Invest1 publisher3 min readPublished
ServiceTitan keeps 1,000 shared accounts after delisting a $100 million partner
ServiceTitan switched off Podium's integration after nine years and, SaaStr says, kept essentially every shared account. Inside Salesforce's quarter, the data segment added $730 million and the record segment about $450 million.
The Investor · Invest desk
What happened
- ServiceTitan switched off Podium's integration across roughly 1,000 shared customers, ending a nine-year partnership.
- Podium had built $100 million of AI agent ARR in under 24 months, much of it sold inside ServiceTitan's own customer base.
- SaaStr reports ServiceTitan kept essentially all of the delisted accounts, because contractors' jobs, invoices, customer history and technician schedules live in ServiceTitan.
- ServiceTitan's April 2026 API terms confine calls to a predefined certified scope and bar AI systems from choosing endpoints on their own.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- constraint With agents barred from picking endpoints, a vendor building on ServiceTitan either ships inside the certified scope or sells around the platform, and the integration route now runs through ServiceTitan's approval.
- contradiction SaaStr argues the record is the most expensive database a company owns, then shows 35 of Salesforce's 52 trillion ingested records never moved in at all, so the largest system of record is growing fastest on data it does not hold.
- cost At $3,000 per gigabyte per year list, agent-scale data stored inside Salesforce costs orders of magnitude more than the same gigabyte on S3, and the customer either pays that or routes the data around the platform.
- decision Anyone pricing a vertical SaaS name has to value renewal separately from expansion, since on SaaStr's reading the delisting protects future revenue without booking any.
Salesforce's own quarter is where the split shows up in dollars. Data 360, Headless Platform and Other went from $2.95 billion to $3.68 billion [9], an increase of $730 million, or 24.7% [1]. Agentforce Apps, the record layer, reported $6.91 billion on 7% constant-currency growth [8], which implies a prior-year base near $6.46 billion and roughly $450 million of added revenue [2]. The smaller segment added about 1.6 times the dollars of the larger one [3]. Total revenue was $11.13 billion, up 13%, and SaaStr puts organic growth closer to 8-9% once Informatica is excluded [10]. The 7% is constant currency and the $6.91 billion is as reported, so treat the $450 million as approximate.
Podium's side is simpler. A hundred million dollars of agent ARR across 1,000 shared accounts is $100,000 per account per year [4], and that is a ceiling, because SaaStr says much of the $100 million sat inside ServiceTitan's base, not all of it [3]. ServiceTitan collected none of it when the integration went off. In SaaStr's reading, blocking a partner prevents the loss of future dollars and adds none, and the leverage behind that shows up as high renewal rates, pricing power and net revenue retention above 110% [19][20].
The post compares growth rates without revenue bases or multiples. Databricks past $7 billion of ARR growing over 80% [6] and Snowflake up 34% on product revenue with 126% net revenue retention [5] sit beside ServiceTitan's 25% [4] as though the bases were comparable. The retention claim is also contemporaneous with the cut-off, so the 1,000 contractors have not yet renewed with Podium gone [2].
The pricing argument and the Zero Copy numbers pull in opposite directions. Salesforce lists additional storage at $125 per month per 500MB, or $3,000 per gigabyte per year, against a fraction of a cent per month for the same gigabyte on S3 [11]. At agent volumes the data goes elsewhere: of the 52 trillion records Data 360 ingested in the quarter [13], 35 trillion arrived through Zero Copy and never physically moved into Salesforce, a motion up 277% year over year [14], which is about 67% of the total [5]. Salesforce processed 28.6 trillion tokens for customers in the same period, up 152% quarter over quarter [12]. It allocates API calls per user per day, and agents do not have users [15].
I would put the distinction more narrowly than retention against growth. ServiceTitan grew 25% with the record, the April 2026 API scope and the delisting all in place [4][16], and Veeva grew 16% last quarter while running a full CRM migration underneath its customers [18]. Locked-in accounts can be sold something new. The counter-case is that the perimeter taxes the specialists who were doing the agent selling, and Podium built its $100 million in under 24 months largely through the integration ServiceTitan has now closed [3]. SaaStr calls 25% roughly the best case for a vertical system of record in 2026 [4]; if the perimeter does what ServiceTitan's marketplace policy says it is for [17], the next growth number should be above that.
What to watch
- Whether the 1,000 delisted accounts renew with ServiceTitan once a full contract cycle has passed with Podium switched off.
- Any ServiceTitan agent revenue booked out of those accounts, which would show the perimeter converting into expansion.
- Zero Copy's share of Data 360 ingestion in Salesforce's next quarter, against the roughly 67% it ran in this one.