For most of 2026, the dominant AI story has been about who is spending. The hyperscalers — Microsoft, Amazon, Google, Meta — are on pace to deploy roughly $690 billion in capital expenditure this year. Nvidia just reported $81.6 billion in a single quarter. OpenAI filed its IPO paperwork. Every headline has pointed in the same direction: more infrastructure, bigger models, faster spending.
What has received considerably less attention is what that spending does to the companies sitting directly underneath it — the enterprise software platforms that sold per-seat subscriptions to the same enterprise customers who are now being told by Google that AI will replace their existing workflows at a fraction of the cost.
That story moved significantly on Wednesday evening, when Salesforce reported Q1 fiscal 2027 results that beat on nearly every headline metric and still sent the stock lower after hours.
The Numbers Were Good. The Market Wasn’t Sure What to Do With Them.
Salesforce posted Q1 FY2027 revenue of $11.1 billion, up 13% year-over-year, beating Wall Street’s consensus of $11.06 billion. Non-GAAP EPS came in at $3.88, up 50% year-over-year, well clear of the $3.12 estimate. The company entered a $25 billion accelerated share repurchase. Current remaining performance obligations hit $33.6 billion, up 14% year-over-year — the forward revenue indicator analysts watch most closely.
And yet shares slipped roughly 3% in after-hours trading, because second-quarter guidance came in below what the street had modeled. For a company trading 32% below its January levels heading into the print, beating on Q1 and dipping on Q2 guidance is the kind of outcome that illustrates exactly how difficult the market finds this company to value right now.
“Agentic AI is the biggest growth opportunity for our customers, and for Salesforce.”
— Marc Benioff, Chair and CEO, Salesforce, Q1 FY2027 Earnings Release, May 27, 2026
The number that actually matters in tonight’s report is Agentforce ARR, which reached $1.2 billion in Q1, up 205% year-over-year. Agentforce and Data 360 combined hit nearly $3.4 billion in ARR — more than 200% above the year-ago level. Salesforce has now delivered 3.8 billion Agentic Work Units to date, a metric the company introduced to track AI agent activity rather than human seat counts. The transition to outcome-based measurement is deliberate: it is how Benioff is arguing to the market that the per-seat model isn’t dying — it’s being replaced by something better for Salesforce’s customers and, eventually, for Salesforce’s revenue.
The Sell-Off Was About a Model Problem, Not a Business Problem
To understand what Salesforce is navigating, the context matters. On February 3, 2026, roughly $285 billion in market capitalization was erased from global enterprise software stocks in a single session. By mid-March, the cumulative damage across the sector had reached an estimated $2 trillion. The catalyst was the accelerating evidence that AI agents — software that executes multi-step workflows autonomously — could replace the human users that per-seat licensing revenue was built around.
Analysts and financial media began calling it the “SaaSpocalypse.” It was not, strictly speaking, an overreaction. The underlying question it raised is a legitimate one: if a company can automate the work of fifty employees using an AI platform for a fraction of fifty software licenses, what happens to the SaaS vendor’s revenue per customer? The bear case on Salesforce is that Agentforce is cannibalizing its own seat base. The bull case is that Agentforce creates net-new value on top of the existing platform. Q1 FY2027 offered the first real data point in a while, and the answer appears to be: mostly bull, with some residual anxiety about Marketing Cloud and Tableau, both of which showed weakness in the quarter.
Bank of America analyst Tal Liani reinstated an Underperform rating at $160 on May 18, calling the shift an AI-driven structural reset. Options traders priced in an 8.7% post-earnings move — nearly double Salesforce’s historical average swing of 3.96% — going into the Wednesday print. The market was genuinely uncertain which way this was going to break, which is unusual for a company of Salesforce’s scale.
ServiceNow Is Running the Same Experiment With Different Results
The contrast with ServiceNow is instructive. Where Salesforce has been the poster child for the SaaSpocalypse repricing — down roughly 32% YTD versus ServiceNow’s approximately 14% decline — ServiceNow has used the same agentic AI narrative to put distance between itself and the fear. In Q1 2026, ServiceNow reported subscription revenue of $3.671 billion, up 22% year-over-year, and raised its full-year guidance. Current RPO hit $12.64 billion, up 22.5% year-over-year.
The Now Assist figure is the specific number worth tracking. On the Q1 earnings call, management disclosed that Now Assist is on a trajectory to exceed its $1 billion ACV target for 2026, with AI-specific commitments now forecast at $1.5 billion for the full year. Now Assist customers spending more than $1 million in ACV grew more than 130% year-over-year. The distinction the market appears to be drawing between ServiceNow and Salesforce is that ServiceNow’s AI products are additive to its existing workflow automation platform, where the agent does more of what ServiceNow was already doing. Salesforce faces the harder question of whether Agentforce is genuinely additive or partially substituting for the human-user seats that underpin its core CRM business.
Google Just Made This More Complicated
Into this environment, Google landed a pricing reset at I/O on May 19 that deserves more attention from enterprise software investors than it has received. Alphabet restructured its AI subscription lineup, cutting the top-tier Gemini plan from $250 to $200 per month and introducing a new $100 tier targeting developers and enterprise buyers. More directly, Google told enterprise customers explicitly that migrating 80% of their workloads from competing AI tools — naming Claude and ChatGPT — to Gemini could save them $1 billion annually in AI spend.
That is not a pricing adjustment. That is a stated competitive strategy against the same enterprise budget lines that Salesforce’s Agentforce and ServiceNow’s Now Assist are pursuing. It also comes as OpenAI, heading into its September IPO window, is offering flat-fee enterprise licensing at competitive rates to lock in customers before going public. The enterprise AI market is now experiencing simultaneous pricing pressure from Google at the platform level, Microsoft through Copilot bundling in M365, and OpenAI through pre-IPO deal structures. The SaaS vendors that built on top of these models face margin compression from two sides: from the hyperscalers setting the underlying price floor, and from their own customers asking whether the agent can do the work without the seat.
What the Renewal Cycle Will Actually Show
The numbers that will resolve this debate are not available yet. They will show up in Q2 and Q3 enterprise renewal data — specifically in net revenue retention rates and cRPO growth trajectories. Salesforce’s Q4 FY2026 report showed that more than 60% of Agentforce bookings came from existing customer expansion, which is the evidence bulls need. Q1 FY2027 raised that metric slightly, with more than 50% of Agentforce and Data 360 bookings from existing customers — a slight moderation worth watching, but not alarming given the faster absolute ARR growth. The question is whether the expansion rate holds as Agentforce moves from early adopter enterprise accounts into the broader mid-market base, where the competitive dynamics with Google, Microsoft, and OpenAI are more acute and the switching costs lower.
For investors who held through the SaaSpocalypse decline, tonight’s Salesforce print offered a partial answer: the business is not broken, and Agentforce is generating real, fast-growing ARR. What it did not offer is confirmation that the seat-based model is fully protected. Q2 guidance came in light, and the stock reacted accordingly. The honest read is that this story has a few more quarters left to run before the market has the data it actually needs to price it with conviction. The valuation framework for agentic AI revenue remains unsettled — and Wednesday’s earnings did more to narrow the debate than end it.
What to Watch Next
- Salesforce Q2 FY2027 results, expected late August — the first quarter where full-year guidance compression becomes either a confirmed trend or a one-time miss. Watch net revenue retention rate and cRPO growth direction more than headline revenue; they will tell you whether Agentforce is expanding customer wallet share or redistributing existing budget.
- ServiceNow Q2 2026 results, expected late July — Now Assist is forecast to exceed $1 billion in ACV for the full year. The midpoint update will confirm whether the $1.5 billion AI commitment figure management disclosed in Q1 is tracking or facing enterprise procurement delays.
- Any named enterprise customer disclosing a vendor switch in AI tooling — specifically any Fortune 500 that moves budget from Claude or ChatGPT to Gemini following Google’s price cut, or from Salesforce to a Microsoft Copilot-native workflow. That announcement, when it arrives, will be the clearest signal about which platform is winning the replacement cycle rather than the expansion cycle.
- Microsoft Copilot monetization disclosure at Build or in Q4 FY2026 results — Microsoft has disclosed a $37 billion annualized AI run rate, but the Copilot paid-seat trajectory remains the most consequential undisclosed metric for understanding whether AI-native enterprise tools are growing the market or cannibalizing existing seat counts across the sector.
- Google Cloud Q2 2026 results and enterprise AI revenue line — Alphabet’s stated promise of $1 billion in enterprise savings will only be credible if Google Cloud accelerates its own enterprise revenue growth in Q2. Watch for whether the Gemini pricing cut shows up as faster subscription growth or simply as margin compression without corresponding share gains.