Tax
IRC 643(f) and QSBS Trust Stacking: The Statute Treasury Just Named
Treasury officials named IRC 643(f) as the tool that could collapse aggressive multi-trust QSBS stacks. Here's what the statute says and what still holds.
By 409.AI Team - 2026-08-20
# IRC 643(f) and QSBS Trust Stacking: The Statute Treasury Just Named
If you've talked to a wealth planner in the last twelve months about your founder shares, the pitch probably went something like this: the QSBS exclusion is worth $15 million per taxpayer after last year's tax bill, and each irrevocable non-grantor trust is a separate taxpayer, so you gift portions of your stock to three or four trusts and multiply the exclusion. A $60 million shelter instead of $15 million. Sometimes eight figures more than that.
You've probably also heard, in passing, that "the IRS is watching this." That warning is now specific. On May 20, 2026, Kenneth Kies, Treasury's Assistant Secretary for Tax Policy and acting IRS Chief Counsel, told a BakerHostetler seminar in Washington: "We don't like stacking, OK?" Two weeks earlier, Treasury attorney-adviser Evan Adams said the same thing at a different conference. The *Wall Street Journal* followed on June 29 with reporting that Treasury and the IRS are drafting guidance to shut down the more aggressive versions of the structure.
Nothing is final. No proposed regulation, no notice, no revenue ruling. But the statutory hook Treasury is signaling has a number: IRC § 643(f). Before you sign a term sheet with anyone selling you a multi-trust QSBS plan, you should know what that section actually says and where it plausibly applies to what you're being sold.
What QSBS stacking is trying to do
The [One Big Beautiful Bill Act](https://409.ai/articles/qsbs-one-big-beautiful-bill-act-section-1202-changes) raised the Section 1202 lifetime exclusion cap from $10 million to $15 million per taxpayer, per issuer, and lifted the aggregate gross-assets ceiling to $75 million, for QSBS issued after July 4, 2025. The cap is per taxpayer. Stacking exploits that.
The mechanics look like this. A founder holds 10 million shares of Newco stock she expects to sell for $80 million. Her exclusion cap is $15 million (or 10x her basis, if that's larger). Anything above the cap is a long-term capital gain. So she gifts 25% of the stock to an irrevocable non-grantor trust for her spouse, 25% to a similar trust for each of her two children, and keeps 25%. Four taxpayers. Four $15 million exclusion caps. On paper, $60 million of gain sheltered instead of $15 million.
That's the picture the pitch decks show. It's a real planning technique, and there are versions of it that plainly work. The version Treasury is pointing at is different: same grantor, same practical beneficiary, multiple trusts stacked to bank exclusion caps that were never designed to be duplicated.
What IRC 643(f) actually says
Section 643(f) is short. It gives the Treasury Secretary authority to treat two or more trusts as one trust for federal income tax purposes if two conditions are met:
1. The trusts have substantially the same grantor or grantors and substantially the same primary beneficiary or beneficiaries; and 2. A principal purpose of establishing the trusts (or of contributing property to them) is the avoidance of federal income tax.
For purposes of the test, spouses are treated as one person. The final regulations under the section were adopted in February 2019 at 26 CFR § 1.643(f)-1 and carry the same two-prong structure.
Read those elements against a stack of four trusts created in the same year, funded with the same QSBS from the same founder, benefiting the same nuclear family, and structured around a per-taxpayer cap that would not exist if the trusts were combined. The fit is not a stretch. It's why practitioners have flagged 643(f) as the natural weapon for a Treasury crackdown ever since OBBBA raised the cap.
There is a real textual quibble. Section 643(f) begins "For purposes of this subchapter" (Subchapter J, which covers estates and trusts), and Section 1202's exclusion sits in Subchapter P. Some tax lawyers argue Treasury has to reach for a different authority to collapse trusts specifically for QSBS purposes. Section 1202(k) is the more natural home for anti-abuse regulations under 1202, and it authorizes the Secretary to prescribe rules "to prevent the avoidance of the purposes of this section through splitups, shell corporations, partnerships, or otherwise."
The debate over which subsection the guidance sits under matters if you plan to litigate. For a founder deciding whether to set up a four-trust stack today, both roads point in the same direction.
Where the line probably runs
The Treasury officials' public comments have all noted that ordinary family planning is not the target. One non-grantor trust per family member, funded from a real economic transfer, administered by an independent trustee with meaningful discretion, is old law. That structure existed long before OBBBA. Neither 643(f) nor 1202(k) was written to blow it up.
What the officials keep coming back to is what they call the "same-beneficiary" problem. If a founder creates three trusts and all three name the same child as the primary beneficiary, or if the trusts have overlapping beneficiary classes so wide that they're economically fungible, you've built the structure 643(f) was drafted to reach. The trusts are separate on paper and unified in fact. That's the case where the second-prong "principal purpose is tax avoidance" analysis writes itself, because no non-tax reason for the duplication survives ten minutes of questioning.
A cleaner structure typically has: one trust per genuinely distinct beneficiary; an independent, non-related trustee with real discretion; distinct terms across trusts that reflect the beneficiaries' actual circumstances; and a non-tax purpose (creditor protection, generation-skipping planning, second-marriage protection, professional-liability shielding) documented at the time of formation, not reverse-engineered from a memo written after the sale.
The other planning traps founders miss
Two adjacent issues matter as much as 643(f) if you're planning around the exclusion.
First, the holding-period clock. The five-year clock runs from the date the stock was originally issued for the trust that receives the gift, so long as the transfer qualifies under Section 1202(h) as a gift, and the recipient trust is treated as having acquired the stock at the same time and in the same manner as the founder. If your original shares came from a [SAFE that converted](https://409.ai/articles/safes-qsbs-holding-period-conversion-section-1202), the clock started at conversion, not at the SAFE signing. Founders sometimes gift stock into a trust and then sell twelve months later, thinking each trust starts a fresh clock. It doesn't.
Second, every gift into a QSBS trust is a taxable gift for federal gift-tax purposes, and every gift needs a defensible valuation. If your transfer into three trusts uses inflated share values, you burn more of your lifetime gift exemption than you needed to. If it uses values that can't be supported, you invite an examination that has nothing to do with 643(f) at all. The [2026 lifetime gift and estate tax exemption is $15 million per person](https://409.ai/articles/gifting-startup-equity-2026-estate-tax-exemption), so a founder moving tens of millions of pre-exit stock into trusts needs the exemption arithmetic to work. That means a supportable [fair market value](https://409.ai/articles/409a-valuation-vs-fair-market-value) of the shares at gift date, prepared for gift-tax purposes rather than repurposed from a stale 409A. A private company's gift-tax valuation is not the same document as its [409A safe-harbor appraisal](https://409.ai/articles/409a-safe-harbor-price-vs-qualified-appraiser). Different standard, different discounts, often a different result.
What to actually do
If you're being pitched QSBS stacking today, three questions are worth asking your planner in writing before you fund anything:
Who is the primary beneficiary of each trust, and is that person genuinely different across trusts? If the answer is "my kids as a class" for all of them, you have the 643(f) problem. If each trust names one distinct child and reads differently to reflect it, you're on firmer ground.
What non-tax purpose is documented for each separate trust? Creditor protection, second-marriage planning, and generation-skipping are all real. "Multiplying the QSBS exclusion" is not a non-tax purpose, and a memo that says otherwise becomes an exhibit for the government.
What is the plan if Treasury issues guidance with a retroactive effective date? Most practitioners expect any final rule to apply prospectively, but the government has floated the theory that stacking has always been prohibited under existing anti-abuse doctrines. A defensive structure should be able to survive that argument on its own merits, not because it was grandfathered.
The QSBS exclusion is a genuine gift to founders. It is also one of the most-watched provisions in Subchapter P right now, and Treasury has stopped speaking about the aggressive versions in the abstract. If your plan can pass the 643(f) reading described above, do it and document it well. If it can't, the $60 million shelter on the whiteboard is a bet you may not want to place with a Treasury official already using the number of the statute out loud.
If you're preparing to claim the exclusion, a [QSBS attestation](https://409.ai/products/qsbs) built on defensible fair-market-value work is the record you'll want in front of you the first time anyone asks.