Salesforce Looks Cheap Against Its AI Software Peers — the Bookings Are the Catch

Salesforce has spent 2026 in the market's doghouse, and then one report in late August changed the language. The stock that began the year near $254 and bled to a 52-week low of $146.32 and jumped more than 20% in a single day on August 27 — its second-best session ever — after the company paired a strong earnings beat with an expanded partnership with Anthropic, the OpenAI rival behind the Claude chatbot. Even after that pop, the shares still sit below where they started 2026. The question for anyone watching from the sidelines is whether the August report really changed the investment case, or whether it just gave a beaten-down software giant a flashy AI headline.

The reason Salesforce got crushed in the first place matters for reading what just happened. For much of 2026, investors dumped software names on a fear Wall Street has taken to calling "SaaSpocalypse" — the worry that generative AI would let companies replace per-seat software subscriptions with far cheaper AI agents, hollowing out the SaaS revenue model. Salesforce is the purest big-cap hostage to that fear: it sells seats. Marc Benioff spent the August call doing his best to bury the narrative, saying the dire predictions "have none of them come true for us" over the previous two quarters.

What replaced the fear is a concrete monetization story. Salesforce's Q2 for fiscal 2027, for the period ended July 31, came in at $11.35 billion in revenue, up 10.8% year over year, with non-GAAP earnings of $5.90 per share against a $3.27 consensus. More important than any single quarter, management pointed to the pipeline: current remaining performance obligations (cRPO), a leading indicator of contracted revenue to come, accelerated to 14% growth. The centerpiece of the bull case is Agentforce, Salesforce's AI agent platform, whose annualized recurring revenue has passed $1.5 billion, up more than 240% year over year; combined with the Data 360 data product, ARR is nearly $3.9 billion, up over 210%. Management raised full-year revenue guidance to $46.1 billion to $46.4 billion.

Here is where a systematic screen matters more than the story. On the metrics that actually drive a rating, Salesforce looks like the relative value in large-cap AI software — not because it is cheap in absolute terms, but because it is cheap against the peer set. It trades around 4.6 times trailing sales and roughly 21 times trailing earnings, with an enterprise value of about 18 times EBITDA. Compare that to the neighbors in the same narrative trade: Snowflake fetches roughly 21.6 times sales despite an unconventional earnings profile, Oracle trades near 6.8 times sales and 27 times earnings, and Microsoft, by the comparison consensus, sits around 25 times forward earnings. If you want exposure to the AI-monetizes-SaaS idea, Salesforce is the discount aisle.

The profit-quality read backs up the valuation, with one important asterisk. On the operating side the grades are strong: a 77.6% gross margin, a 20.4% operating margin, and — the cleanest number in the report — free cash flow of $15.15 billion over the trailing twelve months, a 34% margin with cash flow growing over 21% year over year. Here is the asterisk: reported GAAP earnings of $4.29 per share were inflated by $2.43 per share, or $2.61 billion, of one-time strategic investment gains. The tell is visible in the multiples themselves — the forward P/E of roughly 29 sits above the trailing 21, which only happens when analysts expect next year's earnings to come in below this year's flattered figure. The cheap trailing multiple rests partly on non-recurring gains. The free cash flow does not, which is why the cash number is the honest anchor.

The momentum and revision picture is what actually flipped over the past quarter, and that is the part worth treating as a genuine change in the report card rather than noise. The stock is up roughly 27% over the past 20 sessions, sitting well above both its 50-day ($197) and 200-day ($201) moving averages, with a relative-strength index around 64 — strong, but not yet screaming overbought. Salesforce has now beaten estimates six straight quarters, and guidance went up, not down. A name that traded from eroding to improving on both momentum and revisions in a matter of weeks is, in a systematic framework, more actionable than a static top scorer.

What the factor stack actually argues for is a placeholder in the growth-and-Tech sleeve, sized with the honest caveat acknowledged. The bull case depends on the pipeline becoming revenue: Agentforce and CloudForce (the Anthropic co-branded offering) carry a 60% to 80% premium on premium editions, and only about 5% of knowledge workers have upgraded so far — that is the runway, and also the risk. If the 240%-growth AI bookings convert into broadly visible revenue, roughly 21 times trailing earnings with an 11%-grossing base business and triple-digit agent growth is a reasonable relative price for a mega-cap. If CloudForce adoption stalls, the cheaper-looking multiple turns out to have been dressed up by one-time gains and the fear that drove the stock to $146 was not entirely wrong. The balance sheet — net debt of about $28 billion against that $15 billion of free cash flow, plus an aggressive buyback taking the diluted count down to roughly 821 million shares — can carry the company through either scenario. The mistake would be buying the headline "cheap" without the cash-flow proof; the reasonable position is owning the story on the cash flow, with the 10.8% revenue base growth and the 240% agent-bookings growth as the two numbers that will decide whether the re-rating holds.