Invest1 publisher3 min readPublished Updated
August re-rated software faster than any growth rate could move
A single month of trading undid eight months of AI-disruption pricing in B2B software. That is a multiple moving, not a business changing. The winners were sorted by whose meter the agent workloads run through.
The Investor · Invest desk
What happened
- By SaaStr's count, roughly $2 trillion of market cap left B2B software eight months ago on the view that AI would eat the category, and most of it has since come back.
- The iShares Expanded Tech-Software ETF closed September 3 at $106.81, about 40% above its April low near $74 and 16% higher over August, while State Street's XSW touched an all-time high in late August.
- Inside that index recovery sits a 117-point spread, with CrowdStrike up 83% on the year and monday.com down 36%, both of them selling software to the same B2B buyers.
- Twilio and monday.com each grew 22% in their June quarters and still finished 109 points apart on the tape.
- Airtable, valued at nearly $12 billion at its 2021 peak, sold to Bending Spoons for under $1.3 billion.
Compiled by The InvestorSomething wrong?How this is made
Why it matters
- exposure The meter that collects AI upside also concentrates it, so the names that re-rated on agent workloads now carry single-account risk a beat-and-raise cannot offset: Datadog fell 19% on August 6 after disclosing that its largest customer was trimming usage.
- constraint Part of what got re-rated at Twilio is revenue booked at cost, since $71 million of pass-through carrier fees is 4.7% of a $1.50 billion quarter and accounts for most of the wedge between 22% reported and 17% organic growth, so the multiple is being paid on gross flow.
- contradiction Jefferies told clients that displacement fears are overblown, while the private market cleared Airtable at under 11% of its 2021 peak, so the public tape and the private marks disagree on whether AI disruption was repriced or merely deferred.
- decision Holding software as a sector now expresses no view on AI, because the index round trip averaged constituents that moved in opposite directions, which pushes the work down to underwriting individual billing models.
Atlassian's whole year happened in one month. It gained 88% across August and closed September 3 at $194.68, up 20.1% on the year [25], and dividing 1.201 by 1.88 puts it roughly 36% down through the end of July [27]. The quarter that set it off grew revenue 28% with cloud up 31% and RPO up 44% [26], solid numbers that didn't shift 88 points in four weeks on their own. What re-rated was the multiple.
Salesforce ran the same trade at lower amplitude: down 29.5% year to date in early April, a 52-week low of $146 on June 22, then about 40% over August after the August 26 print, closing at $264.43 against a $253.62 start [22]. Eight months of violent tape for $10.81 a share [24]. Underneath it, current remaining performance obligation grew 14% in constant currency while organic revenue grew 6.4% once Informatica's $456 million comes out [23], which is a growth rate that justified neither the April price nor the September one.
SaaStr's read of the year, whose price tally runs to the September 2 and 3 closes, is that two variables did most of the sorting, whether AI workloads land on your meter and what multiple you carried into January [8]. The real question is whether an agent workload arrives on your invoice without anyone buying a seat, because an enterprise deploying agents generates more logs to monitor, more queries to run and more traffic to route whether or not it hires [20].
CrowdStrike is the cleanest test of that and also the murkiest disclosure. ARR of $5.84 billion grew 25% while Falcon Flex ARR passed $2.29 billion, up 101% [11][12]; back the growth rates out and the prior-year book was about $4.67 billion of which roughly $1.14 billion was Flex, so the non-Flex remainder went from about $3.53 billion to $3.55 billion, up half a percent [13]. Either the rest of the portfolio stopped growing or customers are migrating existing spend into Flex commitments, and the disclosure does not distinguish. George Kurtz credits the momentum to the world's AI adoption expanding the attack surface [21], which is the demand-side account; the accounting account is that one consumption vehicle now carries the entire increment.
Two things would break the view that the meter rather than the growth rate set these multiples. If the consumption names decelerate into the January quarters while their multiples hold, August was momentum, which is close to what Jefferies actually told clients over the Labor Day weekend when it said displacement fears were overblown [4], and momentum is a position rather than a valuation. If HubSpot or monday.com, both down more than 33% [6], re-rate upward on seat growth alone, then the sorting variable was the January entry multiple and AI was the cover story for a mean reversion. Against that, Snowflake's product revenue of $1.49 billion grew 37% in a third consecutive quarter of acceleration with the guide lifted to $6.07 billion [17], which is the one number in the group that re-forecast a business instead of re-pricing one.
What to watch
- January-quarter prints from Snowflake and Datadog, which show whether consumption growth holds or whether August was a momentum trade.
- Whether Salesforce's organic growth converges toward its 14% constant-currency cRPO once Informatica laps out of the comparison.
- Whether CrowdStrike separates Falcon Flex from non-Flex ARR growth, which would settle whether the remainder is flat or simply migrating.