Total business applications in August 2026 were 531,728, according to the Census Bureau's latest Business Formation Statistics release. Two months earlier, in June, the number was 531,423. In between, applications jumped 8.1% in one month and fell 7.8% in the next, almost entirely on the back of one industry's whiplash. The number that moved in the same direction for two straight months, with no spike and no reversal, was one nobody wrote a headline about: applications that plan to pay someone.
What Business Formation Statistics track, release to release
Business Formation Statistics (BFS) is the Census Bureau's monthly count of applications for an Employer Identification Number, filed on IRS Form SS-4, built in research collaboration with the Federal Reserve Board, the Atlanta Fed, the University of Maryland and the University of Notre Dame. The headline Business Applications (BA) series counts nearly everything filed. Underneath it sit two narrower series worth knowing here: High-Propensity Business Applications (HBA), restricted to filings whose characteristics (a corporate structure, a stated hiring plan, a NAICS code with a historically high employer-conversion rate) predict the filer will actually become an employer, and Business Applications With Planned Wages (WBA), a stricter cut limited to filers who told the IRS they expect to pay wages within four quarters. WBA is the smallest and slowest-moving of the four series the Bureau publishes, which turns out to be exactly why it matters below. We covered the July release's Retail Trade spike here, and this piece picks up where that one left off: the next release it flagged as due September 11, 2026 has now arrived.
Three releases, one round trip
| Month (release date) | Total applications | Change vs. prior month |
|---|---|---|
| June 2026 (Jul 9, 2026) | 531,423 | +1.1% |
| July 2026 (Aug 12, 2026) | 578,926 | +8.1% |
| August 2026 (Sep 11, 2026) | 531,728 | -7.8% |
Run the arithmetic straight through and August's total sits just 305 applications above June's, a difference of 0.06%. A reader who checked in only twice, once in June and once now, would conclude nothing happened. A reader who checked every month lived through an 8.1% jump and a 7.8% drop that mostly cancelled each other out. Both readings are true of the same data, which is the actual lesson: a single month's change in this series tells you less than it looks like it does. The Census Bureau says as much in its own release notes, flagging every one of these monthly percentages with a symbol meaning statistical significance "is not applicable or not measurable."
Almost all of that round trip traces to one industry. Retail Trade (NAICS 44-45) applications went from 97,096 in June to 141,241 in July, a 44.5% jump, the largest move of any sector Census tracks, then fell to 104,087 in August, down 26.0% in a single month. Retail didn't fully give back July's gain: 104,087 is still 7.2% above where it stood in June. But it gave back most of it, and the industry breakdown is why the national total swung the way it did. No other sector moved anywhere near that far in either direction.
The number that didn't bounce
Applications With Planned Wages didn't spike in July and didn't fall off a cliff in August. They fell in both months, on their own, with nothing to bounce back from: 35,695 in June, 35,024 in July (down 1.6%), 34,263 in August (down another 1.9%). Over the same two months that left the headline total essentially flat, the applications that told the IRS they plan to pay someone dropped 4.0%. That's a small move and a real one, because it pointed the same direction twice in a row, which the total, Retail Trade and even High-Propensity Business Applications all failed to do.
High-Propensity Business Applications add a messier data point rather than a clean second confirmation: 149,714 in June, up to 151,857 in July, then down to 145,387 in August, a net decline of 2.9% across the two months even though the direction flipped in between. Worth saying plainly: not every series lined up as neatly as Planned Wages did, and pretending otherwise would be the same mistake as reading too much into one loud month.
Why the quiet number is the more honest one
Software's own corner of this data behaved the same steady way. The Information sector, Census's broadest category for software, media and telecom, ticked from 12,999 in June to 13,223 in July (+0.5%) to 13,180 in August (+0.3%), a 1.4% rise across two months that never showed up in a headline because it never moved fast enough to be one. Software's share of new filings hasn't moved much all year, which the July release already showed. What this round adds is narrower: boring and steady turned out to be the more reliable read of where things actually stand than volatile and dramatic. If a headline about business-formation numbers surging or crashing crossed your feed this summer, there's a good chance it was reporting on retail, not on the market a software build competes in.
What this means for a funded founder deciding whether now is the time
A non-technical founder scoping a funded build this fall doesn't need a national business-formation index to decide whether to start. The risk with a series like this one isn't that it's wrong, it's that it's easy to reach for as evidence for a decision it was never built to inform: "the market's heating up, we should move fast" after a headline about a monthly jump, or "things are cooling off, maybe we wait" after the next month's reversal. Both readings, this time, would have been reacting to a retail swing that had nothing to do with software. The steadier numbers, applications planning to pay someone and the Information sector's own count, moved a few percent either way over two months and would never make a headline on their own. That's usually a sign a number is telling you something real rather than something loud.
Where this actually would matter: if the product you're funded to build serves new retail formations directly, point-of-sale software or inventory tools sold into new store openings, that Retail Trade swing is closer to your own market signal, and it is worth tracking month to month rather than dismissing as noise. For most funded, non-technical founders in this dataset, whose businesses have nothing to do with retail formation rates, the honest read is that neither the July spike nor the August pullback tells you anything about your own timeline. What does is the scope you're actually funded for and how long it takes to reach a working version, which is a conversation about your build, not about a national index.
A worked example
Picture a two-person, seed-funded home-services marketplace, funded in the spring to build a scheduling and dispatch platform connecting independent contractors to homeowners. Neither founder has a technical background. One of them reads the July headline about business applications jumping 44.5% and takes it as a sign competitors are about to flood in. The other reads the August headline about a 7.8% pullback and relaxes. Neither reaction is really about their business: the jump was retail, and the pullback was mostly retail unwinding. A home-services marketplace isn't measured by either series. What would actually tell this founder something is a number closer to home, how many home-services businesses with planned wages are filing in their own state and county, not a national retail-driven swing. Chasing the wrong index either way costs the same thing: attention spent reacting to a number that was never about them, instead of on the scope and timeline of the one build they're actually funded to ship.
If a headline about business-formation trends made your own build feel more or less urgent this summer, the more useful question is usually about your own scope: what actually goes into scoping a fixed-price build, and whether the same funded, pre-revenue population the SEC's Form D filings describe matches where your company sits today. We scope every custom software build against your own funded plan and timeline, not against a national index, with transparent pricing and an itemised estimate within 48 hours of a call.
Sources
- U.S. Census Bureau: Business Formation Statistics, August 2026 (Release CB26-148, published September 11, 2026; accessed September 17, 2026)
- U.S. Census Bureau: Business Formation Statistics Monthly Data Release, August 2026 (published September 11, 2026)
- U.S. Census Bureau: Business Formation Statistics, July 2026 (Release CB26-130, published August 12, 2026)
- U.S. Census Bureau: Business Formation Statistics, June 2026 (Release CB26-115, published July 9, 2026)