Because a surprising number of people own this stock (or its ETFs) without quite knowing what it sells, here is a simple breakdown first.
Salesforce rents software seats to companies' sales, service, and marketing teams. When a salesperson logs a call, a support rep opens a ticket, or a marketer fires off a campaign, the system they're clicking around in is often Salesforce (the customer database plus the workflow built on top of it). "CRM" literally stands for Customer Relationship Management, sold as a subscription, priced per user, per month ("per seat"). Over the years they bolted on Slack (chat), Tableau (dashboards), MuleSoft (data plumbing), and now Agentforce (AI "agents" that do some of that work automatically). That's the whole company: \~95% of revenue is subscription, almost all of it enterprises renting seats.
I've done a deep dive on the 10-Q, filed May 28, 2026, covering Q1 2026 (the quarter ending April 30, 2026) & some more!
The interesting thing about Salesforce right now is that it's the mirror image of the hot growth stocks people usually dissect here, a profit-and-cash machine the market is treating like a melting ice cube.
\--------------------------------------
**TL;DR**
* Q1 2026 looked strong on the surface, revenue **$11,133M** (**+13.27% YoY**), operating margin **21.08%**, diluted EPS **$2.42** (**+52.2% YoY**), but read the fine print! The **13%** top line is flattered by acquisitions (Informatica alone added \~**$444M**), and net income is flattered by a **$558M** one-time investment gain and a partly debt-funded **$25B** buyback. The cleaner read is cRPO **+14%** and a **\~21%** operating margin.
* **The case for it:** Cheap vs its own history (**\~3.7x** EV/sales, **\~18x** trailing earnings), expanding margins, \~**$14**.**4B/yr** of free cash flow, a shrinking share count, and a real (if unpriced) AI option in Agentforce.
* **The case against it**: Organic growth is high-single-digits and leans on M&A to look double-digit; the company tripled its debt to **\~$39.5B**, issuing **$25B** of senior notes to fund a buyback of stock at prices well above today's; and the core per-seat model is exactly what AI agents threaten. Management itself flags "decreases in the number of users at our customers" as an attrition risk.
* At **$158.37** (June 26 close), **\~819M** shares post-buyback implies **\~$130B** market cap, **\~$157B** EV. My scenario math lands a base-case fair value of **\~$140–173**, so today's price sits in the upper half of fair. The game again is whether AI is a tailwind it monetizes or a tax on the seats it sells.
\--------------------------------------
**The numbers are strong, but read them carefully!**
Q1 2026 (quarter ended April 30, 2026) vs the year-ago quarter:
* Revenue **$11,133M**, up **13.27%** YoY (**$9,829M** a year ago)
* Income from operations **$2,347M**, a **21.08%** operating margin, up from **19.76%** (**$1,942M**)
* Net income **$2,107M** (**+36.73%**)
* Diluted EPS **$2.42** vs **$1.59** (**+52.2%**)
* Effective tax rate **\~23%**, a clean and normal rate (no distortion here)
Three flags I need to mention, because the headline overstates the underlying business:
1. The **13%** growth is partly bought, not earned. The filing lists three recent deals baked into the quarter: *"our April 2026 acquisition of Qualified.com... our November 2025 acquisition of Informatica... and our October 2025 acquisition of Regrello."* It even quantifies one: the Informatica acquisition *"contributed approximately* ***$444 million*** *of total revenues"* this quarter, about **4.5** points of the **13.3%** growth. Strip the M&A and organic growth is roughly **\~9%**, consistent with the full-year trend (FY ending Jan 2026 revenue was **$41,525M**, **+9.58%**).
2. Net income is flattered above the operating line. Below operating income, the quarter carried a **$558M** net gain on strategic investments (vs a **$63M** loss a year ago, a **$621M** swing). It's two roughly equal halves: a **$268M** realized gain from exiting one private holding and a **$268M** unrealized mark-to-market gain on another (the filing also nets out **$119M** of impairments). Either way it's non-operating and largely non-recurring. It was partly offset by interest expense jumping to **$317M** from **$68M**. Salesforce took on new debt, including a **$6.0B** term loan drawn in March 2026, so total debt is now **\~$39.5B**. The cleaner profitability read is the **21.08%** operating margin, not net income.
3. The EPS lift came with a debt-funded buyback. Per the MD&A: *"Our* ***$25 billion*** *Accelerated Share Repurchase ('ASR Agreements') executed in March 2026 resulted in the repurchase of approximately 103 million shares in the period and benefitted our diluted net income per share by $0.14."* Where did the **$25B** come from? The same filing is explicit: *"In March 2026, we also issued unsecured Senior Notes with an aggregate principal of $25.0 billion... We used the net proceeds from the March 2026 Notes to fund an accelerated share repurchase program."* So Salesforce funded the entire buyback with newly issued debt, maturities stretching to 2066, to retire shares in March 2026, when the stock was well above today's **$158**. The diluted share count is genuinely down **\~10%** YoY (**870.7M** vs **969.2M**), which is real and shareholder-friendly; just know it was bought on the balance sheet, not purely with cash.
**What's really driving it: cRPO and buybacks**
For a subscription business the number that leads revenue is the contracted backlog, and that's the genuinely healthy signal:
* Current Remaining Performance Obligation (cRPO) **$33.6B**, **+14%** YoY. Contracted revenue to be recognized over the next 12 months.
* Total RPO **$67.9B**, **+11%** YoY
* Subscription & support is **\~95%** of revenue, recurring and sticky
* Operating cash flow was **$6.7B** in the quarter (Q1 is the seasonal collections peak) against capex of just **$145M** (**1.30%** of revenue). An asset-light cash machine throwing off roughly **$14B/yr** of free cash flow.
* Capital return is now core: the **$25B** ASR plus **$365M** of dividends in the quarter. Share count has fallen **971M** (Jan 2024) → **962M** → **929M** (Jan 2026) and again after the March ASR, to **\~819M** baseline following the upfront delivery of the recent ASR.
The honest framing is Salesforce has shifted from a growth story to a margin-expansion and capital-return story. Management is explicit: *"We are also focused on reducing our operating expenses to improve our operating margin."* While this mirrors the classic activist playbook (originally pressured by firms like Elliott and Starboard), it represents a stable, mature enterprise engine focused on efficiency over landing raw new logos.
**The biggest structural risk: the seat model, meet AI (their own words)**
Here's what should keep a CRM bull up at night, and it's the crux of the "SaaS crash" debate. Salesforce charges per human seat. The promise of AI agents, including its own Agentforce which is priced per "Agentic Work Unit (AWU)", is that software does work humans used to do. If that's real, customers may need fewer seats. Management flags exactly this in RiskFactors (Item 1A):
>"It is difficult to predict attrition rates given our varied customer base, the number of multi-year subscription contracts, and our shift toward consumption-based pricing models. Our attrition rates may increase or fluctuate as a result of various factors, including... decreases in the number of users at our customers... and economic downturns."
So Salesforce is simultaneously (a) selling AI as its next growth driver and (b) acknowledging a shift away from per-seat toward consumption-based pricing, quietly hedging the model it was built on! They also list the risk of *"any failure to expand our services and to develop and integrate our existing services in order to keep pace with technological developments."* Agentforce is both the answer to the AI threat and the thing that could cannibalize the seat count.
**The AI/Agentforce narrative is real, but mind the relabeling**
This quarter Salesforce reorganized its subscription reporting into two buckets:
* **Agentforce Apps** (**$6.91B | +9% YoY**):
* The rebrand: this is a renaming of the legacy core portfolio (Sales, Service, Marketing, and Commerce clouds) with Slack added in.
* The growth: despite the new AI-centric name, this bucket grew just **9%**, reflecting the maturity of the core CRM business.
* **Data 360, Headless Platform, & Other ($3.68B | +25% YoY)**:
* Constituents: includes Data 360, MuleSoft, Tableau, and the newly acquired Informatica.
* M&A impact: the high **25%** growth was heavily bolstered by Informatica, which contributed **$444** million of revenue this quarter.
Isolating "pure" AI revenue:
* **ARR vs. revenue**: the headline **$1.2 billion** Agentforce ARR (**+205%** YoY), a figure management reports in its earnings materials, not a line item in the 10-Q, is an annualized contract value, not the revenue actually recognized this quarter.
* **Monetization lag**: because Agentforce uses a consumption-based model, actual recognized revenue typically lags ARR by one to two quarters. Currently, Agentforce ARR is only **\~3%** of total annualized subscription revenue (**$1.2B** against **\~$42.4B** annualized).
**The takeaway**: don't read "**$6.9B** of Agentforce" into that first bucket, it's the rebranded core apps. The AI excitement is a product story and a fast growing but "tiny" ARR line, not yet a material recognized-revenue engine. Treat Agentforce as an unpriced option, not a current growth driver. Anyone modeling it as today's growth is modeling a rebrand plus a 3%-of-revenue contract base, not the income statement.
**Where the big money actually moved (13F filings, positions as of March 31, 2026)**
The cleanest fact first. Across the large 13F filers, aggregate shares held were roughly flat (**\~500M** shares, **-4.4%** YoY) while the dollar value of those holdings fell **\~33.5%** YoY. When shares are flat but value craters, it's the price that left, not the institutions. This wasn't a stampede for the exits.
Underneath that:
* The index complex is steady-to-higher: Vanguard **\~86.3M** → **88.1M** shares over eight quarters, BlackRock **74.6M** → **79.7M**, State Street \~flat at \~**50M**, Geode **19.5M** → **22.2M**. Passive money isn't going anywhere.
* The notable active move is a value shop stepping in: Harris Associates, a deep-value manager, cut to just **0.33M** shares in early 2025, then rebuilt to **14.92M** shares by March 2026. Value money buying the de-rating.
* What I won't consider as clear signals are the big jumps from Arrowstreet (**0.7M** → **12.7M**, a quant) and Morgan Stanley (**19M** → **31.7M**) that look like factor/inventory flow, not conviction. Flagging them here but they might not be a result of long-term hold.
**Caveat the lens honestly**: 13Fs only capture large institutional filers and reflect March 31, 2026 positions , before the most recent leg down, so this is neither every owner nor fully current. But the direction of the biggest holders is informative: indexers steady, a value manager accumulating, no broad institutional exit.
**Valuation is cheap vs its own history, but fairly priced once you count the debt**
At **$158.37** (June 26 close), with shares cut by the March ASR to **\~819M** (cover page), market cap is ≈ **$130B**. Add **$39.5B** total debt (including the **$6.0B** term loan drawn in March 2026), subtract **$11.8B** cash & marketable securities implies net debt of **\~$27.7B** and enterprise value ≈ **$157B**.
On a trailing-twelve-month (TTM) basis through Q1 2026 (Q2 25 + Q3 25 + Q4 25 + Q1 26): revenue **$42,829M**, net income **$8,023M**, diluted EPS **$8.63**.
* P/E (TTM) ≈ **18.4x** ($158.37 ÷ $8.63). Strip this quarter's one-time **$558M** gain (\~$0.49/sh after tax) and "core" TTM P/E is closer to **\~19.5x**.
* P/S ≈ **3.0x** ($130B ÷ $42.8B);
* EV/Sales ≈ **3.7x** ($157B ÷ $42.8B).
For context, that's a fraction of the double-digit sales multiple the market paid Salesforce at the 2021 software peak, the re-rating is the whole story. Today's multiple is roughly in line with slower-growth mature software, which is exactly the point of contention.
**In conclusion, here's what I keep going back and forth on:**
Salesforce's entire model is renting human seats. Its biggest bet, Agentforce, is a software designed to do work humans used to do. Those two things point in opposite directions: if Agentforce really works, does Salesforce sell more (every customer bolts agents on top of their seats) or less (customers quietly cut the seats the agents replace)?
**Figures from the 10-Q filed May 28, 2026 ( Q1 2026, ownership from 13F filings as of March 31, 2026; price as of the June 26,2026 close.**
**I expanded this into a** [full write-up on my Substack](https://open.substack.com/pub/secaura/p/the-case-on-salesforce-stock-crm?r=1jjx11&utm_campaign=post-expanded-share&utm_medium=web) **with the P/S de-rating chart, the debt-maturity schedule, the reverse-DCF, and the bull/bear scenario grid.**
**Disclaimer: This is an analysis of the SEC filings for educational. This is not financial advice. Do your own due diligence (DD).**
