SaaS & Salesforce Intelligence Digest

Dan Ives, now at Yorkville Ives, initiated Salesforce at Outperform with a $290 target, about 27% above Friday’s $229.13 close. His argument is that CRM has been swept into the SaaS-disruption narrative unfairly: autonomous agents need trusted customer data, permissions, workflow logic and an audit trail, and Salesforce owns more of that stack (Data 360, MuleSoft, Informatica, Slack) than almost any application vendor. The target is well below the $475 he carried at Wedbush, and that is the telling number: even the Street’s loudest bull has reset expectations. The stock did not respond. CRM fell 2.4% on the week while IGV rose 3.9%.

Benioff said on Friday that the AIForce agent platform is now SIForce (“The era of Super Intelligence is here”), following Elon Musk in adopting the term from the President’s September 29 executive order. The order covers federal agencies only, so this is a choice, not a requirement. The likely logic is public-sector access: Salesforce is building out its Missionforce government business and wants to match federal procurement language. The cost is credibility with enterprise buyers, who already struggle to map Salesforce’s fast-changing product names to what they are paying for. Renaming a product does not change what it does.

Sierra (Bret Taylor’s company) and Meta announced the Personal Agent Protocol on October 6, an OAuth-based standard for how a consumer’s personal AI agent signs in, identifies itself and completes tasks with a business. Launch partners include Shopify, Stripe, Walmart, and contact-center vendors Genesys and NiCE. Salesforce is not on the list. This is the scenario bears worry about: the customer arrives as an agent, and the rules for that handshake are written by a Salesforce competitor and the company that just launched an enterprise platform aimed at its customers. Salesforce’s own October 9 post on personal agents shows it sees the issue. It needs a seat at the table before the v0.1 spec ships later this month.

Oppenheimer’s Dreamforce customer survey (Outperform, $275 maintained) found that 85% of respondents have tested or deployed Agentforce, but nearly half are still in tests or pilots. About 40% of customer data sits with a third-party provider rather than Data 360. Wide trials with few full rollouts usually means revenue arrives later and in waves. For Salesforce’s FY27 math, the question is conversion rate and timing, not awareness. That is why the next print’s Agentforce ARR (above $1.5B in Q2) and paid-deal count matter more than any Dreamforce adoption headline.

JPMorgan expects the software rebound to extend through third-quarter reports, arguing that AI is lifting infrastructure software and that disruption fears for application software have gone too far. Its picks are Microsoft, ServiceNow, Snowflake and Cloudflare, with no pure application vendor among them. The tape agreed: IGV rose 3.9% this week to $112.67, Snowflake 8%, Atlassian 10% and ServiceNow 5%, while Salesforce fell 2.4%. Money is coming back to software, but to data and platform names first. Application vendors will have to prove their case in their own October and November prints.

Snowflake rose 7.4% on Friday to $368.89, a new 52-week high, after releasing Agentic Marketplace Discovery in public preview, which lets customers find and compare AI and data listings inside Snowsight. The feature is modest. The move reflects the market’s view that the data layer captures agent spending first: product revenue grew 37% to $1.49B last quarter. For Salesforce, this is the competitive framing to watch. If agents are grounded in Snowflake or Databricks rather than Data 360, the CRM becomes one source among many. Oppenheimer’s 40% third-party data figure makes the same point.

HubSpot is eliminating about 660 roles (7% of its workforce) and replacing its Hub-based product teams with teams that own customer outcomes end to end, with fewer management layers. It expects $65M–$75M in charges, mostly in Q4, and left Q3 and full-year guidance unchanged. CEO Yamini Rangan says the cuts are not driven by AI efficiencies. JPMorgan downgraded the stock to Neutral on October 8, but HUBS still finished the week up 6%. The structural signal matters more than the headcount: a CRM vendor is dropping its product-line structure because “customers don’t think in Hubs.” Salesforce’s cloud-by-cloud organization faces the same question.

OpenAI and Atlassian expanded their partnership on October 6: GPT-6-family models will power agents across Atlassian’s platform and Rovo, grounded in its Teamwork Graph. The same day, watchTowr published a proof of concept for CVE-2026-21589, a critical (9.3) pre-auth file-read flaw across eight Data Center products, and exploitation attempts began within two hours. Cloud is already patched. TEAM still rose 10% for the week. The lesson for enterprise buyers is that self-hosted software now carries a patch window measured in hours, which strengthens every vendor’s push to move customers to cloud.

The minutes of the September FOMC showed the 25bp hike to 3.75%–4.00% was unanimous and that most officials see another increase by year-end as likely. Futures still price only about a 20% chance of an October 28 hike, after weak September payrolls, with December the consensus timing. September CPI arrives Wednesday, October 14. Forecasts call for headline inflation to rise to roughly 3.6%–3.7% on record gasoline prices, with core near 2.4%. A core print of 0.3% or more month on month would put October back in play. For software multiples, that one number matters more than any company headline next week.