Built by Claude FableThe Ultimate Bull Case
This is deliberately the bull case, not the base case. Every assumption below sits at the optimistic end of a defensible range, and the sum-of-the-parts assumes a multiple re-rating that has not happened. The scored view is 6.15/10 with a half-size BUY -- see the overview. The purpose here is to size the upside honestly so the risk/reward can be judged, and to state plainly what has to be true. The bear case lives on the Concerns & Risks page.
The entire bull case rests on one observation: nobody is underwriting a good outcome. All 33 published FY27 models, all 32 FY28 models and all 6 FY30 models assume Salesforce is a ~9–10% grower forever. The stock trades at 14.9x guided FY27 non-GAAP EPS and 13.5x FY28 consensus — a mature-industrial multiple attached to a 77%-gross-margin, ~95%-recurring oligopoly franchise.
That means the upside does not require heroics. It requires Salesforce to stop being priced as a melting ice cube. Two things move the number, and they compound:
| Lever | Today | Bull FY30 | Why it moves |
|---|---|---|---|
| Revenue growth | ~8.7% organic | 12-13% | Agentforce consumption becomes a second revenue pool on the installed base |
| Non-GAAP op margin | 34.3% guided | 38% | Consumption revenue carries no incremental sales headcount |
| Share count | 839M diluted | ~730M | Continued buyback at an 8.8% FCF yield funded at ~3% after tax |
| Multiple | 14.9x EPS / 11.4x FCF | 22x EPS / 20x FCF | Re-rating to a software multiple once Rule-of-50 is re-established |
Rule of 40 today is roughly 8.7% + 34.8% = 43.5% on organic growth and non-GAAP operating margin. In the bull case it becomes 12.5% + 38% = 50.5% — and software trading at a Rule-of-50 does not clear at 14.9x.
This is the pillar that matters most, because it is the one the market is assigning close to zero value.
Salesforce charges per human seat today. Agentforce introduces a second, orthogonal meter: charging for agentic work performed, measured in Agentic Work Units and sold as Flex Credits. If that meter scales, revenue decouples from headcount — which simultaneously destroys the bear thesis and re-rates the multiple.
The early traction is genuinely steep:
| Agentforce ARR (organic) | $1.2B, +205% Y/Y |
| Agentic Work Units delivered | 3.8B, +111% Q/Q |
| Tokens processed | 28.6T, +152% Q/Q |
| Premium AI SKU bookings (A1E, A4X) | +60% Y/Y |
| Share of AI bookings from existing customers | Over 50% |
Why the distribution advantage is decisive. Over half of Agentforce bookings come from the installed base. Salesforce does not need to win new logos to scale this — it needs to raise the price of relationships it already owns, into budgets it already sits inside, against a data set only it holds. That is the cheapest revenue in enterprise software.
The bull path: $1.2B of Agentforce ARR growing to $7B by FY30. That sounds heroic until you note it requires ~80% annual growth against a business currently compounding at 205%, and that it would represent only ~10% of total revenue. Management has three named monetisation routes — seat upgrades (working now, +60% bookings), consumption credits (early), and the data layer — so the path does not depend on a single mechanism.
The bear case requires seats to shrink. The disclosed evidence says they are growing — in exactly the products most exposed.
- Sales and Service both showed year-over-year seat growth, "with humans and agents both expanding on the platform." These are the two ~$10B-scale franchises and the two most obviously automatable.
- 7 of the top 10 deals added new seats.
- Top-10 deal TCV was ~$800M, 2.5x the same 10 deals a year prior, with incremental annual bookings on those deals +60%.
- 98 deals above $1M of net new annual order value.
- Sales, Service and Slack together are more than 60% of net new AOV.
The single statistic bears cite against this — Agentforce Apps growing +7% cc — is a composition artefact. That bucket contains Marketing (ExactTarget) and Commerce (Demandware), which management has named as a drag on every one of the last six calls for reasons that predate AI by a decade. Strip them and Sales + Service is plausibly growing 8–10% cc. See the thematics page for the full decomposition.
In the bull case, Marketing and Commerce get fixed or divested. At roughly $5.6B of annualised revenue running flat-to-down, they are suppressing the reported growth rate of a 65%-of-revenue segment by an estimated 2–3 points. Removing that drag alone re-rates the perceived growth profile without selling a single new licence.
Agents are only as good as the data they act on, and enterprise customer data is famously filthy. Salesforce bought the remediation layer.
- Data 360 ingested 52 trillion records, +136% Y/Y
- 35 trillion via Zero Copy, +277% Y/Y — federating data without moving it, which is the only architecture large enterprises will accept
- 12 terabytes of unstructured data processed
- Informatica adds $1.1B of Cloud ARR plus governance, lineage and master-data management
This is the pillar where the acquisition case is strongest and least appreciated. Informatica is not a growth-optics purchase; it is the compliance and lineage substrate that makes agentic actions auditable. In regulated industries, "which agent changed this record, on what authority, using what data" is the gating question for deployment. Salesforce now owns the answer.
The bull path: Data 360 plus Informatica from roughly $8B annualised to $13B by FY30 (~18% CAGR), at a higher incremental margin than the applications business because it is consumption-metered.
Slack was widely written off as a $27.7B mistake. In an agentic world it becomes the surface where humans supervise agents — the interface layer, not a chat app.
- Slack MCP surpassed 1 million active users within six weeks of launch. Model Context Protocol is the emerging standard for connecting agents to tools; Salesforce shipped it into an installed base instantly.
- Slack sits inside the Agentforce Apps bucket and grows double digits, materially faster than the +7% cc segment average.
The bull path: Slack from roughly $2.4B to $4.0B by FY30 (~19% CAGR) as the human-agent supervision layer, with attach economics rather than standalone seat economics.
At an 8.8% FCF yield financed at roughly 3% after tax, every dollar of repurchase is accretive by ~580bps. Management has already demonstrated willingness at scale.
| FY27Q1 capital returned | $27.5B ($27.1B buyback + $365M dividends) |
| ASR size / shares retired | $25B / 103M shares (11% of the count) |
| EPS benefit disclosed | $0.23 non-GAAP / $0.14 GAAP in Q1 alone |
| Diluted share count | 871M, down 10.2% Y/Y |
The bear reading of the ASR is that it is financial engineering to defend EPS while growth fades. The bull reading is simpler arithmetic: if the shares are genuinely worth $600 and management is buying them at $194, this is the highest-return capital deployment available to them, and the leverage is serviced ~10x over by operating income. The bull case assumes the count reaches ~730M by FY30, which alone adds ~13% to per-share values.
| Item | The upside |
|---|---|
| Public sector | ARR already above $2B, +23% Y/Y, with public-sector AWUs up nearly 400% Q/Q. Long contracts, low churn, and the segment least exposed to seat compression because headcount is politically fixed. |
| The $7.77B investment portfolio | Treated as an earnings-quality negative elsewhere, and rightly. But it is also a real, marked, liquid asset worth ~$9.50/share that most valuation work ignores entirely. |
| Margin runway | Sales & marketing is still 34% of revenue. Consumption revenue requires no incremental quota-carrying headcount, so incremental margins on Agentforce should run far above the corporate average. |
| Backlog visibility | $33.6B cRPO covers 73% of the FY27 revenue guide before the year starts, and $67.9B total RPO means the downside scenario unfolds over years, not quarters. This is what makes a half-size starter defensible into a binary print. |
| Asset-light AI transition | Capex is ~1.5% of revenue against a 34.7% FCF margin. Unlike the hyperscalers, Salesforce funds its AI build from the income statement -- there is no capex supercycle to swamp free cash flow. |
FY30 (fiscal year ending January 2030) bull revenue build, against a street that models $60.5B:
| Pillar | FY27E annualised | FY30 bull | CAGR | Driver |
|---|---|---|---|---|
| Sales + Service (Agentforce core) | $21.0B | $27.0B | 9% | Seat growth plus premium SKU upgrade |
| Agentforce consumption / Flex Credits | $1.2B | $7.0B | 80% | The second meter; AWU monetisation |
| Data 360 + Informatica | $8.0B | $13.0B | 18% | Zero Copy, governance, lineage |
| Slack | $2.4B | $4.0B | 19% | Agentic supervision interface, MCP |
| MuleSoft + Tableau + Other | $6.5B | $7.5B | 5% | Stabilised, not fixed |
| Marketing + Commerce | $5.6B | $5.6B | 0% | Assumed fixed or divested, not grown |
| Professional services | $2.2B | $2.4B | 3% | Run near break-even by design |
| Total revenue | $46.9B | $66.5B | 12.4% | vs street $60.5B |
From revenue to price
| FY30 bull revenue | $66.5B |
| Non-GAAP operating margin (from 34.3% guided FY27) | 38.0% |
| Non-GAAP operating income | $25.3B |
| Less interest expense (~$39.3B debt, held flat) | ($1.4B) |
| Taxed at the 20.5% non-GAAP rate | $19.0B |
| Diluted shares (from 839M FY27E) | 730M |
| FY30 bull non-GAAP EPS | ~$26.00 |
| At 18x (still a discount to software) | $468 |
| At 22x (Rule-of-50 re-rating) | $572 |
| At 25x (full software multiple) | $650 |
Cross-check on free cash flow. At a 36% FCF margin on $66.5B, FY30 FCF is ~$23.9B. At 20x FCF that is a $478B enterprise value; less ~$27B of net debt gives $451B of equity, or $618/share on 730M shares. The two methods land within 8% of each other, which is the most that should be claimed for a four-year projection.
Bull target: ~$600, range $486–782. From $210 that is roughly +185% over three and a half years, or a ~34% IRR.
| Condition | Confidence | First checkpoint |
|---|---|---|
| 2H FY27 organic revenue reaccelerates as promised | Medium -- unquantified indicator | 2026-08-26 |
| Agentforce consumption becomes a disclosed revenue line, not just an ARR headline | Medium | FY28 guidance |
| Sales and Service seat counts keep growing | Higher -- already disclosed twice | 2026-08-26 |
| Marketing and Commerce stop being a drag | Low -- unfixed for six quarters | FY28 |
| Non-GAAP operating margin reaches 38% | Higher -- 1,940bps already delivered | FY28-FY29 |
| The market re-rates from 14.9x to ~22x | Low -- requires two clean quarters first | FY28 |
2. Agentforce ARR growth decelerates below ~100% before reaching $3B. The second-meter thesis needs to clear the seat-erosion drag, and it cannot do that from $1.2B at 50% growth.
3. Sales or Service seat counts turn negative. This is the falsification of the entire bull architecture, not a trim.
4. A large adverse mark on the $7.77B strategic portfolio breaks the beat streak and exposes how much of the reported EPS growth was never operating. A 10% reversal is ~-$0.70/share.
5. Microsoft prices Dynamics plus Copilot to zero inside E5 renewals. Salesforce has no bundle to answer with.
The honest summary: the margin and buyback legs of this case are high-confidence, and the growth and re-rating legs are not. That split is exactly why the position is sized at half now and half after the print rather than full today — you are paid to wait at 14.9x, and the first checkpoint is five days away.