServiceNow is converting the same demand into workflow contracts at a pace that requires no data centers of its own. One business monetizes the picks and shovels; the other monetizes the foreman who tells the miners where to dig. Both are profitable. Only one of them writes checks for billions of dollars of infrastructure every quarter.
The Capex Question Nobody Can Dodge
The single biggest overhang on Microsoft heading into the print was not whether Azure would grow - most investors expected it to - but whether the bill for that growth would keep rising faster than the revenue. Management had guided for capital spending above $50 billion this quarter, a figure that had spooked the stock in the weeks leading up to earnings. The actual number, $39.37 billion including leases, came in roughly 14% below even that lowered bar. That is the kind of surprise that moves a multitrillion-dollar market cap: investors are no longer asking whether AI demand is real. They are asking whether each incremental dollar of capex produces an incremental dollar of Azure revenue, and when the gap between the two stops widening.
The mechanism here is simple but unforgiving. Hyperscale cloud is a scale business with long lead times: data centers must be built before the demand arrives, which means capex leads revenue by quarters, not days. If Azure growth stays near 48% while capex growth moderates, the ratio improves and the stock rerates higher. If capex reaccelerates while Azure growth decelerates back toward the low 40s, the multiple compresses again regardless of how large the absolute revenue number is. Wednesday's print bought Microsoft time, not a verdict.
Context matters. In the prior quarter Microsoft had already pushed Azure growth to 43% and crossed $100 billion in annual Azure revenue for the first time, while quarterly capex hit $41 billion and free cash flow fell 23%. The market's fear heading into Wednesday was that the denominator problem was structural - that Microsoft would have to keep spending at an accelerating rate just to hold a 40-something-percent growth rate. The $39.37 billion figure, against a $45.93 billion fear level, suggests the opposite: that fleet efficiency gains and faster capacity delivery are starting to bend the capex curve before revenue growth bends. That is the second-order insight buried in the headline beat - the market was pricing a linear extrapolation of capex, and the company delivered a curve that inflected sooner.
ServiceNow faces the mirror-image problem. It does not own the data centers, so it does not carry the capex risk - but it rents the compute from the very hyperscalers whose pricing power is the flip side of Microsoft's bet. If Azure, AWS, and Google Cloud raise the price of inference and training to recoup their infrastructure outlays, that cost eventually lands somewhere in the stack. ServiceNow's defense is that its pricing power comes from workflow governance, not raw tokens: AI-native SKUs carry price premiums in the range of 20% to 30%, and roughly half of net new business is already non-seat-based, meaning revenue is tied to outcomes and transactions rather than per-user licenses that commoditize quickly. That is the moat management is selling, and the numbers so far support it. The risk is that the moat is wide but shallow if the underlying token costs rise faster than the premium customers will pay for governance.
The Counter-Thesis: What If the Market Has This Backward?
The strongest argument against the "ServiceNow is the capital-light winner" reading is that the market is not actually punishing ServiceNow for its model - it is punishing the stock for its multiple. At roughly 27 times forward earnings on 24% revenue growth, ServiceNow is priced for perfection in a quarter where the entire software complex is being repriced against hyperscaler capex. A clean beat with a guide that merely meets expectations leaves no room for error, and the 3% decline reflects multiple compression more than fundamental disappointment. By that logic, Microsoft - trading at a lower multiple on a business that still controls the infrastructure layer - is the better risk-reward, and ServiceNow's fall is a valuation event, not a thesis event.
The strongest argument against Microsoft, in turn, is that the capex relief may be temporary. The company has committed to an AI infrastructure buildout that rivals the gross domestic product of a mid-sized nation, and a single quarter of lower-than-feared spending does not change the multi-year trajectory. If Azure growth decelerates from 48% toward the mid-40s while the full-year capex commitment holds, the ratio problem returns. The falsifying signal for the bullish read on Microsoft is straightforward: if Azure growth prints below 44% in fiscal second quarter while capex reaccelerates above $45 billion, the "capex under control" narrative breaks. For ServiceNow, the falsifying signal is equally concrete: if cRPO growth - now at 20.5% - decelerates below 18% for two consecutive quarters, the claim that AI is driving durable pipeline expansion, rather than one-time large deals, comes apart.
There is also a third possibility that neither bull case fully absorbs: the two models are not competitors at all, but symbionts. Microsoft needs the workflow layer to make its infrastructure indispensable to enterprise buyers; ServiceNow needs the infrastructure layer to remain reliable and competitively priced so its governance premium is not eaten by token costs. In that world, the earnings divergence on Wednesday is not a zero-sum verdict but a rotation within the same trade - investors moving from the layer with the clearer capital answer to the layer with the clearer demand answer, and back again as each quarterly ratio prints.
What to Watch Next
In the short term, the reaction is a sentiment and positioning story: software stocks have been under pressure as investors rotated out of names that do not own AI infrastructure, and Microsoft's beat gives the group a tactical reprieve. The medium-term question is fundamental: does Azure's 48% growth hold through the first half of fiscal 2027 as management has guided, and does ServiceNow's AI ACV trajectory convert into subscription revenue at the pace implied by a $1.7 billion annual target? ServiceNow's own guidance gives the market a concrete yardstick - fourth-quarter subscription revenue of $4.245 billion to $4.255 billion, up 22.5% to 23%, with a non-GAAP operating margin of 32%, and full-year subscription revenue of $15.78 billion to $15.79 billion with a 36% free cash flow margin. Microsoft's yardstick is simpler and harsher: keep Azure above 45% while keeping capex from reaccelerating.
The base case is that both companies continue to grow into their valuations, with Microsoft capturing the infrastructure spend and ServiceNow capturing the enterprise workflow premium - a world where the AI buildout is large enough to enrich both layers. The upside case for Microsoft is that capex intensity peaks sooner than feared and Azure growth stays above 45% for four consecutive quarters, compressing the ratio and rerating the stock. The downside case is that capex reaccelerates while Azure growth decelerates, reviving the "spending ahead of returns" critique that pressured the stock earlier in the year. For ServiceNow, the upside case is that AI ACV keeps compounding and the non-seat-based mix shift lifts margins toward the 36% free-cash-flow target; the downside case is that hyperscaler pricing power squeezes the capital-light model from below while the premium multiple leaves no cushion for a guide that merely meets estimates.
"We are delivering greater capability, choice, control, and cost efficiency for customers across our entire stack," Microsoft Chairman and CEO Satya Nadella said in a prepared statement on the results. "By helping organizations build and own their own intelligence, and earning trust through agent safety and security, we are translating our innovation into growing adoption and real-world impact across every industry."
Wednesday's earnings did not resolve the central question of the AI cycle. They sharpened it: Microsoft proved it can grow Azure at 48% without spending as much as feared, and ServiceNow proved it can grow AI revenue without building the infrastructure at all. The market's verdict on which model wins the next leg will come down to one ratio - Azure revenue per dollar of capex - and one conversion - ServiceNow's AI contract value into recurring subscription revenue. Until those two numbers speak, the trade is not between right and wrong. It is between two different bets on where the money in artificial intelligence actually ends up.
Explore more exclusive insights at nextfin.ai.
