QSBS Stacking: The Cap Is Per Taxpayer, and Treasury Noticed
The section 1202 exclusion is capped per taxpayer, so founders gift shares to separate non-grantor trusts and multiply it. The mechanism is one sentence of statute. Whether it survives is a different question.
The Section 1202 exclusion is capped at $15 million per taxpayer, per company. Read that sentence twice, because the planning technique known as stacking is nothing more than taking the words per taxpayer literally.
It works. It is in the statute. Congress reopened Section 1202 in 2025, expanded it, and left the provision that makes stacking possible untouched. And in May 2026 the Treasury official who oversees tax policy told a room of practitioners that the government does not like it and is writing guidance.
This is what the technique is, what it is worth, when it is worth nothing, and where it is going.
1 The mechanism is one sentence of statute
Section 1202(h)(2)(A) says that when QSBS is transferred by gift, the recipient is treated as having acquired it in the same manner as you did, and as having held it for the period you held it.
Two things follow. The shares stay QSBS in the recipient's hands. And your holding period comes with them, so a gift made in year four does not restart the clock.
Combine that with a cap measured per taxpayer, and the arithmetic is obvious. A founder whose gain will exceed $15 million gifts part of the position to other taxpayers, and each one brings its own cap.
2 Only one kind of trust is a separate taxpayer
This is where most explanations get vague, and it is the only technical question that matters.
Grantor trust. You are treated as the owner for income tax purposes. The trust is ignored; its income lands on your return. Same taxpayer, same cap, no benefit.
Non-grantor trust. The trust is its own taxpayer with its own employer identification number, filing its own Form 1041. Separate taxpayer, separate cap.
Stacking requires the second. A trust is a grantor trust under sections 671 to 679 if you retained almost any meaningful string: the power to revoke, income that can reach you or your spouse, the power to substitute assets of equivalent value. Cutting all of them is what makes the trust a separate taxpayer, and it is not a drafting formality. It means the property genuinely is not yours any more.
There is an uncomfortable consequence. The intentionally defective grantor trust is the workhorse of estate planning precisely because it is a grantor trust: you keep paying the income tax, which is effectively a further tax-free transfer to the beneficiaries, while the assets sit outside your estate. Excellent for estate tax. Worth nothing at all for Section 1202.
3 The leverage is a timing mismatch
Gift tax measures what you gave away on the day you gave it. The Section 1202 cap measures gain at exit. Those two dates can be years and several orders of magnitude apart, and the whole technique lives in that gap.
For 2026 the lifetime exemption is $15,000,000 per person and the annual exclusion is $19,000 per donee, both from Rev. Proc. 2025-32. The annual exclusion does not touch the lifetime amount, so early funding can be partly free, though a gift to a trust only qualifies for it if the beneficiary holds a present interest.
The lifetime exemption and the Section 1202 cap are both $15 million in 2026. That is a coincidence of the current figures. They are unrelated, indexed separately, and measured at different dates.
4 When it is worth nothing at all
Stacking multiplies the dollar cap. It does nothing to the other limb.
The per-issuer cap is the greater of the dollar amount or ten times your basis in the stock. A founder who incorporated out of an LLC carries, under Section 1202(i), a basis equal to the fair market value contributed. Convert at $6 million and the cap is $60 million without a single trust, without spending a dollar of exemption, and without giving anything away.
So the first question is not which trust. It is which limb of the cap binds. If ten times basis already covers the expected gain, the conversation ends there, and anyone who starts it with trusts has the order wrong.
5 The vehicles, and what each one costs you
Every one of these is drafted by counsel. The table is to make the trade-offs legible, not to choose for anyone.
Non-grantor trust per beneficiary. The standard structure and the hardest to attack. Uses lifetime exemption. You have no access.
SLANT, a spousal lifetime access non-grantor trust. Your spouse is a beneficiary, so the household retains indirect access. Section 677 would normally make that a grantor trust, which is defeated by requiring an adverse party — someone with a real beneficial interest who would personally lose by the distribution — to approve distributions.
ING, the incomplete-gift non-grantor trust. You can be a beneficiary of your own non-grantor trust, through a distribution committee of adverse parties, and the gift is incomplete so it consumes no exemption. These are now on the IRS no-rule list, which is a signal in itself.
Intentionally defective grantor trust. The standard estate planning vehicle, and no help here at all, because the grantor is the taxpayer.
One further point that matters for anyone in a high-tax state: a non-grantor trust sited in a state with no income tax can also keep the gain out of the state net. For a California founder facing 13.3% on a gain the federal government is excluding entirely, that second effect is sometimes larger than the first. California's sourcing rules are aggressive and this needs real counsel rather than a template.
6 What Treasury has actually said, and actually done
These are different, and the difference is the whole story.
May 2026. A Treasury attorney-adviser raises concerns publicly. On May 20, Kenneth Kies, Assistant Secretary for Tax Policy and acting IRS Chief Counsel, tells a conference that the government does not like stacking and that guidance is being developed. His described focus is arrangements that go beyond one trust per family member.
June 2026. The Wall Street Journal reports on the technique and on the government's objections, describing a proposal in which two co-founders would gift stock to eighteen trusts.
September 2026. Commentators in Tax Notes urge Treasury to use its Section 1202(k) authority against the aggressive structures, while accepting that well-built stacking appears to work under current law.
Since then. No notice. No proposed regulation. No revenue ruling. Conference remarks bind nobody.
If and when something arrives, it will rest on one of two authorities, and neither is comfortable.
Section 643(f) treats two or more trusts as one where they have substantially the same grantor and substantially the same primary beneficiaries and a principal purpose is avoiding income tax. The implementing regulation was finalized in 2019 and treats spouses as one person, so matching trusts created by each spouse for the same children can be collapsed. The difficulty is textual: Section 643(f) applies "for purposes of this subchapter", meaning Subchapter J, while Section 1202 sits in Subchapter P.
Section 1202(k) directs Treasury to write regulations preventing avoidance of the section's purposes through split-ups, shell corporations, partnerships "or otherwise". That is a cleaner foundation, rarely used, and any rule built on it has to coexist with the gift provision Congress left in place.
Underneath both sits a harder problem. After the Supreme Court's 2024 decision in Loper Bright, an agency no longer receives deference for its reading of an ambiguous statute. A regulation forbidding an outcome the words plainly permit is on weaker ground than it would have been three years ago.
7 What is exposed, and what is not
Treasury's complaint has a shape. It is not that founders make gifts to their children.
The timing item carries independent weight. A gift made once a sale is a practical certainty runs into assignment of income and the step transaction doctrine, both of which long predate this debate and neither of which requires a new regulation to apply.
8 Could any of it reach backwards
Section 7805(b) generally stops a regulation applying before the earliest of its final publication, its proposal, or a notice describing what it will contain. On that reading, the first notice sets the date and completed planning is left alone.
Two caveats are worth holding. Section 7805(b)(3) permits retroactive rules to prevent abuse. And the government has suggested it may simply argue that the most aggressive structures never worked, under doctrines already on the books.
9 The part that gets skipped
A gift reported on Form 709 with the adequate disclosure the regulations require, supported by a qualified appraisal, starts a three-year limitations period under Treasury Regulation 301.6501(c)-1(f).
Leave the disclosure out and that clock never starts. The IRS can revalue the gift years later, after the company has sold and the proceeds have been spent, and the exposure is measured against the value then rather than the value you reported. On early-stage stock that later sells for tens of millions, this is a larger practical risk than the stacking debate itself, and it has nothing to do with Treasury's announcement.
10 What this means if you are holding QSBS now
- Find out which limb of the cap binds. If ten times basis covers your expected gain, stop. No trust improves it.
- If the dollar cap binds, the window is now rather than later. Guidance is usually prospective, and gift value only goes up.
- Build it as estate planning, because that is what survives. One trust per real beneficiary, genuine trustees, a reason that would exist if the tax benefit vanished.
- File the 709 properly and get the appraisal. This is the cheapest risk reduction available and the one most often skipped.
- Do not do it on the eve of a deal. That structure was never safe, with or without a regulation saying so.
Our QSBS calculator will tell you in six questions whether your shares qualify and which limb of the cap applies, which is the fact that decides whether any of this is worth discussing. The structuring itself belongs with your attorney, and our founder exit tax planning work is the modeling that sits underneath it.
General information for owner-operators, not tax or legal advice. Parikh Financial does not draft trusts and does not provide attest services. Thresholds and guidance change; confirm current requirements for your own facts before acting.
Frequently asked
Questions, answered
Is stacking still legal?
Yes, under current law. The cap applies per taxpayer, a non-grantor trust is a separate taxpayer, and section 1202(h)(2)(A) preserves qualified status for gifted shares. Treasury officials criticised aggressive stacking in May 2026 and said guidance was being developed. No notice, proposed regulation or ruling has been issued.
Does it work with a grantor trust?
No. A grantor trust is disregarded, so its income is treated as yours and there is no second taxpayer. Only a non-grantor trust, filing its own Form 1041 under its own EIN, carries its own cap. The intentionally defective grantor trust that is so useful for estate tax is no help here at all.
When is it worth nothing?
When ten times your basis already covers the expected gain. The cap is the greater of the dollar amount or ten times basis, and stacking multiplies only the dollar amount. Establish which limb binds before anyone mentions a trust.
How much exemption does it use?
Gift tax measures value on the day of the gift; the cap applies to gain at exit. Gifting $300,000 to each of three trusts uses $900,000 of exemption and can support $45 million of additional exclusion. For 2026 the lifetime exemption is $15,000,000 per person and the annual exclusion is $19,000 per donee.
What makes a structure risky?
More trusts than real beneficiaries, overlapping beneficiaries, one trustee and identical terms across all of them, gifts made once a sale is in view, and no purpose beyond the exclusion. Section 643(f) lets the IRS treat trusts sharing substantially the same grantor and primary beneficiaries as one.