The funnel is re-accelerating. The stock still prices a seat that may disappear.
Salesforce is going headless: Agentforce, Slackbot and “Claudeforce” let agents work the record without a human screen. cRPO growth accelerated to 14% and the stock jumped about 23% on Aug 27, yet it sits far below its Dec 2024 peak. The market is pricing one question: when the interface dissolves, does the system of record keep the toll?
Gray = actual month-end closes (Oct 2024–Sep 2026, plus Oct 15 close); dots mark intraday 52-week extremes. Log price axis. Colored paths = scenario midpoints at +1/+2/+3/+5 years; clay circle = Street 12-month consensus ($285.68). Scenario weights 30/45/25 bear/base/bull — tilted bear because the AI-disintermediation outcome is binary and the balance sheet is now levered.
Re-weight the scenarios
Drag bear and bull (base takes the remainder). The blended target, dotted chart line and prob-weighted card rows update live.
Four analyst lenses, four answers
Synthesized perspectives, each with its own 12-month target built as EPS × multiple.
The Compounder
cRPO re-accelerating to 14%, Agentforce ARR up 240%, Slack at its fastest bookings growth since acquisition, and management guiding to a second-half organic re-acceleration. Core-plus-AI deserves a market multiple, not a distressed one.
The Cash Counter
~$15.2B trailing FCF on a $203B cap is a ~7.5% yield for a business growing 11%. The ASR has already retired over 100M shares. Net debt of roughly $28B is covered by that cash flow and a stake in Anthropic management says it may sell.
The Seat Squeeze
Per-seat software shrinks when agents do the work, and frontier labs (Anthropic, Meta) are moving into enterprise workflows. Reported EPS leans on investment gains; strip them and the multiple is not cheap. Flat EPS and a de-rating to ~13x gets you the June lows.
The Leverage Check
Debt went from about $14B to about $39B in two quarters; book equity is smaller than goodwill. Interest cut FY27 FCF growth guidance from 9–10% to 4–5%. Fine at today’s cash flow, unforgiving if growth stalls. Fair value near the Morgan Stanley neighborhood.
Sell-side 12-month price targets
Selected firms, sorted low to high. The dashed line is today’s $246.74.
Targets and ratings as reported Aug 27–Oct 12, 2026 (BofA double-upgrade to Buy on Oct 12, $160→$285). Aggregators disagree on the mean: stockanalysis $285.68, MarketBeat $278.12, TipRanks $277.47 (range $165–$400). No sell-rated firm target was verified, so none is charted.
Where the funnel leads
Each price is in-year clean EPS × exit multiple. Returns are versus $246.74. Illustrative frameworks, not forecasts.
Bull case — assumptions & math
FY32 EPS ≈ $29.8 × 24x → $715 · EPS CAGR ≈ 17%Base case — assumptions & math
FY32 EPS ≈ $24.2 × 18x → $436 · EPS CAGR ≈ 12%Bear case — assumptions & math
FY32 EPS ≈ $15.0 × 12x → $180 · 5-yr CAGR ≈ −6%Why “clean” EPS?
$246.74 / $13.6 ≈ 18.1x · vs 14.8x on $16.69 reportedRevenue, capex, free cash flow & debt ($B)
The bull and bear theses both live in the gap between the FCF bar and the debt bar.
FY25–FY26 from 8-K Ex. 99.1 (Feb 25, 2026): revenue $37.9B/$41.5B, capex $0.66B/$0.59B, FCF $12.4B/$14.4B, debt $8.4B/$14.4B. FY27E: revenue guide midpoint $46.25B, capex ~1.5% of revenue, FCF ~+4.5% (management: ~$15B). FY27E debt is the Jul 31, 2026 actual: $33.3B notes + $6.0B term loan (10-Q). Capex is tiny; the risk is the $25B debt-funded ASR, not asset intensity.
EPS path underpinning the targets ($)
Non-GAAP EPS excluding strategic-investment gains: the ladder the base case is built on.
FY25 reported $10.20 (includes a $0.10 loss, so $10.30 clean); FY26 reported $12.52 (includes $0.83 of gains, so $11.69 clean). FY27E ~$13.6 is my estimate (guide $16.69 less ~$3.1 of gains). FY28E onward is the base case at ~13% growth. Fiscal years end January.
The business is still growing
Q2 FY27, year over year. Core lines in olive, frontier AI/data lines in clay.
Revenue +11%, cRPO +14% (cc), FCF +81%, AI+Data ARR +210%, Agentforce ARR +240%, Data 360 records ingested +355%. Frontier lines grow off small bases ($3.9B of ARR against $46B of revenue); organic revenue growth still needs the guided second-half re-acceleration to show up.
Bull vs. Bear
Is Salesforce the system of record agents must pass through, or the seat license they make unnecessary?
▲ THE BULL CASE
- Bookings are accelerating. cRPO +14% cc, near-record-low attrition, and Q3 guided to ~14% before two pending deals.
- AI is now a real line. $3.9B ARR, Agentforce bookings in premium editions more than doubled Q/Q.
- Cash machine. ~$15B FCF; $60B+ cumulative buybacks; dividend raised 5.8%.
- Frontier labs as partners. Claudeforce and AIforce make Salesforce data the thing agents call; Anthropic reported 1.5M actions in a month.
- Long-range targets above the Street. $63B+ FY30 revenue vs $59.2B LSEG consensus.
- Optionality. Management says it will likely sell the Anthropic stake to repay buyback debt.
▼ THE BEAR CASE
- Seat compression. Agents doing the work reduce the licenses a per-seat model depends on.
- Organic growth is ~8%. Informatica adds about 3 points; reported 11% is not organic.
- EPS quality. Q2 non-GAAP EPS +103% includes a $2.6B investment gain; clean EPS is far smaller.
- Leverage. Debt up ~$25B; FY27 FCF growth guidance halved; book equity below goodwill.
- Platform risk. Anthropic and Meta moving into enterprise workflows can bypass the interface.
- M&A churn. Informatica integration plus pending Contentful and Fin deals.
Risk map — likelihood × impact
Over a 3–5 year horizon. The hot corner is likely and high-impact.
Seat compression
Likely × HighAgents replace human seats faster than consumption pricing replaces the revenue.
Model-vendor bypass
Possible × HighFrontier labs own the interface and route work around Salesforce’s record.
Agent data breach
Tail × HighA major incident in agent access to customer data stalls enterprise rollouts and reprices trust.
Debt burden
Likely × Medium~$39B of debt at a higher fixed cost leaves less room if growth slows.
Sector de-rating
Likely × MediumSoftware sentiment swings; the stock fell to $146 in June despite rising estimates.
M&A churn
Possible × MediumIntegration of Informatica, Contentful and Fin distracts from organic re-acceleration.
GAAP noise
Likely × LowInvestment mark-to-market swings headline EPS in both directions.
The jargon, decoded
Hover the dotted terms above, or scan these definitions.
- cRPO
- Contracted revenue Salesforce expects to recognize within 12 months — the best leading read on near-term growth.
- Agentforce ARR
- Annualized value of active AI-agent subscriptions at period end. Since Q2 it includes Slackbot and Headless 360.
- Headless
- Software whose data and workflows are used by agents through APIs, with no human screen required.
- ASR
- Accelerated share repurchase: a block buyback paid up front, here $25B funded by new debt, with 103M shares delivered immediately.
- FCF yield
- Free cash flow ÷ market cap. About 7.5% here.
- Non-GAAP EPS
- Earnings excluding stock comp, amortization and restructuring. Salesforce leaves investment gains in, which is why we strip them.
- Exit multiple
- The P/E assumed at the horizon. Multiply by that year’s EPS to get the target price.
- Prob-weighted
- Each scenario price times its probability, summed into one expected value.