Salesforce reported its fourth-quarter results on Wednesday and, on paper, delivered almost everything analysts had hoped for. Revenue of $11.20 billion came in line with Wall Street's $11.19 billion estimate. Earnings per share of $3.81 crushed the $3.05 consensus. Free cash flow hit $5.32 billion, a 48% margin, beating forecasts by $250 million. The company simultaneously announced a $50 billion share repurchase authorisation and raised its dividend.
None of it has been enough to shake the stock loose from a months-long decline. Salesforce closed Thursday at $191.75, down roughly 39% from its 52-week high of $313.70, caught in what Wedbush analysts described in a note published this morning as an "AI Ghost Trade" sell-off, a broad rotation out of enterprise software names that has punished the sector regardless of underlying results.
Wedbush maintained its Outperform rating on Salesforce but cut its 12-month price target from $375 to $325, a reduction it attributed to a lower valuation multiple across the sector rather than any change in its view of the company's competitive position.
Agentforce numbers are growing fast from a low base
The most scrutinised figure heading into Wednesday's print was Agentforce revenue, and the result was $800 million in annual recurring revenue, up 169% year-over-year. Since the platform launched, Salesforce has now signed 29,000 Agentforce deals, a figure that grew 50% in a single quarter. The proportion of those accounts actually running in production also grew 50% sequentially, which matters more than deal counts, because it indicates customers are moving from evaluation into live deployment.
Combined Agentforce and Data Cloud ARR reached $2.9 billion, up 200% from $1.4 billion a year earlier. Large deal activity accelerated: contracts worth more than $1 million grew 26% year-over-year, while deals above $10 million increased 33%.
The caveat analysts were quick to raise is that $800 million in ARR, while growing rapidly, remains a small fraction of a company with $41.5 billion in total annual revenue. The question is not whether Agentforce is working; Wednesday's figures suggest it is. The question is how quickly it becomes material to the overall growth trajectory.
The data moat argument
Sitting beneath the Agentforce commercial story is a set of infrastructure figures that Wedbush argues represents Salesforce's deeper competitive advantage. During the quarter the company processed more than 19 trillion tokens through its LLM gateway. Data Cloud ingested 112 trillion records, up 115% year-over-year, including 53 trillion via its Zero Copy architecture, up 310%. A further 18 terabytes of unstructured data was processed.
Salesforce's core enterprise pitch is that agentic AI is only as reliable as the data underpinning it, and that enterprises already trusting Salesforce with their customer data are naturally positioned to extend that trust to AI agents running on top of it. The token and record volumes reported Wednesday give that argument more substance than it had a year ago.
Guidance met expectations but didn't exceed them
For fiscal year 2027, Salesforce guided total revenue of $45.80 billion to $46.20 billion, bracketing the Street's $46.11 billion estimate. EPS guidance of $13.11 to $13.19 was essentially in line with consensus of $13.15. The company expects operating margin of roughly 34.3% and operating cash flow growth of 9% to 10%.
The guidance was not a disappointment in absolute terms. The problem, for investors hoping to see evidence of an inflection, is that the 10% to 11% constant-currency growth outlook for the year ahead includes roughly 3 percentage points of contribution from the Informatica acquisition, which closed during the period. Strip that out and organic growth sits in the high single digits, a respectable rate for a business of this scale but not the acceleration the stock's bulls have been waiting for.
The company also raised its long-range fiscal 2030 revenue target to roughly $63 billion, incorporating Informatica.
A $50 billion buyback at 14.6x forward earnings
The size of Wednesday's capital return announcement is worth pausing on. A $50 billion share repurchase authorisation, replacing all prior unused programmes, from a company trading at 14.6x forward earnings, is a direct statement from management about where it believes the stock should be trading. The quarterly dividend increase to $0.44 per share, up roughly 6% year-over-year, reinforces that posture.
Current remaining performance obligations of $35.10 billion, up 13% year-over-year in constant currency and above the Street's $34.53 billion estimate, point to a revenue pipeline that is building rather than softening. Total RPO reached $72.40 billion, ahead of consensus.
The argument the stock is losing right now
Wedbush's bull case is essentially that Salesforce has been caught in the crossfire of a sector-wide re-rating and that the gap between business performance and share price will eventually close. At current levels, the stock trades at roughly 15x fiscal 2026 earnings, a steep discount to its own history and to peers including ServiceNow, which commands substantially higher multiples despite comparable growth profiles.
The bear case does not require doubting the Agentforce numbers. It simply holds that 29,000 deals and $800 million in ARR, however impressive the growth rates, have not yet demonstrated the revenue acceleration that would justify a re-rating, and that in-line guidance for a year that includes an acquisition-driven boost is not a catalyst. Professional services revenue declined 3% year-over-year during the quarter, a drag that has persisted for several periods.
Wedbush's $325 price target implies roughly 70% upside from Wednesday's close, a return that depends heavily on whether the next several quarters of Agentforce results shift that conversation.