Salesforce's $5.66B Braze Deal Anchors SaaStr Argument That $1B ARR Is the New Safe Scale for Public B2B Software

There is no model release in this story. The AI products that come up are:

• Rovo, Atlassian's chat and agent layer across more than 20 apps. Its features are exposed to outside agents through MCP, an open protocol that lets AI agents call tools in other software (video 11:22).
• Fin, the customer-service AI agent Salesforce is buying.
• Agentforce, Salesforce's AI agent platform (TechRepublic).

No version numbers are disclosed.

• How the essay's data was built: ten public companies with $400M–$1B in revenue, one quarter each, compared against larger companies. It is a picked sample, not a full screen of the market, so the NRR–growth pattern it shows is a correlation in that group only (SaaStr).
• Pricing on usage: JFrog, which charges by consumption, grew cloud revenue 53% and has 121% NRR. Its CEO says AI-written code creates more software artifacts for JFrog to store and secure (Barchart/JFrog).
• Atlassian's product lessons from the linked talk:
• Watching what people type into chat shows which features to build (06:51).
• Every feature built for humans should also be exposed as a tool agents can call (13:36).
• Atlassian's working rule is that anything a human can do in the product, an agent should be able to do too (19:32).

• Braze deal terms come from Braze's own announcement and are also in an SEC filing (SEC). The 58% premium is reported by Yahoo Finance. Braze's full-year revenue guidance of $910–913M (Finviz) is consistent with SaaStr's figure of about 6.2x revenue.
• Company metrics (ARR, NRR, free cash flow) are self-reported in each company's earnings release.
• Market-wide figures in the essay (median growth of 11.8%, the Rule of 40 count, valuation multiples by growth group) cite Aventis and SEG. They were not independently checked here.

Salesforce is paying for both deals from its own cash, which supports Lemkin's point that scale funds acquisitions. Fin was a private company, so it is not an example of a sub-$1B public company being bought. HubSpot, which is above the $1B line, is also slowing down. That cuts against reading the $1B line as protection from slowdown; Lemkin's argument is that it funds the response, through buybacks and cash to rebuild the product (TIKR).

• HubSpot customer additions: SaaStr says 7,000 net new customers against an expected 9,000–10,000. TIKR reports that management cut the outlook to about 5,000–6,000 (TIKR).
• Asana NRR: SaaStr's figure of 97% overall matches the "core" customer number from the earlier Q4 FY26 report, where overall NRR was 96% (MarketBeat).
• Sprinklr cash flow: Q2 free cash flow was $13.1M (CMSWire). SaaStr's $79M first-half figure was not confirmed.
• Atlassian staffing pilots: in about 10 projects staffed with "AI builders" (people whose main job is shipping code with AI tools), work was fast at first and then slowed down badly (22:12). This is an internal account from a company employee.

• Fix retention before chasing new customers. In the sample, every company growing faster than 20% had NRR of 105% or more (SaaStr).
• Move away from per-seat pricing if customers' headcounts are flat or shrinking.
• Being profitable without growth doesn't protect the stock price. PagerDuty and Sprinklr both make money and trade at low valuations.
• For product teams: ship chat inside the product to learn what users want, then add AI to existing workflows instead of rebuilding from scratch (16:33).
• Hiring: Atlassian says it now hires more juniors and fewer mid-level staff, pairing juniors who try more AI tools with seniors who are good at catching low-quality AI output (25:28).