OpenAI PPU Holder: Planning Around Tax Uncertainty
An OpenAI engineer holds Profit Participation Units with no clear precedent for tax treatment. Here is the framework we used to plan around the ambiguity.
An engineer at OpenAI joined in 2023 and received a grant denominated in Profit Participation Units, or PPUs. Unlike RSUs, options, or restricted stock, PPUs are a contractual entitlement to a share of OpenAI’s profits (subject to a cap) rather than an equity interest in the entity. Because OpenAI’s parent entity is a capped-profit LLC owned by a 501(c)(3) nonprofit, traditional equity mechanics do not apply. His grant value on paper looks large, rumored tender pricing suggests $1.4M, but he has no precedent for how the IRS will treat delivery when it happens. His question to us: how do I plan for a tax event I cannot characterize?
Situation
His grant structure (as disclosed in his offer letter and subsequent internal communications):
- 300 PPUs vesting over 4 years, 25% cliff at year one, then quarterly.
- Fully vested by late 2027.
- Each PPU entitles him to a pro-rata share of profit distributions after investor return caps are met.
- Tender offers have been conducted by OpenAI at implied PPU values of $3,500 to $4,800 per unit, with strict eligibility (employees only, tenure minimums).
- No option-style strike price. No dividend right until caps are satisfied.
Tax treatment is genuinely unclear. Three plausible characterizations:
-
IRC §83 property compensation. If a PPU is treated as “property” transferred in connection with services, delivery at vest produces ordinary income equal to FMV. This is how RSUs work. But PPUs are arguably not property because they are not stock and do not carry ownership rights.
-
IRC §409A deferred compensation. If PPUs are rights to future compensation tied to company performance, they fall under §409A. Delivery is taxed as ordinary income on receipt, but §409A imposes brutal penalties (20% additional tax plus interest) if the arrangement does not meet strict timing and form rules.
-
Partnership interest (IRC §707 or §83 profits interest). If OpenAI’s capped-profit LLC treats the PPU as a profits interest under Rev. Proc. 93-27, it could be non-taxable at grant and produce self-employment income as profits are distributed.
OpenAI has not publicly taken a position, and guidance from its general counsel in 2024 suggested Form W-2 reporting at delivery, but with substantial uncertainty. The planning question: do I withhold conservatively, do I set aside cash at the tender-offer level of value, or do I treat PPU tender proceeds as ordinary income and pay through withholding?
What we modeled
We ran four scenarios around the 2025 tender:
| Assumption | Effective rate | Action |
|---|---|---|
| W-2 ordinary income, §83 analogous | 54.1% (top fed + CA + NIIT on other income) | Rely on OpenAI’s withholding; plan like RSU |
| §409A with compliance | 54.1% | Same as above; no penalty |
| §409A non-compliant | 74.1% (base + 20% §409A penalty + interest) | Worst case; set aside reserves |
| Profits-interest treatment | 20-23.8% long-term gains plus SE tax (if applicable) | Best case; hold and defer |
He could not choose among scenarios. OpenAI’s characterization would determine his treatment. What he could do is prepare for the worst-case cash need while the better outcomes remained possible.
The 2025 tender cleared at approximately $4,200 per PPU. He had 75 vested units (25% cliff plus three quarters). He participated with 40 units, generating $168,000 of gross tender proceeds. OpenAI withheld $36,960 (22% federal) plus $17,196 (10.23% California), total $54,156. That implied W-2 characterization.
What he did
He assumed §83 / W-2 treatment and treated the $168,000 as ordinary income. He set aside an additional $32,000 in a savings account specifically earmarked for “PPU tax surprise,” in case the characterization shifted to require higher withholding or if §409A penalties applied. He made a $28,000 Q4 estimated payment to cover the federal gap, calculated against a 37% effective bracket.
He also consulted with a tax attorney (not his CPA) to get a written opinion on §409A compliance. The attorney confirmed that OpenAI’s PPU plan appeared to use short-term deferrals paid within 2.5 months of vest, which generally satisfies the §409A short-term deferral exception under Treas. Reg. §1.409A-1(b)(4). This reduced the probability of the 20% penalty scenario meaningfully.
For the remaining 35 vested units not tendered, and for future tranches, his plan was to continue participating in each tender (subject to liquidity needs) and assume ordinary income treatment until OpenAI or the IRS provides clearer guidance. He declined to hold PPUs past vesting in expectation of capital gains treatment, given that the IRS has issued no guidance supporting that view.
What he wishes he had done differently
He did not file Form 8275 (Disclosure Statement) with his return. Because the tax treatment of PPUs is unsettled, a taxpayer can file Form 8275 to disclose the position taken, which prevents accuracy-related penalties if the IRS later takes a different view. The incremental cost was $0; the downside protection was meaningful. His CPA did not raise it, and he did not know to ask.
He also did not track his PPU vesting against the §409A short-term deferral rule carefully. If OpenAI’s tender payment slipped past March 15 of the year after vest, §409A compliance could fail. He now maintains a spreadsheet with each vest date and the corresponding 2.5-month deadline, so he will notice if a future tender creates a compliance issue.
Finally, he held off on charitable giving in 2025 because he was worried about cash flow. Given the income spike, a donor-advised fund contribution would have generated a deduction at his 37% marginal rate, while letting him grant out the DAF to charities over multiple years. $30k of appreciated public stock contributed to a DAF would have saved approximately $11,100 of federal tax plus $3,990 of California tax, for a net benefit of $15,090. He donated $8k in cash instead, saving $3,900. The opportunity cost was ~$11k.
Frequently asked
Why doesn’t OpenAI issue standard RSUs or options?
OpenAI’s parent entity is a 501(c)(3) nonprofit that holds the governance rights of OpenAI, LP (now OpenAI Global, LLC), a capped-profit entity. Standard equity structures do not translate cleanly to a capped-profit LLC, and traditional options would create characterization issues under §409A. PPUs are a custom instrument designed around those constraints.
Will PPUs eventually get IRS guidance?
Maybe. The IRS has historically issued guidance on novel equity structures only after widespread adoption. OpenAI’s structure is novel enough and the employee count large enough that guidance is plausible, but no proposed regulations have been published as of mid-2025.
Should I treat PPU tender proceeds as capital gains?
Almost certainly not. The compensatory nature of the grant and the absence of an investment risk at the time of grant argue strongly for ordinary income treatment. Taking a capital gains position on PPU tender proceeds without a written tax opinion and a Form 8275 disclosure would expose you to substantial understatement penalties.
If I leave OpenAI before my PPUs fully vest, what happens to unvested units?
Forfeited. Standard vesting forfeiture rules apply. Vested but undelivered PPUs depend on the plan’s post-termination provisions.
Does OpenAI’s PPU qualify for QSBS?
No. §1202 QSBS applies to stock in a domestic C corporation. PPUs are not stock.
Composite scenario drawn from common patterns in our advisor network's casework. Names, companies, and exact numbers are illustrative. Not tax, legal, or investment advice.