534,947 New Business Applications Were Filed in September. Only 29,022 Are Projected to Become Employers.
Census data shows a steady stream of new startups, but the number likely to put anyone on payroll is a small slice. Here is how to read the gap.
In this article 9 sections
The Census Bureau counted 534,947 business applications in September 2026, up 0.5% from August. It projects that only 29,022 of that month's applicants will become employer businesses with payroll within four quarters. The gap shows that most new tax IDs belong to solo ventures, side projects and ideas that never hire anyone.
Upstart Wire covers entrepreneurship and strategy with an eye on the data behind the stories. This feature reads the Census Bureau's latest Business Formation Statistics for founders, investors and anyone sizing up a market. It covers what the numbers measure, what the regional split shows, and how to use the gap between applications and employers.
What Did the Census Bureau Report for September 2026?
The Census Bureau's Business Formation Statistics release of October 14, 2026 reported 534,947 seasonally adjusted business applications for September. That is a 0.5% increase from August. The Bureau notes that statistical significance is not applicable or not measurable for these month-to-month changes, so treat small moves as noise.
Three subseries sharpen the picture. High-propensity applications, which the Census Bureau ties to the likelihood of becoming a real business, came in at 147,331, up 1.3%. Applications with planned wages reached 35,786, up 4.3%. Applications from corporations were 40,951, down 0.2%.
The forward-looking number is the one that matters most for hiring. The Bureau projects 29,022 business formations within four quarters, up 1.4%, and 39,703 within eight quarters, up 1.0%. A formation here means a business that shows its first payroll tax liability.
How Many New Business Applications Actually Become Employers?
A small share do. By our math, the projected 29,022 employer startups equal about 5.4% of September's 534,947 applications. Measured against the 147,331 high-propensity applications, the figure rises to about 20%. These ratios are our own calculation from the Census Bureau's published numbers.
The reason is simple. An Employer Identification Number is easy to get, and many people file one for reasons that have nothing to do with hiring. A freelancer, a landlord forming an LLC, a creator opening a bank account and an online reseller all need a tax ID.
The Census Bureau describes the series as applications for tax IDs through filings of IRS Form SS-4, while formations are the first instance of payroll tax liabilities for those applications. Its Business Formation Statistics overview notes that applications tied to internet sales are excluded from the high-propensity series as of the January 2026 release. That change tightens the definition of a serious startup.
The takeaway for founders is practical. A headline about record startup activity usually counts applications. The count of businesses that hire is much smaller and moves more slowly.
What Does the Regional Split Show?
The South dominates, and the West is cooling. The table below lists September applications and the monthly change by region, from the Census Bureau's release.
| Region | September 2026 applications | Change from August |
|---|---|---|
| South | 235,028 | +1.2% |
| West | 130,070 | -1.8% |
| Midwest | 94,891 | +2.6% |
| Northeast | 74,958 | +0.3% |
The South accounts for roughly 44% of the national total, by our calculation. The West makes up about 24%, the Midwest about 18% and the Northeast about 14%.
Projected employer formations tell a slightly different story. The Northeast is projected up 3.0% and the West up 2.4%, while the South is down 0.2%. Applications and payroll-bound startups do not always move together, which is a reason to watch both series.
One caution applies. The Bureau warns that month-to-month changes are often not statistically significant. A single month's regional swing is a prompt to look closer, not a trend.
Should Founders Treat the Gap as a Warning or an Opportunity?
It is both. The gap reveals how hard it is to turn an idea into an operating employer. It also reveals how few of your competitors will survive long enough to matter.
On the warning side, many applicants stop at the filing stage. A tax ID does not create revenue, customers or a product. Founders who treat registration as progress often discover that the hard work has not started.
On the opportunity side, thin competition at the employer level means real businesses can stand out. If you reach payroll, you are in a smaller group than the headline count suggests. That is a useful frame when sizing a local market or pitching a lender.
Our read at Upstart Wire is that the planned wages figure is the best early signal. Applicants who say they plan to pay wages have committed to a more demanding path. At 35,786, that group is under 7% of all applications.
How to Use This Data in Your Own Planning
Start with your market. If you sell to small businesses, count the employers, not the applicants. A directory of new tax IDs overstates your addressable market by a factor of ten or more.
Next, use the data to calibrate your own milestones. Moving from idea to first customer is the step many applicants never complete. Our 90-day roadmap for first-time founders lays out a sequence for getting there.
Finally, watch competitors the way professionals do. Public filings and signals reveal who is actually operating. Our guide to reading competitors' public filings shows what to look for beyond a registration record.
Investors and lenders can use the same lens. A credit officer who sees a new tax ID knows little. A credit officer who sees payroll filings, a lease and repeat customers knows a lot. Founders who can show those signs early, rather than a registration certificate, tend to get a faster and warmer hearing.
Money decisions come next once you plan to hire. Pay is a hard question in year one, which we covered in how much to pay yourself as a founder. And your legal structure affects both taxes and hiring, as shown in our comparison of LLC vs S-Corp.
Why Is the Monthly Number So Bumpy?
Monthly application counts swing more than most people expect. The Census Bureau reported 578,926 applications for July 2026, an increase of 8.1% over June. September's 534,947 sits well below that peak, which means a sharp pullback happened in between.
Several things can cause swings like that. A policy change or a deadline can pull filings forward. A viral side-hustle trend can add thousands of solo filers in a week. Seasonal adjustment smooths some of it, but not all.
The lesson is to look at several months together. A three-month average tells you more about the direction of entrepreneurship than any single release. It also keeps you from building a pitch deck around a spike that fades.
The employer-bound series is steadier. Projected formations moved only 1.4% in the latest month. If you care about real hiring businesses, that line is the more reliable guide.
What to Check Before You File for a Tax ID
If you are one of those applicants, a short checklist keeps the filing from becoming the finish line. Confirm these points before and after you apply:
- Decide whether you need an EIN now, or only after you hire or open a business bank account.
- Pick a legal structure with a tax adviser, because it changes what you owe.
- Write down the first three customers you expect and how you will reach them.
- Set a date by which you will either hire, hire a contractor or revisit the plan.
None of these steps is complicated. They are designed to turn a filing into a decision. The data suggests that many applicants skip the decision, which is why so few reach payroll.
What the Data Cannot Tell You
These figures describe filings, not outcomes. The Census Bureau's projection estimates how many applicants will show payroll, but it does not say how many will survive, grow or profit. Actual formations are only available with a long lag, because the Bureau confirms them from administrative payroll data.
The next annual update arrives with the November 12 release. During it, the Bureau revises the formation series and recomputes seasonal factors, which can shift the numbers you see today. Treat any figure here as provisional until then.
The data also excludes much of what happens informally. Many people earn money without filing for a tax ID at all. Applications are a useful barometer of formal entrepreneurship, not a full census of work.
Conclusion
September's numbers show steady startup activity and a much smaller group heading toward payroll. Reading the gap correctly protects you from inflated market sizes and from mistaking filings for progress. The most useful signals are high-propensity applications, planned wages and projected formations.
Before your next planning meeting, replace any "new businesses" figure with an employer count. Upstart Wire will track the November 12 release and the annual revisions that come with it.
