Regional capital coverage tends to report one number: total dollars raised. It is the least useful number available, and this piece is an argument for why — made with our own data, against our own headline figure.
Our deal record currently holds 274 transactions across Minnesota, Wisconsin, North Dakota, and South Dakota, almost all sourced from SEC Form D filings. Aggregate disclosed capital: $4,328,598,423.
That number is real and almost useless. Here is what is underneath it.
Nearly half the money never reached an operating company
Form D is filed by anyone selling securities under a Regulation D exemption. That includes venture-backed startups. It also includes real estate partnerships raising for a single building, bank holding companies recapitalizing, private funds closing, and public companies doing PIPEs.
Sorting our record into pooled vehicles, banks, and real estate on one side and operating companies on the other:
| Deals | Disclosed capital | Median | |
|---|---|---|---|
| Pooled vehicles, banks, real estate | 90 | $2,203,909,263 | $3,835,575 |
| Operating companies | 184 | $2,124,689,160 | $2,530,685 |
Fifty-one percent of the capital went to something that is not an operating business. Any story that reports $4.33 billion as a measure of regional company formation is off by roughly a factor of two before it starts.
North Dakota’s record is one 2018 real estate deal
This is the finding that most changes the picture.
North Dakota shows $711,605,237 across 14 filings — which, per filing, makes it look comparable to Wisconsin. It is not.
$673,177,030 of it is a single February 2018 filing by Edgewood Properties, LLLP. That one entry is 95 percent of the state’s recorded capital.
Remove it:
| North Dakota | Deals | Disclosed capital | Median |
|---|---|---|---|
| As recorded | 14 | $711,605,237 | $2,224,731 |
| Excluding Edgewood Properties | 13 | $38,428,207 | $1,679,464 |
| Operating companies only | 8 | $13,454,462 | $1,332,232 |
Eight operating-company filings. Thirteen and a half million dollars. That is North Dakota’s visible private capital record in our data, and it is a more honest description of the state than any per-filing average.
South Dakota is smaller still: 6 operating-company filings, $7,798,218, median $555,000. For context, the Dakota Seed Fund closed on $500,000 in March 2026 to write $25,000–$75,000 checks — which, against a state median raise of $555,000, is a fund sized correctly for its market rather than a rounding error.
Operating companies, by state
| State | Deals | Disclosed capital | Median | Mean |
|---|---|---|---|---|
| Minnesota | 120 | $1,892,652,069 | $3,530,800 | $15,772,100 |
| Wisconsin | 50 | $210,784,411 | $1,577,250 | $4,215,688 |
| North Dakota | 8 | $13,454,462 | $1,332,232 | $1,681,807 |
| South Dakota | 6 | $7,798,218 | $555,000 | $1,299,703 |
The Minnesota–Wisconsin gap is the one worth sitting with. Minnesota records 2.4 times Wisconsin’s deal count but nine times its operating-company capital, and its median raise is more than double ($3.53M against $1.58M).
Two states with comparable populations and research universities are not running the same market. Wisconsin is doing a similar volume of meaningfully smaller deals. Whether that reflects sector mix, investor base, or company stage is not something Form D can answer — but the asymmetry is large, consistent, and worth knowing before anyone tells you the Upper Midwest is a single ecosystem.
Most raises here are small
Distribution of operating-company raises:
| Raise size | Share of filings |
|---|---|
| Under $500,000 | 13% |
| Under $1,000,000 | 27% |
| Under $2,000,000 | 44% |
| Under $5,000,000 | 65% |
| Under $10,000,000 | 78% |
Forty-four percent of operating-company raises in this region are under $2 million. Fewer than a quarter clear $10 million.
The mean across all 274 filings is $15.8 million. The median is $2.8 million. When a distribution’s mean is more than five times its median, the mean is describing a handful of outliers, not the market. Nearly four fifths of what happens here is a raise small enough that most institutional funds would not return the call.
The most active issuer files constantly
Vireo Growth Inc. filed seven times between July 2025 and June 2026, totaling $266,893,990 — more than the entire operating-company total for Wisconsin, North Dakota, and South Dakota combined.
Other repeat filers: Centaur Health Holdings (3 filings, $89.2M), Gravie (2, $200.0M), HBS Research Farms (2, $26.9M), Flywheel Exchange (2, $18.6M).
Repeat filings are easy to misread. A company filing three times has not necessarily raised three rounds — Form D is amended, offerings are extended, and a single raise can generate multiple filings across tranches. Counting filings as rounds inflates activity. We record each filing separately because that is what exists, and we say so rather than merging them on an assumption.
What Form D will not tell you
Two of our 274 entries carry a round label. Two.
That is not an oversight. Form D does not report the round. It reports the issuer, the amount sold, the exemption claimed, and the number of investors. It does not say “Series A,” it does not name who invested, and it does not distinguish a priced equity round from a convertible note or a fund closing.
Our record uses exempt-offering for 235 entries for exactly this reason.
Calling those “venture rounds” would be an inference dressed as a fact. The 36
debt-financing entries are labeled from the filing’s own indication; the two
venture-round entries — Claros Technologies
and Quercus Biosolutions — carry round labels
because the companies announced them, not because a filing said so.
This is the practical limitation of Form D-based coverage, and the reason the useful work is pairing filings against announcements. Gravie’s June 2026 filing reports $50 million from exactly one investor; its July announcement names General Atlantic as lead and declines to state an amount. Neither document references the other, and we record the linkage as an inference rather than folding it into the data — but the filing supplies a figure the company chose not to publish, and that is the whole value of reading both.
Where the money concentrates
Filing counts by city: Minneapolis 57, Milwaukee 22, Madison 17, St. Paul 9, Minnetonka 8, then Middleton, Fargo, Edina, and Plymouth at 6 each.
Minneapolis alone accounts for more filings than Milwaukee and Madison combined. The suburban Twin Cities entries — Minnetonka, Edina, Plymouth, Wayzata, Golden Valley, together 30 filings — are a real feature of this market and one that city-level coverage elsewhere tends to miss entirely.
Method, and what would make these numbers wrong
Stated plainly, because a reader should be able to discount this appropriately:
The time series is not usable. Our record holds 162 filings dated 2025 and 73 dated 2026, against 39 for every year before 2025 combined. That is a collection artifact — the corpus was built from a recent sweep — not evidence that activity multiplied. Do not read year-over-year growth into it. We are not publishing a trend line, because we do not have one.
Recent months are incomplete. July 2026 shows 2 filings. Form D is due within 15 days of first sale and filings surface on EDGAR with a lag, so the most recent weeks always undercount.
The operating-company split is our judgment, not a filing field. We classified by the issuer’s self-reported industry and by entity-name conventions — LLLP, “Properties,” “Bancshares,” “Fund.” It is a defensible heuristic and it is still a heuristic. Reasonable people would sort some entries differently.
Two entries distort Minnesota badly. SharpLink Gaming ($425.0M, May 2025) is a public company, and Mill City Ventures III ($450.0M, July 2025) is a specialty finance company; both clear the operating/vehicle line awkwardly. Excluding SharpLink alone drops Minnesota’s operating-company total from $1.89 billion to roughly $1.47 billion. Medians are stable against this; totals are not, which is why every table above reports both.
Amounts are as filed. Form D reports the amount sold as of the filing date, not the size of the round. A company that files early in a raise understates it permanently unless it amends.
None of this makes the record useless. It makes it a record of filings rather than a record of rounds — and read that way, with the vehicles separated out and the outliers named, it is the most complete picture of private capital in these four states that we are aware of anyone publishing.
Figures computed from the Northern Dealbook deal record as of August 7, 2026. Every underlying entry links to its source filing.