Analyst Adam Wood moved Salesforce from Overweight to Equal Weight and slashed his price target from $287 to $185 on July 21 — one of several downgrades this month and the most consequential, given Morgan Stanley’s prior bull stance. The call centers on weak Agentforce channel checks translating into a lower conviction that seat-based revenue holds through the disruption window. At $185, the new target sits barely above the stock’s $170 trading range — a bear that no longer sees much air between price and fair value.

The Morgan Stanley downgrade didn’t move the bulls. Wedbush’s Dan Ives still carries a $475 target — implying roughly 180% upside from current levels — and Piper Sandler’s Brent Bracelin holds an Overweight rating with a $395 target. The spread between the Street’s highest and lowest CRM targets is now roughly $290, one of the widest dispersions among large-cap software names. That gap is the market openly disagreeing on whether Agentforce monetization or seat erosion wins the next four quarters.

Salesforce fell as much as 3.9% intraday before closing at $170.06 on July 21, down 2.2% on the day. That leaves the stock almost exactly where it sat a week earlier ($170.77 on July 17) — the post-June-low recovery has stalled just above the $160 BofA bear-floor target. At -38% year-to-date, Salesforce remains the worst-performing mega-cap software name in 2026, and the Morgan Stanley downgrade landed on a stock that had already stopped climbing.

Following the OpenAI/ChatGPT catalog integration that went GA earlier in July, Salesforce is rolling out Google Search (including AI Mode) and Gemini app integration for Agentforce Commerce through the rest of summer 2026, alongside Storefront Next for spinning up production-ready storefronts in under thirty minutes. The strategy compounds: wherever product discovery happens next — a chat window, a search result, an AI assistant — Salesforce wants to be the commerce backend underneath it, monetizing the same shift in buyer behavior that bears cite as the threat to per-seat software.

Salesforce’s new Agent Script gives builders conditional logic and deterministic controls over how agents behave, pairing the creativity of generative AI with code-like predictability. It’s a direct answer to the enterprise objection that’s slowed Agentforce deals all year: unpredictable agent behavior in regulated workflows. Whether it moves the KeyBanc-style channel checks that flagged weak adoption is the thing to watch heading into Q2 FY27 earnings.

Context for the next print: Salesforce’s last reported quarter (Q1 FY27, ended April 30) showed Agentforce ARR at $800 million, up 169% year-over-year, with 29,000 cumulative deals closed, up 50% quarter-over-quarter. That was the last hard evidence the bulls have that outcome-based and agentic revenue is scaling fast enough to offset seat-based erosion. Late-August Q2 FY27 results are the next checkpoint, and the market’s reaction this week suggests conviction has thinned since May.

ServiceNow reported Q2 revenue of $3.99 billion (up 24% YoY) against a $3.93 billion estimate, with adjusted EPS of $0.90 beating the $0.86 consensus, and non-GAAP operating margin of 29.5% — 300 basis points above guidance. AI-related annual contract value passed $1 billion, ahead of the company’s own long-term pace target. Shares had slipped nearly 5% into the print on pre-earnings jitters tied to the IBM warning; the rebound after hours is the clearest evidence yet that margin discipline, not just top-line growth, is what the market now demands before it forgives an AI-exposed software name.

Away from the IBM-and-Salesforce headlines, the quieter earnings prints tell a steadier story: Box grew revenue 11% — a reliable band for a mature content-management platform — while DocuSign’s FY2027 guidance of $3.5 billion (roughly 9% growth) signals management confidence even past its hypergrowth phase. Neither number supports the “SaaS is being obsoleted” framing driving this month’s selloffs; they support a slower, more differentiated growth phase that the market isn’t yet pricing separately from the AI-disruption names.

IBM’s July 14 warning and 25% single-day crash — its steepest one-day loss in decades — has hardened into a working thesis rather than a one-off scare. Nicholas Mugalli’s framing that this is “the SaaS reckoning arriving not with cancellations, but with deprioritization” is now the reference line for the quarter, and commentators including Jacob Bourne are warning of a “triple whammy” — hardware-first budgets, investor punishment of legacy names, and pressure from AI-native competitors — with Palantir and ServiceNow explicitly named as exposed. ServiceNow’s beat this week is the first data point pushing back on the thesis; whether it holds depends on what the rest of the sector reports through August.

Zoomed out past this month’s selloffs, the global SaaS market is projected to grow from roughly $375.6 billion in 2026 to $1.48 trillion by 2034 — an 18.7% CAGR driven by AI-enabled tooling and rising demand for integrated enterprise applications. That structural growth case is what every bull thesis on Salesforce, ServiceNow, and peers ultimately rests on: the equity market is pricing near-term seat-compression risk while the spending data still describes an industry compounding, not one being replaced.