Salesforce CRM — Executive Report

As of 27 Aug 2026, price $252.05 (+22.6% today), post Q2 FY27 print

---

1. Net income over time — the actual shape of the story

GAAP net income per quarter, FY22 Q3 through FY27 Q2. Two step-changes: the FY24 margin reset, and the FY27 Q2 spike.

Year-over-year comparison. Prior-year bars faded.

Key numbers

Metric	Q2 FY27	Q2 FY26	Change
Revenue	$11.345B	$10.237B	+10.8%
GAAP net income	$3.53B	$1.90B	+84%
GAAP operating margin	20.55%	22.78%	−223 bp
Gross margin	76.65%	78.10%	−145 bp
Diluted shares	821M	962M	−14.7%

This is the single most important observation in the entire report: net income exceeded operating income by ~$1.2B. Operating margin fell 223bp year-over-year. The $3.5B print is not an operating achievement — it is a non-operating/tax item plus a 14.7% smaller share count.

---

2. The guidance raise is largely an accounting echo, not a business upgrade

• Adjusted EPS: $5.90 vs $3.27 consensus → beat of +$2.63
• FY27 adj EPS guide raised from $14.06–14.12 to $16.67–16.71 → +$2.60
• Q3 adj EPS guide $3.42–3.44 vs $3.37 consensus → +$0.06
• Q3 GAAP EPS guide $1.81–1.83 vs $1.90 consensus → below street

Revenue guide moved only $46.1–46.4B from $45.9–46.2B — +$200M, or +0.4% — and the company simultaneously flagged a reduced FX tailwind. If you are bullish on this print, be honest about what you're bullish on: it is not upgraded forward operating estimates.

---

3. Segments & KPIs — where growth actually comes from

Salesforce switched to a new "FY27 Revenue Disclosure Framework" this year, collapsing Sales / Service / Marketing / Platform / Integration cloud lines into two buckets. You can no longer track Sales Cloud vs Service Cloud growth separately. Companies do not usually simplify disclosure when the sub-lines look strong — treat this as reduced transparency, not neutral housekeeping.

Segment	Q2'27	Q1'27	Q2'26	YoY
Agentforce Apps	7.2B	6.9B	6.7B	+7.5%
Data 360 / Headless / Other	3.6B	3.7B	3.0B	+20.0%
Prof. Services & Other	525M	540M	546M	−3.8%
RPO	66.3B	67.9B	59.9B	+10.7%

Reads:

• The core application franchise is a ~7.5% grower. Everything above that comes from the data layer, which is where Informatica (closed Q4 FY26, $6B of borrowings) sits. Some of that +20% is bought, not built.
• Professional Services is shrinking and losing money at an accelerating rate: gross profit −$103M on $525M revenue (−19.6% margin) vs −$51M a year ago. Implementation is being subsidised to land agentic deals.
• RPO +10.7% is in line with revenue — no backlog acceleration. Sequentially RPO fell 2.4% (Q1'27 67.9 → Q2'27 66.3); last year the same sequential step was −1.6%.
• Geography: Americas $7.4B (+10.4%), Europe $2.8B (+16.7%), APAC $1.2B (+9%). Europe leads — but that's the leg most exposed to the FX tailwind management just told you is fading.

AI KPIs (from the Q1 FY27 call and Q2 release): Agentforce ARR >$1B; AI + Data ARR $3.4B in Q1, "nearing $4B" per Q2 coverage; 28.6T tokens processed (+152% QoQ); 3.8B agentic work units (+111% QoQ); Slack AWUs +350% QoQ.

Be sceptical here. Usage metrics compounding 100–350% per quarter against revenue growing 11% per year means one of two things: the usage is overwhelmingly bundled/free, or revenue per unit of consumption is collapsing. Both are consistent with the −145bp gross margin. Token counts are not revenue.

---

4. Margins, capital structure and the shift to financial engineering

Metric	Q2 FY27	Q2 FY26
Total debt	$41.7B	$11.2B
Net debt	$30.3B	−$4.1B (net cash)
Total equity	$38.4B	$61.3B
Goodwill	$59.3B	$51.4B
Buyback yield	22.6%	2.0%

In twelve months Salesforce went from a net-cash balance sheet to $30.3B net debt (~2.0x EBITDA) and negative tangible book, in order to fund a $27B buyback and the Informatica deal. Consequences:

• ✅ EPS is structurally boosted: 14.7% fewer shares. Buyback yield 22.6% vs industry median 0.7%.
• ✅ ROE optically improved 3.2% → 7.1% — but largely because the denominator shrank by $23B.
• ⚠️ ROIC is unmoved: 2.25% this quarter vs 2.55% a year ago. Capital returns did not improve the business.
• ⚠️ Interest expense and refinancing risk are now real variables that didn't exist a year ago.

Cost lines: R&D $1.69B = 14.9% of revenue (up from 14.5%); SG&A 40.4% (flat); SBC $906M = 8.0% (up from 7.7%). Cost discipline has stalled; the margin story that carried FY24–FY26 has ended.

Cash flow remains excellent: TTM FCF $15.2B vs $12.5B (+21.6%), P/FCF 10.0x. Dividend $0.44/qtr = 0.70% yield at $252 — immaterial; the buyback is the capital return.

---

5. Valuation

At quarter-end (price ~$184): P/E ttm 16.8x, P/S 3.45x, P/FCF 10.0x, EV/EBITDA 12.4x, P/B 3.95x. Versus the packaged-software industry medians: P/S 3.25x, P/FCF 9.90x, EV/EBITDA 8.5x, P/B 3.31x.

So the "cheap software name" argument, which was genuinely true at $150 in June, is gone. After a 22.6% day CRM trades at a premium to its industry median on every ratio, and its EV/EBITDA is roughly 1.9x the peer median. Note also that the FY27 adjusted P/E of ~15x flatters the business because it embeds the one-off Q2 item; strip a normalised Q2 and forward multiple is materially higher.

Consensus estimates for FY28 in the data ($2.56, $2.62 quarterly) are stale and have not been rebuilt post-print — do not anchor on them.

---

6. Technical picture

• Today: gap open 230.05 from a 205.62 close, high 254.48, close 252.05, on 55.4M shares (~4.6x normal). Unfilled gap $205.62–$230.05.
• Reclaimed the 200-day SMA (200.25) around 19–20 Aug, then extended. 50-day at 178.87 — price is 41% above it.
• Weekly: closed 252.05 vs weekly 200-SMA at 236.79 — first weekly close above that long-term line in this cycle, after a downtrend where the weekly 50-SMA (208) sat below the 200 all year.
• Daily RSI(14) = 79.97 (most stretched reading in the dataset); weekly RSI 69.1.
• Cycle low: $147.58 (week of 22 Jun). From there to today: +70.8%.
• Relative strength vs S&P 500 (1-year percentile): bottomed at 4 in mid-June — CRM outperformed only 4% of the index — recovered to 39.8 on 24 Aug. Still below median. The bounce has not yet made this a leader; it has made it less of a laggard.

Levels worth watching: 254.48 (today's high), 230.05 (gap floor / today's low), the 205.62–230.05 gap, 200.25 (200-day).

---

7. Alternative data & positioning — the most interesting part

Institutions were leaving before this print. Q2 13F filings show heavy net exits: Capital World −8.0M shares, Morgan Stanley −8.4M, BlackRock −6.9M, Capital International −4.5M, State Street −4.1M, Norges −3.2M, Geode −2.4M. Full liquidations: Bridgewater, Soros, Baillie Gifford, Temasek, Ontario Teachers, Parnassus (−1.9M / $305M — the same firm whose CIO publicly warned on 18 Aug that the per-seat model is at risk), Carmignac, AKO, Veritas (down to 51K shares). On the other side: JPMorgan +10.4M (new), Sixth Street +6.6M (new, ~$1.0B), Alyeska +4.7M (new), Qube +3.3M, Susquehanna +3.1M, Crosslink +2.2M, Pacer +2.1M.

⚠️ One row — CalSTRS +186.7M shares / $29.2B — is almost certainly a filing/aggregation artifact (that would be >20% of the company). I am discounting it entirely rather than reporting it as a fact.

TV and sell-side sentiment was at a trough exactly at the lows. Cramer "Not Recommending" and Carter Worth bearish on 21 Jul; Sechan "Sell" 11 Jun. Morgan Stanley downgraded to Equal-Weight on 21 Jul and cut its target from $287 → $185 — then today raised it to $235 while still at Hold. CLSA initiated Hold at $165 on 20 Jul. That is what capitulation looks like in the data.

Insiders were the contrarians. Zero open-market insider sales in the entire Feb–Aug window (everything else is scheduled RSU vesting/conversion). Two directors bought stock outright in March: David B. Kirk 2,570 shares for $500.2K (≈$194.63/sh) and Laura Alber 2,571 shares for $500.3K (≈$194.59/sh). Clean, discretionary, sizeable, at roughly $195 — now up ~29%.

Retail was not the buyer. Retail participation is tiny (0.12%–0.48% of daily volume) and retail net-flow sentiment has been persistently negative (−3 to −6) through all of August. This move is institutional.

Short volume ran 1.1–2.9M/day (roughly 10–20% of volume) and declined into the print (1.5M, 1.1M, 1.4M on 24–26 Aug) — some pre-emptive covering, no squeeze setup.

Dark pool (ATS) volume peaked at 43.4M the week of 15 Jun — precisely the capitulation week — and normalised to 24–28M through late July, consistent with institutional accumulation at the lows.

Congress: Ro Khanna churning $1K–$50K lots repeatedly; Salazar and Armstrong small purchases. No information content.

---

8. Options market — expectations after the move

Spot $252.49. IV is a flat ~39–41% across Sep/Oct/Nov (the earnings premium is out), with 1-DTE still 63–75%. Today's flow was concentrated in near-dated upside: 31.2K contracts at the 250 strike, 22.4K at 260, 11.2K at 240 for tomorrow's expiry. Standing open interest for 18 Sep is stacked below spot — 17.9K at 200, 11.4K at 210, 8.8K at 220 — i.e. a large block of calls that were far out of the money before this print and are now deep in the money. Upside tails exist at 300 (2.4K OI) and 320 (2.9K OI) but are lottery-sized. Roughly 40% IV on 50-day contracts implies about a ±$29 one-standard-deviation move over that window.

---

9. Company-disclosed risks that matter to this thesis

Straight from the 10-K, filtered to what is actually live:

• "The markets and monetization strategies for certain offerings, including Agentforce and Data 360, remain relatively new and uncertain." Salesforce explicitly says it has limited experience determining optimal pricing for consumption-based contracts and may see lower consumption than expected. The entire bull case rests on a monetization model the company itself calls unproven.
• AI-native competitors are named as a distinct competitor class that may "bypass traditional business process workflows or displace established user interfaces." Corroborated in the news flow: OpenAI's "Presence" enterprise-agent product (22 Jul), Curative publicly cancelling a $600K Salesforce contract (10 Aug), Jack Dorsey building a Slack competitor, McAfee replacing ServiceNow with Agentforce ITSM (it cuts both ways).
• AI compute cost is flagged as potentially raising the cost profile of offerings — visible now in that −145bp gross margin.
• Informatica integration + debt service are named risks; $6.0B was drawn under the Informatica credit agreements in Nov 2025.
• Subscription revenue lag: downturns in new business don't show for several quarters. This cuts both directions and is why RPO matters more than reported revenue.

---

10. Bottom line — and where the bias risk sits

What is genuinely, verifiably good: Free cash flow +21.6% TTM to $15.2B. A 14.7% share-count reduction. Five consecutive beats. Directors buying with their own money at $195. Positioning that was washed out into the print. A decisive weekly close above the 200-week moving average. Twenty-nine analyst target hikes in one day.

What the bull case is quietly leaning on that it shouldn't:

1. Net income of $3.5B is $1.2B larger than operating income. Operating margin fell 223bp YoY. The headline is not operating leverage.
2. The FY27 EPS raise (+$2.60) is ≈ the Q2 beat (+$2.63). Forward guidance was raised by essentially nothing, and Q3 GAAP EPS guidance is below consensus.
3. Revenue guidance moved 0.4% while FX turned into a headwind. "H2 organic reacceleration" is a management claim, not yet a datapoint.
4. Core applications grow 7.5%. The 20% grower is the segment that absorbed a $6B-debt-funded acquisition.
5. Gross margin has fallen four quarters straight. Professional services loses money at −19.6% margin and is getting worse.
6. Disclosure got less granular exactly as the sub-segments came under scrutiny.
7. Valuation is no longer the argument. At ~15.8x EV/EBITDA versus an 8.5x industry median, buying today is buying the narrative, not the discount.

The honest framing: this is a high-quality, cash-generative franchise whose reported earnings are being flattered simultaneously by a one-time below-the-line item, a $27B debt-funded buyback, and a segment reclassification — while the underlying operating engine grows revenue ~11%, applications ~7.5%, and is losing gross margin. That can still be a good investment at the right price. It is a very different proposition from "net income nearly doubled."

Two things I could not verify and will not guess at: the specific driver of the ~$1.2B non-operating benefit (the Q2 FY27 transcript isn't available yet — only the Q1 FY27 call is indexed), and the actual dollar terms of the Anthropic "Claudeforce" partnership announced alongside the print.

I can't make the decision for you, and none of this is trading advice — position sizing and risk are yours.

---

Useful next steps I can run:

Pull the Q2 FY27 earnings call transcript as soon as it's indexed to identify the exact driver of the $1.2B below-the-line benefit

Create an alert on CRM for a close back below the gap floor at $230, which would signal the post-earnings move failing

Compare CRM against NOW, ADBE, MSFT and ORCL on gross margin trend and organic revenue growth, to see whether AI-driven margin erosion is company-specific or sector-wide