The QSBS Question Most Founders Get Wrong

There is a particular shape of conversation that happens, with some frequency, in the months after a founder's first meaningful sale of company stock. The numbers come back from the accountant. The federal exclusion was smaller than expected, or applied to less of the proceeds than the founder believed it would. The state took its share regardless. The non-excluded portion got taxed at a rate that surprised everyone, including, sometimes, the accountant.

The reaction in the room tends to be the same: I thought QSBS covered this.

Section 1202 of the Internal Revenue Code — the qualified small business stock provision that lets founders, in the right circumstances, exclude a portion of their gain from federal taxation — is one of the most powerful planning tools available to early-stage shareholders. It is also one of the most consistently misunderstood. The misunderstanding is not, mostly, the founder's fault. The provision has changed twice in living memory: first when the 100 percent exclusion was made permanent in 2015, and again when the One Big Beautiful Bill Act, signed into law on July 4, 2025, restructured the per-issuer cap, the holding period schedule, and the gross-asset threshold for stock issued after that date.

Most of what is currently circulating about QSBS — in podcasts, in legal-marketing blog posts, in the ambient knowledge that founders absorb from their networks — was written under the older regime. It is not wrong, exactly. It is stale. And the founders who rely on it tend not to discover the staleness until a return is being filed and the math comes back differently than it should have.

This piece is about four of those gaps. None of them are exotic. All of them turn on details that the wider commentary has not, yet, caught up to.

The dual-regime problem

The cleanest way to understand the post-OBBBA QSBS landscape is to recognize that, for many active founders, two different versions of the rule now apply to the same cap table.

Stock issued on or before July 4, 2025 remains under the prior regime. The exclusion is binary: hold for five years and exclude one hundred percent of gain, up to ten million dollars of qualifying gain per issuer or ten times basis, whichever is greater. Stock issued after that date sits under a different schedule. The dollar cap is fifteen million per issuer — indexed for inflation beginning in 2027 — with the ten-times-basis alternative carried over unchanged from the prior regime. The exclusion is no longer binary; it is tiered, with fifty percent available at three years, seventy-five at four, and the full one hundred percent at five and beyond.

A founder whose company has issued multiple rounds of stock since incorporation may, without realizing it, hold lots from both regimes. Common stock issued at founding, restricted shares granted in a 2023 funding round, and additional shares acquired in a 2026 employee program would, on a single tender at exit, be governed by different sets of rules. The accountant preparing the return will need to track the issuance date of each lot and apply the corresponding regime — a task complicated by the fact that the issuance date for §1202 purposes is not always the issuance date the founder remembers, particularly where stock was acquired through option exercise or restricted-stock transfer.

There is a related point worth flagging here, because it is the way the wider commentary most often misleads: pre-OBBBA QSBS cannot be "upgraded" to the new regime through a tax-free exchange. Stock acquired in a §351 nonrecognition exchange or a §368 reorganization, in exchange for QSBS issued on or before July 4, 2025, remains under the prior regime. The grandfathering is durable, and so are its limits. The pre-OBBBA founder who hopes a future restructuring will sweep their old stock into the new caps will find that the statute does not permit it.

The practical effect is that the dual-regime founder may have a fully qualifying pre-2025 lot eligible for the older one-hundred-percent exclusion at the older ten-million cap, alongside a partially qualifying post-2025 lot subject to the tiered schedule. Whether prior pre-OBBBA exclusions on the same issuer's stock reduce the new fifteen-million cap is a question on which the practitioner consensus and the published guidance are not yet fully aligned. Combining the two regimes without distinguishing them, as much of the wider commentary continues to do, is the most common error in this area, and the most consequential. The dual-regime founder who plans for a single, unified exclusion has, almost by definition, planned for the wrong number.

A question worth asking

On the founder's cap table, how many discrete lots of stock exist, and was each one issued on or before July 4, 2025, or after?

The tiered exclusion timing trap

For stock issued after July 4, 2025, the elimination of the binary five-year cliff has changed the math on early exits in a way that is rarely modeled honestly at the time of acquisition.

Under the prior regime, the planning question was simple. Hold for five years and the exclusion is fully available; sell earlier and it is not. The decision was bright-line. Under the new schedule, every additional month of holding period now changes the after-tax outcome of a sale. A three-year hold produces a fifty-percent exclusion. A four-year hold produces seventy-five. A five-year hold produces one hundred. The non-excluded portion, in each case, is subject to a maximum federal rate of twenty-eight percent — above the long-term capital gains rate that applies to ordinary appreciated stock — and the 3.8 percent net investment income tax under §1411 applies on top of it. The all-in federal rate on the non-excluded portion can therefore approach thirty-two percent before any state tax.

The consequence is that a founder weighing a partial exit at year three or year four is no longer making a binary decision about whether QSBS applies. They are making a calculus decision about how much of the gain they are willing to surrender to time. For a founder selling into secondary at year three to fund a house purchase or to diversify, the math is no longer "do I qualify or not." It is "how much exclusion am I leaving on the table by selling now, and is the cost of waiting twenty-four months worth what I would gain."

The wider commentary, written largely under the older regime, tends to treat the holding period as a hurdle rather than a curve. It is a curve now. Founders who frame it as a hurdle tend to make decisions either earlier or later than the underlying math would suggest.

The trust-stacking question

For founders whose anticipated gain on exit will exceed the per-issuer cap — fifteen million for post-2025 stock, ten million for pre-2025 stock — the question of whether trusts can be used to stack additional caps is where some of the most aggressive QSBS planning happens, and where some of the largest mistakes are made.

The provision permits non-grantor trusts to be treated, for §1202 purposes, as separate taxpayers, each with its own per-issuer cap. A founder who transfers QSBS-eligible stock to multiple non-grantor trusts before exit can, in theory, multiply the available exclusion. The IRS has not, to date, formally challenged this position in published guidance — but the absence of a challenge is not the same as endorsement. There is no published Revenue Ruling, final regulation, or judicial opinion squarely confirming the structure under §1202. The practice rests on practitioner consensus rather than authoritative pronouncement.

What constrains it materially are two doctrines, applied with some force by the IRS in recent examination cycles.

The first is the multiple-trust rule of §643(f), which permits the IRS to aggregate trusts where two conditions are both met: the trusts have substantially the same grantor and substantially the same primary beneficiary, and a principal purpose of their separate existence is the avoidance of federal income tax. Both prongs are required. Trust planning structured around separate primary beneficiaries — different children, for instance, rather than a single grantor's spouse and the same grantor's spouse — sits outside the rule's reach as a doctrinal matter, though the factual record always matters.

The second is the reciprocal trust doctrine, which applies most often where a founder and a spouse establish substantially mirror-image trusts for each other's benefit. The doctrine is well-established in the estate-tax context, where it originates in United States v. Estate of Grace. Whether and how it applies to income-tax planning, and to §1202 specifically, has not been resolved in published IRS guidance or court opinion. Practitioners flag it as a potential analytical risk for mirror-image structures established in close proximity to a known exit, and it is the kind of risk that prudent counsel will pressure-test before recommending the planning.

The wider commentary on QSBS trust stacking, particularly the more aggressive variants, often does not engage these doctrines at the level of detail required. It is one of the cleanest examples in the tax-planning world of where the spread between what is technically possible and what is doctrinally defensible is large. The founder who relies on the technically-possible end of that spread, without independent counsel pressure-testing the doctrinal end, may be relying on a structure the IRS will not respect when it matters.

State non-conformity, and what California buys you

The federal exclusion is, in many cases, only part of the founder's tax bill. The state where the founder resides at the time of the sale, and the state where the company was incorporated and operates, may both impose their own tax on the gain — and may or may not conform to the federal §1202 exclusion in doing so.

California, for reasons of its own, does not conform to §1202. A founder selling QSBS-eligible stock while resident in California is fully excluded from federal tax, on the qualifying portion, and fully taxed by California on the same gain. The disjunction is not subtle. A founder with a $20 million gain on QSBS-eligible stock issued after 2025 might exclude $15 million federally, owe federal tax at the §1202 rate on the remaining $5 million, and owe California state tax on the entire $20 million at the state's top marginal rate. The federal exclusion is real. The California tax is also real. Combining them honestly produces a different number than the one many California founders are first quoted by their accountants.

California is not alone. Alabama, Mississippi, and Pennsylvania remain fully non-conforming. Hawaii allows a partial exclusion only. New Jersey was non-conforming for years but has now conformed for tax years beginning on or after January 1, 2026. The map shifts, sometimes from one tax year to the next, and a founder whose state of residence changes during the year of sale may find themselves in a different conformity regime than the one their planning assumed.

The strategies that exist for managing the state-conformity problem — pre-sale residency moves, gifts of stock to trusts established in non-conforming or favorable states, sales structured through entities incorporated outside the high-tax jurisdiction — all carry the same set of doctrine concerns described in the trust-stacking section above. They work, in some cases. They are also among the most heavily examined by aggressive state revenue authorities, and the procedural questions surrounding them — what counts as a change of residency, when a trust is properly seated in another state, how a sale through a separate entity is characterized — are areas of considerable ongoing litigation.

The point is not that the planning is impossible. The point is that the federal exclusion, alone, is not the whole answer for founders in non-conforming states. The wider commentary tends to focus on the federal side because the federal side is national; the state side is jurisdictional. But the founder's bill is paid jointly, and a planning conversation that does not contemplate both sides is, by definition, incomplete.

What this piece doesn't pull

Several threads connected to QSBS have been left for separate pieces. The interaction between QSBS and installment-sale treatment, which is structurally messy, is one. The treatment of QSBS in mergers and acquisitions where stock is exchanged for buyer stock — and the question of whether the acquired stock can inherit QSBS character at all — is another. The §1045 rollover provision, which permits gain on the sale of QSBS to be deferred into other QSBS within sixty days, is a third, and the relevant escape valve for founders considering a sale before any of the new tiered thresholds have been reached.

Each of these is a substantive enough subject to deserve its own treatment. The point of this piece is not to map all of QSBS, but to surface the four gaps where the wider commentary is, at the moment, most consistently behind the underlying law.

Where this lands

A founder who has read this far with a meaningful QSBS position is, ideally, holding stock that has not yet been sold. The decisions described above mostly compound forward — they sit between the current moment and a future closing — and the room to make them is widest before the transaction is being negotiated, narrowest after the LOI is signed, and effectively closed once the sale is complete.

The Check on this site is not a §1202 audit. It is shorter, six questions, scoped to the kinds of tax drag that show up in an investment portfolio rather than to the structural decisions of a pre-exit shareholder. For founders inside the window described in What a Signed LOI Actually Starts, the questions there are closer to the right ones. For founders past the close, the Check reframes the relevant questions to the years that come after — the year with little or no W-2 income, the years of newly liquid capital, and the long compounding question of how much of the rest of it will be lost to the tax it doesn't appear, on first glance, to owe.

Take the Check →