Which product actually funds a manufacturer?
The menu is short once you stop treating every acronym as interchangeable. Conventional cash-flow paper underwrites adjusted earnings and a debt-service coverage ratio. SBA 7(a) for manufacturing acquisitions is the guaranteed change-of-ownership product inside a published $5 million box. SBA 504 for plants and heavy equipment buys owner-occupied dirt and qualifying iron through a CDC. Asset-based lending lives on a borrowing base. Equipment finance follows serial numbers and useful life. Mezzanine and unitranche stretch leverage when senior is already maxed. Seller notes and earn-outs fill gaps; they are not senior. An ESOP financing stack is a trustee-and-leverage conversation, not a 7(a) slogan.
Census CBP 2023 counted 284,452 U.S. manufacturing establishments, and 68.4% of them employ fewer than 20 people (Source: Census CBP, 2023). Most of those files never need mezzanine. Many of them do need a honest answer about whether SDE after a replacement wage can carry a 7(a), or whether the floor is worth more on orderly liquidation value than on a thin earnings multiple. Buying a manufacturing business still starts with a buybox. Financing does not invent a buybox.
Jason Taken will not tell a first-time operator that every machine shop "does SBA." HedgeStone Business Advisors will put the product next to the artifacts the committee will request: three years of financials, classed inventory, a machine list with serials, and concentration on a trailing-12 and trailing-36 basis. See how lenders underwrite manufacturers.
When should a buyer stop forcing 7(a) onto a larger file?
Stop when enterprise value plus fees plus the working-capital need exceed what a $5 million 7(a) can carry, even after you peel the plant onto a 504. Combined SBA-backed financing can reach $10 million under the Administrator's 2026-05-18 announcement and the 2026-07-04 policy notice that outstanding 7(a) does not reduce 504 capacity (Source: SBA / NAGGL). That is still two products. Occupancy, affiliation, and a 10% injection still apply. It is not a $10 million cash-out on a thin equity check.
A 504 stack that CDC guides still describe as roughly 50% bank first lien / 40% CDC / 10% equity only works if the buyer will occupy at least 51% of an existing building or 60% of new construction (Source: SOP 50 10 8). A leased plant with a landlord who will not sell does not become a 504 story because someone liked the rate rumor.
Above that box, rebuild with investor equity, conventional senior or ABL, equipment paper, and a seller note that is not pretending to be standby equity. Search-fund and sponsor files often live here. Stanford's 2024–25 median purchase of $16 million at about 6.2x EBITDA and ~25% margin is that world (Source: Stanford GSB). It is not a 12% EBITDA job-shop quote.
How does a seller test a buyer's capital before exclusivity?
Ask what is committed, what is applied for, and what is a hope. An individual buyer who can show injection source and a lender already reading the book is a different animal from a buyer who "will talk to SBA next month." A searcher still raising acquisition equity is not a cash close. An independent sponsor without a committed fund is a process risk, not a personality flaw.
IBBA / M&A Source Pulse Q1 2026 reported that 83% of deals over $5 million drew three or more offers (Source: IBBA). That statistic describes competition in a band, not a reason to grant 90 days of exclusivity to an unfunded letter. How it works keeps customer names last and still asks the capital question early.
| Product | What it underwrites | Published or labeled limit | Common break |
|---|---|---|---|
| 7(a) | Cash flow + supported valuation | $5 million standard; 85/75 guaranty | Project cost exceeds the box; amortizing note sold as equity |
| 504 | Owner-occupied plant / heavy M&E | $5 million / $5.5 million small-manufacturer project | Occupancy miss; landlord will not sell |
| Cash-flow senior | Adjusted EBITDA and DSCR | Lender-specific; manufacturing DSCR often quoted as 1.20–1.30x market practice | Concentration or thin margin |
| ABL | Eligible AR, inventory, sometimes M&E | Market-practice advances 70–85% AR, 40–60% inventory | WIP, ineligible AR, seasonal collapse |
| Equipment paper | Specific serials | Useful life of the cell | Dirty UCC; missing controls |
| Seller note | Gap and tax timing | SBA equity only on full standby, ≤50% of injection | Current pay that wrecks DSCR |
Sellers who want a clean close should also read seller notes and standby before they volunteer paper that the lender will not count.
What breaks a mixed stack after the LOI?
Three collisions show up on manufacturing files. First, the working-capital peg and an ABL borrowing base get written as if they were the same exhibit. The peg is a purchase-price true-up. The base is a monthly availability test. Confusing them is how a buyer wires a price the revolver cannot support in month two.
Second, an earn-out is treated as if it were amortization coverage. Most credit memos ignore earn-outs as a repayment source. Read earn-outs and how lenders treat them before anyone "makes the multiple" with a hope.
Third, the machine list and the UCC search disagree. Equipment lessors, prior 7(a) liens, and landlord waivers are not footnotes. They are whether the new lender can attach. Equipment financing and leasing is the serial-number sibling of this map.

How does the peg sit inside every financing conversation?
Every product still needs a definition of net working capital that a manufacturer can count. Job-cost WIP, percentage-of-completion, customer-owned tooling, and quality-hold inventory are the usual fights. A 7(a) that fits the dollar cap still dies when the cage is light and the true-up is a personality. An ABL that looked ample on a CIM inventory number still starves when WIP is ineligible.
Write the method in the letter of intent. Trailing average, classed inventory after E&O, physical at close, treatment of deposits and unbilled AR. Then let the credit file and the purchase agreement use the same religion. HedgeStone will not market a file that says "customary NWC" as if manufacturing were a retail store.
Which buyer type arrives with which stack?
Individual operators usually arrive with 7(a) plus a standby note conversation. Strategics often arrive with cash and a tighter non-compete. Searchers arrive with investor equity, bank or ABL, and often a note. ESOP processes arrive with a trustee, a feasibility study, and leveraged paper that is not a marketed CIM with a fun name. Family successors often arrive with below-market structure and estate counsel. None of those habits is a moral ranking. They are how valuation clears.
BizBuySell's 2025 year-in-review put manufacturing at a 223-day median time to close (Source: BizBuySell). Capital that is not real is one reason files sit. Product choice is the other. Pick the product. Then staff the file.
What does a $4.2 million EV fab look like on the menu?
A 28-person general fab with $900k SDE, owner-occupied 40,000 square feet, and 22% customer concentration can often live inside 7(a) if the buyer funds injection and the plant occupancy fits. Peel the building onto 504 only if the CDC path is cleaner than stretching the 7(a) across dirt and goodwill. A 90-person certified supplier at $3 million EBITDA is a different menu: equity, senior or ABL, maybe a note. Do not quote the fab's 7(a) story onto the certified supplier because both "make metal." Size band and sub-sector still pick the row.
Jason Taken is a business broker. He does not issue term sheets, set rates, or decide SBA eligibility. Educational information only. Confirm product, SOP, occupancy, and advance rates with the lender, the CDC, and counsel. Data current as of September 2026.
How should a first data room be labeled so the product conversation stays honest?
Create four folders before anyone debates rate folklore. Folder one is earnings: tax returns, YTD, recast with documents. Folder two is iron: serial-level list, maintenance, UCC. Folder three is working capital: AR/AP aging, classed inventory, WIP method. Folder four is occupancy and dirt: lease or deed, Phase I if the history requires it. Product arguments that start before those folders exist are entertainment.
A buyer who wants 7(a) should already know SOP 50 10 8 governs most 2026 applications and that SOP 50 10 8.1 applies to loan numbers issued on or after 2026-10-01. A buyer who wants ABL should already know ineligible AR and WIP haircuts. A seller who wants a note to "be the down payment" should already have read the standby rule. Label the folders. Then pick the product.
How should a uses-of-funds table decide the product before anyone debates a close date?
Write uses of funds on one page: enterprise value, estimated fees, working-capital need, any required payoffs, and whether the building is in or out. Then assign a product to each use. Goodwill and operating cash usually want cash-flow paper or 7(a). Dirt and qualifying heavy iron may want 504 if occupancy is real. Seasonal inventory may want a revolver or ABL. A gap may want a note with a named job. An earn-out is not a use of funds. It is a hope.
Buyers who skip this table pick a slogan and then discover the project does not fit the $5 million 7(a) box (Source: SBA). Sellers who skip this table grant exclusivity to a letter that cannot be funded. HedgeStone will ask for the table before anyone marks a close on a wall calendar. The 223-day manufacturing median (Source: BizBuySell) is already long. Product shopping after week six makes it longer.
A uses table also stops double-counting. Iron that is inside enterprise value should not also appear as a separate equipment draw unless someone carved it out. Inventory that is inside the peg should not also appear as a second "working-capital loan" without a sentence that explains the difference. How lenders underwrite manufacturers will ask for the same arithmetic.
If two products can fund the same use, pick the one whose eligibility you can actually prove. Occupancy you cannot prove is not a 504 idea. A borrowing base you cannot class is not an ABL idea. A 10% injection you cannot show is not a 7(a) idea. SOP 50 10 8.1 will govern numbers issued on or after 2026-10-01. Build the table to the text that will govern, not to a training slide from an older year.