On Tuesday, Citi analyst Tyler Radke raised his CRM price target from $187 to $204, maintaining a Neutral rating and citing channel checks showing modest improvement in Agentforce deployment ahead of the company’s fiscal second-quarter earnings on August 26, 2026. The stock rose as much as 3.82% on Tuesday, extending a recent software-sector rotation that has taken CRM up nearly 15% over the past month. An investor watching that number might conclude the upgrade cycle is finally working. The pattern of 2026 says otherwise.
Seven weeks and counting, Wall Street’s most prominent bull calls on Salesforce have each produced a pop and then given it back. JPMorgan’s Samik Chatterjee resumed coverage at Overweight with a $250 target on August 13, 2026 — the loudest bullish call from a major bank since Guggenheim’s initiation in early July — and the stock reached $201.37 that day before reversing to $196.21 on Friday, erasing most of the gain and finishing below where it traded before Chatterjee’s note landed. That reversal is not a data point. It is a pattern. Until August 26 produces a specific number — cRPO growth rate and Q3 guidance — that confirms Salesforce’s promised second-half revenue reacceleration, the stock has signaled it will treat every analyst upgrade as temporary noise.
For investors still holding CRM, or deciding whether to buy ahead of earnings, the relevant question is not what Citi or JPMorgan think. It is what the August 26 report must actually deliver to turn analyst opinion into durable price recovery.
Every Upgrade Since June Has Reversed
The upgrade-reversal pattern began in earnest after Salesforce’s Q1 FY2027 earnings beat on May 27, 2026 — adjusted EPS of $3.88 against a $3.13 consensus, revenue of $11.13 billion growing 13.3% year-over-year — when the stock initially rose and then fell back as investors focused on the cash flow guidance cut. In the weeks that followed, Citigroup upgraded CRM to Buy in early July. Guggenheim moved from Neutral to Buy with a $228 target. Monness Crespi lifted its target to $222 on August 13. Not one of those moves produced a rally that held.
The most striking illustration came from Morgan Stanley. Analyst Adam Wood downgraded CRM to Equal Weight and cut his price target from $287 to $185 on July 21, 2026, citing the disconnect between strong Agentforce metrics and insufficient progress in contracted revenue. The downgrade was accurate about the market’s near-term read on the stock — CRM fell — but the subsequent upgrade waves from JPMorgan, UBS, and Wells Fargo failed to close the gap that Wood identified. That gap is not about sentiment. It is about one specific metric: cRPO.
Soros Fund Management disclosed via its Q2 13F filing, released August 14, 2026, that it had exited its Salesforce CRM position during the second quarter. The move was not an announcement of collapse; institutional rebalancing rarely is. But it followed the same logic as the 1,621 institutional investors who reduced their Salesforce positions in the most recent reporting quarter against 1,083 who added shares — what the pension fund and long-horizon investor community is expressing is not conviction that Agentforce will fail, but uncertainty about timing that analyst upgrades cannot resolve.
What an Analyst Upgrade Cannot Tell an Investor
When Chatterjee argued at JPMorgan that Salesforce controls “scarce inputs” — enterprise data context and trust — that raw language models cannot replicate, he was making a structural argument about competitive moats. The argument is reasonable. It is also invisible in the data investors can actually see in 2026.
The $85 spread between Morgan Stanley’s $185 bear target and JPMorgan’s $250 bull target reflects a genuine disagreement about timing, not a disagreement about facts. Both the upgrade camp and the downgrade camp agree that Agentforce ARR reached $1.2 billion growing 205% year-over-year. Both agree that legacy products like Tableau and Commerce Cloud showed weakness in Q1, with management explicitly citing both as partial offsets to forward-revenue growth. What they disagree on is whether the combination of Agentforce expansion bookings — more than 50% came from existing customer accounts in Q1 — is outpacing legacy attrition fast enough to accelerate the contracted-revenue pipeline. That question has a specific answer in the data. It is called cRPO.
Current Remaining Performance Obligations measure contracted revenue Salesforce expects to recognize within the next 12 months, capturing booked commitments before they show up in recognized revenue. It is the closest thing enterprise software companies report to a forward-revenue odometer. In Q1 FY2027, cRPO grew 14% year-over-year to $33.6 billion. Analysts at Morgan Stanley and KeyBanc cited the absence of cRPO acceleration — specifically, the failure of Agentforce booking momentum to show up in contracted forward-revenue growth — as the concrete reason they downgraded. Analyst upgrades cannot manufacture that number. The August 26 report will contain it.
The $27.5 Billion Signal No One Knows What to Do With
Salesforce’s leadership has not been passive during 2026’s repricing. In February 2026, the board approved a $50 billion share repurchase authorization. In March, the company executed a $25 billion accelerated share repurchase — the largest accelerated share repurchase in corporate history — financed partly through a $25 billion debt issuance. By the time Q1 FY2027 earnings were reported, Salesforce had returned $27.5 billion to shareholders in a single quarter, including $27.1 billion through repurchases and $365 million in dividends. For perspective: that sum exceeds the company’s total free cash flow generation across the entire prior fiscal year. It reduced the diluted share count by 10% year-over-year.
The signal management was sending is clear: the current stock price represents a historic buying opportunity. Marc Benioff has publicly called the stock “quite low.” The market’s response has been ambivalent. The $25 billion debt used to fund the ASR also forced management to halve its cash flow growth guidance — from approximately 9.5% to 4% to 5% — and trim its GAAP operating margin outlook from 20.9% to 20.6%. A company returning capital at a rate that exceeds its free cash flow generation is betting heavily on its own valuation. The question investors are waiting to answer is whether that bet is confirmed by the fundamental metrics in Q2.
How Agentforce Bills — and Why It Creates Accounting Uncertainty
Understanding what August 26 must deliver requires understanding why Agentforce’s spectacular growth metrics have not yet moved the contracted-revenue needle in a way that satisfies the upgrade camp’s critics.
Agentforce runs on what Salesforce calls the Atlas Reasoning Engine, an event-driven, directed-graph architecture where planning, data retrieval, tool invocation, and response generation execute concurrently through a publish-subscribe asynchronous workflow. When a Veterans Affairs employee asks the system to verify a veteran’s benefits status, the Atlas engine classifies the intent, queries the relevant data sources, executes the verification through an API call, and returns the response — without a human touching the queue. One completed cycle of that process is what Salesforce calls an Agentic Work Unit, or AWU. Enterprises pay per AWU rather than per named user seat.
This is a fundamentally different economic structure from the seat licensing that built Salesforce’s $41.5 billion revenue base. Under seat licensing, a contracted amount books to cRPO at signing and is recognized each month. Consumption contracts — where the customer pays for actual AWU volume rather than a fixed headcount of users — may produce different cRPO dynamics. When a customer commits to a fixed AWU spending commitment, it books similarly to a seat contract. But flex-credit and open-ended consumption arrangements may book lower committed amounts to cRPO even when actual usage significantly exceeds the baseline. This timing lag — between Agentforce ARR growing 205% and that growth fully appearing in the contracted revenue pipeline — is the structural gap Morgan Stanley and KeyBanc identified.
The deeper constraint is what might be called the AWU efficiency paradox. As the Atlas Reasoning Engine improves — better intent classification, faster tool invocation, more accurate routing between specialist Language Action Models — it completes the same enterprise workflow in fewer processing steps. Fewer steps means fewer billable AWUs, even though the customer receives equivalent or greater task completion. A more capable AI agent generates less revenue per workflow than a less capable one, even when customer satisfaction improves. This is a structural feature of consumption billing on an improving AI platform, not a temporary transition problem. Salesforce is currently running three concurrent Agentforce pricing architectures simultaneously — per-AWU, flex-credit pools, and per-user seat editions for Agentforce Sales and Service — which signals the company has not yet settled on which model best aligns its revenue with customer value as AI capability advances.
The Federal Proof Points — and the Counter-Signal
The bull camp’s most durable argument does not rest on ARR percentages. It rests on what type of institution has committed to Agentforce at scale.
In July 2026, the U.S. Department of Veterans Affairs awarded Salesforce a three-year $1.6 billion AELA, deploying Agentforce Public Sector and Agentforce Health across the VA’s network of 170 medical centers and more than 1,100 outpatient clinics. The contract requires Government Cloud Plus Defense at IL5 clearance — a federal security certification that eliminates vendors who have not been evaluated against government-grade standards. That certification was earned through extended vendor evaluation processes, not a product demonstration. Earlier in 2026, the U.S. Army awarded a $5.6 billion, 10-year Indefinite Delivery Indefinite Quantity contract through Computable Insights LLC, Salesforce’s national security subsidiary. The U.S. Air Force deployed Missionforce National Security to manage an $84,000-vehicle fleet across 389 global locations.
Federal agencies do not sign multi-billion-dollar technology commitments on speculation. The VA, Army, and Air Force commitments are evidence that Agentforce performs reliably enough at scale for government-critical workflows.
The counter-signal arrived August 11, 2026, when The Information reported that the U.S. Department of Agriculture — which spends approximately $3 billion annually on IT and uses Salesforce for nearly every function — is tapping C3 AI specifically to reduce its Salesforce footprint. The USDA report matters precisely because Salesforce’s 2026 bull thesis has leaned on federal validation. A federal agency actively reversing course doesn’t erase the VA and Army commitments, but it complicates the “government signals universal confidence” narrative.
What August 26 Must Deliver
Salesforce is scheduled to report Q2 FY2027 results on August 26, 2026, after market close. The analyst consensus across roughly 53 covering firms projects adjusted EPS of $3.27 to $3.28 and revenue near $11.3 billion — representing year-over-year growth of roughly 12.4% in earnings and 10.4% in revenue. Salesforce has beaten adjusted EPS in each of its four prior consecutive quarters, with the most recent miss rate running near zero; another EPS beat is the baseline expectation, not a catalyst.
What actually matters is two specific numbers: the Q2 cRPO growth rate in constant currency, and Q3 revenue guidance. If cRPO growth accelerates from its Q1 pace of 14% — particularly if it shows that Agentforce bookings are outpacing legacy attrition in the forward-revenue pipeline — the bull case gets its first empirical confirmation. If Q3 guidance implies organic subscription revenue reacceleration in the second half of FY2027 as management has committed, the $85 gap between Morgan Stanley’s target and JPMorgan’s becomes narrower. If cRPO decelerates even alongside another clean EPS beat, the market will conclude that the Agentforce transition is not yet delivering contracted-revenue growth at the pace the bull thesis requires — and the discount applied to the stock over the past year will appear earned rather than opportunistic.
Full-year revenue guidance of $45.9 to $46.2 billion stands confirmed. Salesforce’s longer-dated goal is $63 billion in revenue by fiscal 2030 including Informatica, alongside a “Rule of 50” framework requiring the sum of subscription and support growth rate and non-GAAP operating margin to reach 50 by that year. JPMorgan’s Chatterjee specifically identified the 4.7-point gap between Salesforce’s current Rule of 50 score and that benchmark as the rerating mechanism the stock depends on. That rerating requires acceleration — not just stability. The August 26 report will offer the earliest visible evidence of whether acceleration is arriving on schedule or has slipped.
For a company trading at approximately $195 per share after returning nearly $28 billion to shareholders in a single quarter, generating free cash flow at an 8.8% annual yield relative to its current market value, and guiding to record annual revenue — the case that the stock is cheap is straightforward. The case that the stock deserves to be cheap — that the market’s skepticism is rational, not excessive — rests on the same single question: whether the consumption-based revenue model Agentforce requires can accelerate contracted revenue fast enough to validate the bull targets every analyst has been issuing and every rally has been erasing.
August 26 is the first event this year that will give an answer.
This article is for informational purposes only and does not constitute investment advice. Readers should conduct their own due diligence before making investment decisions.
Frequently Asked Questions
Why have analyst upgrades failed to hold Salesforce’s stock in 2026?
Salesforce analyst upgrades — including JPMorgan’s Overweight resumption at $250 on August 13, Guggenheim’s July move to Buy at $228, and Citi’s target raise to $204 on August 18 — have each produced short-term gains that reversed within days. The reason is structural: analyst upgrades express a view about where the stock should trade given the company’s prospects. But the specific metric investors are waiting for — cRPO growth accelerating in constant currency, confirming that Agentforce bookings are flowing into contracted revenue — cannot be delivered by an analyst note. It can only be delivered by the quarterly report. Until August 26 produces that data, the market has consistently treated every JPMorgan’s $250 Overweight coverage note as an entry opportunity to sell into rather than confirmation of durable recovery.
What is cRPO, and why does it matter more than Agentforce ARR for investors right now?
Current Remaining Performance Obligations, or cRPO, is a GAAP metric representing contracted revenue Salesforce expects to recognize within 12 months — including both billed and unbilled amounts under existing contracts. It is the most direct visible measure of whether Agentforce bookings are outpacing the softness in legacy products like Tableau and Commerce Cloud in the forward-revenue pipeline. Agentforce ARR reached $1.2 billion growing 205% — that sounds compelling, but it still represents only about 2.6% of Salesforce’s $46 billion guided annual revenue. Until that 2.6% shows up as an accelerating cRPO growth rate, the upgrade camp’s targets are grounded in a future inflection that the reported numbers have not yet confirmed.
What is the AWU efficiency paradox — and why does it matter for Salesforce’s long-term revenue?
An Agentic Work Unit (AWU) is one completed task cycle through Salesforce’s Atlas Reasoning Engine architecture — the event-driven, directed-graph AI system that processes enterprise requests without human intervention. Enterprises pay per AWU rather than per seat. The efficiency paradox is this: as the Atlas engine improves and completes the same workflows in fewer processing steps, each customer generates fewer billable AWUs per task — even as the AI delivers equivalent or better outcomes. A more capable Agentforce produces less revenue per workflow than a less capable one. This is a structural feature of consumption pricing on an improving AI platform, not a temporary scaling problem. Salesforce’s current three Agentforce pricing architectures — per-AWU, flex-credit pools, and per-user seat editions — suggest the company is actively searching for a model that aligns its revenue with customer value as AI efficiency advances.
Should investors buy Salesforce stock before August 26 earnings?
This article is for informational purposes only and does not constitute investment advice. The factual picture is: Salesforce stock trades at approximately 14 times forward earnings — well below its five-year median price-to-earnings ratio of 75.5 times and below the S&P 500 median — generating free cash flow equal to roughly 8.8% of its market value annually. The analyst consensus across 53 firms rates CRM as a Buy with average 12-month targets of $241 to $250. The risk to buying ahead of earnings is specific: if August 26 Q2 results show cRPO growth decelerating or management lowering its second-half revenue reacceleration outlook, the market has demonstrated throughout 2026 that it will sell the stock regardless of what analysts say. Investors who wait for the August 26 data accept the risk of buying after a potential post-earnings rally; investors who buy before accept the risk that another upgrade-cycle pop reverses if the cRPO number disappoints. UBS, maintaining a Neutral rating, has warned that the bar for second-half acceleration is high after the recent rally, and that any tempering of the narrative would not be tolerated by the market. Either decision should be made with eyes open to that specific binary.
Salesforce (NYSE: CRM) is scheduled to report fiscal second-quarter 2027 results on August 26, 2026, after market close. This article is for informational purposes only and does not constitute investment advice.
ⓒ 2026 TECHTIMES.com All rights reserved. Do not reproduce without permission.
