Wells Fargo’s Michael Turrin raised his price target to $250 from $230 on September 18 while keeping an Equal Weight rating, pointing to Agentforce ARR crossing $1.5B (+240% year over year) as evidence Salesforce is moving from ‘AI as strategy’ to ‘AI as measurable monetization.’ BMO Capital had already lifted its target to $285 from $260 on September 17, keeping Outperform. Both arrived after the six-firm upgrade wave the week before, extending the re-rating into a second week even as the stock itself hasn’t followed.

Salesforce shares closed around $234.30 on September 26 — down modestly from the $237.92 close the week before and still well off the roughly $248 level it traded at before the Fed’s September 16 rate hike. The stock has now spent two full weeks essentially rangebound between $233 and $240 despite the Dreamforce product blitz and a second wave of analyst price-target increases. With consensus still clustered well above $270, the Street’s models and the tape remain in disagreement — the question for Q3 earnings is which one moves to close the gap.

Meta’s Muse, a personal AI agent that connects to a user’s email, calendar, and subscriptions to handle everyday tasks, reached 2.5 million downloads within two weeks of its early-September U.S. launch. It’s the clearest evidence yet that agentic AI has moved past developer novelty into ordinary consumer behavior at real scale — the same shift enterprise software vendors have been betting billions on, now playing out in a Meta app rather than a CRM. For Salesforce and ServiceNow, a consumer-scale proof point that agents genuinely change behavior strengthens the case for their own agentic bets, even as it raises the stakes on who controls the resulting data and access.

Amazon has already restricted Meta’s Muse agent from accessing its services, citing its terms of service — a small skirmish that previews a much bigger fight as agents from Meta, Salesforce, Microsoft, and others all want delegated access into the same email, calendar, and commerce accounts. It is the exact friction Salesforce’s AIforce (with its Trusted Enterprise AI Harness) and ServiceNow’s Action Fabric were built to manage on the enterprise side. Consumer-side, no equivalent governance layer exists yet, and early testing has already surfaced at least one case of Muse pulling in more than a user’s request called for.

Silver Lake’s push to take Workday private in a deal reportedly valued near $51 billion — which would rank among the largest software buyouts ever — is still moving, with CNBC’s David Faber reporting continued financing efforts as of September 17. Workday shares jumped as much as 25% and posted their best single day since 2016 when the talks first surfaced in mid-August; no deal has been announced and talks could still break down. The context matters for the whole SaaS peer set: a private equity buyer is willing to pay a premium for exactly the kind of growth-decelerating, still-profitable software company the public market has spent 2026 repricing downward.

Oracle has fallen roughly 4.7% this month, from the mid-$160s to the mid-$140s, as two pressures converge. Loans backing its Project Jupiter data-center campus in New Mexico are trading at stressed levels on permitting delays and local opposition, and S&P downgraded related credit to just above junk status against roughly $96.3B in long-term debt. Separately, Oracle confirmed another round of workforce cuts, with double-digit percentage layoffs in some teams, sending shares down another 3.8–4.4% on the headlines alone. It’s the clearest single-company illustration of this week’s theme: the bill for the AI infrastructure buildout is starting to show up in credit markets, not just income statements.

Datacom, a major ANZ IT services provider, has begun reselling Rimini Street’s third-party support for Oracle software — a cheaper alternative to Oracle’s own high-margin support contracts. It’s a small deal on its own, but it’s a channel-level signal that Oracle’s most profitable line (support and maintenance) is starting to face real price competition in specific regions, adding a second, slower-moving pressure point alongside this month’s debt and layoff headlines.

This week supplied a clean before-and-after of how capital markets are now grading software companies on balance-sheet risk as much as growth. Workday — a mature, profitable, but growth-decelerating SaaS name — is attractive enough at today’s multiple that a private equity firm is chasing a ~$51 billion take-private. Oracle — aggressively spending on AI infrastructure — is being punished with a near-junk credit view on its debt-funded data-center buildout. Two very different companies, same underlying market verdict: unfunded growth promises are out of favor, and so is unprofitable-on-paper capex, whichever end of the software spectrum it comes from.