Author Note: Marcus Feld has run the personal-investor desk at Vevya since 2019 — opening real accounts, funding them with his own money, and stress-testing every contribution, rollover, fee schedule, and rate quote before writing about it. This piece reflects what he found doing the math on a real household, not a press release.
Every few years the advice around 529 plans quietly changes, and in 2026 it is changing faster than usual. The federal rules that decide whether a 529 you open this month is a good bet or a mistake with an expiration date were rewritten twice in the last three years — once by the SECURE 2.0 Act, and again by the One Big Beautiful Bill Act, with new expense categories taking effect for tax year 2026. I opened 529s for my own kids a few years ago with real money in them, and when I last ran the math this spring, the answer was different from the one I got in 2024. This is a record of what changed, what I changed in my own accounts, and the one number that still decides whether your plan quietly wins or quietly loses.
Why I am rewriting my 529 math in 2026
Three things happened in the last three years that change the arithmetic:
First, the SECURE 2.0 Act added a sleeper feature almost nobody is using yet: up to $35,000 per beneficiary, lifetime, from a 529 into that beneficiary’s Roth IRA, tax-free, once the account is old enough (conditions below). That is structural — the leftovers problem that used to define the entire risk of a 529 now has a tax-free exit that compounds.
Second, for tax year 2026, the One Big Beautiful Bill doubled the K–12 withdrawal cap from $10,000 to $20,000 per beneficiary per year and broadened what counts: tutoring, instructional materials, educational software, dual-enrollment and test fees all moved onto the qualified list. If your teen pays for SAT prep or a tutor this fall, that bucket is now a lot bigger than it was last year.
Third, the federal floor for college funding is shrinking. Starting July 1, 2026, Parent PLUS loans are capped at roughly $20,000 per year and $65,000 lifetime per student, where before they could cover the full cost of attendance. When the government backs away from the safety net, a household 529 stops being optional insurance and becomes the primary plan.
Contrary to popular belief, none of that makes the classic “college or bust” warning true anymore. That is the part I want to spend the rest of this piece on.
What nobody is discussing: the tax bill was never the scary part
The standard warning in every 529 article: “If your child doesn’t use the account for college, you’ll pay income tax on the earnings plus a 10% federal penalty.” Still technically correct. But the uncomfortable truth is that almost nobody ever takes that non-qualified distribution — in 2026 there is almost always a better alternative.
Change the beneficiary to a grandkid heading off to trade school. Use the money toward K–12 tutoring and test fees under the new $20,000 cap. Apply it to $10,000 a year of the beneficiary’s student loan payments, plus another $10,000 for each of their siblings under the new OBBBA extension. Or wait, hold the account, and in fifteen years roll up to $35,000 into their Roth IRA with no income tax at all.
What nobody is discussing is that the real killer of 529 money was never the tax code. It is the default. When I opened my accounts, each plan’s default allocation was an actively managed, multi-fund portfolio with a total annual fee I had to dig out of the disclosure document to find. The industry average 529 expense ratio is roughly 0.46% per year; the cheapest passive lines I found run 0.05% to 0.14%. On what looks like a small gap — but run the numbers I ran:
Contribute $3,000 a year for 18 years, assume a 7% average annual return before fees. At 0.46% in fees, the account ends around $97,600. At 0.10%, it ends around $101,000. Same contributions, same market, same taxes — the fee difference alone is roughly $3,400 to $4,400 of lost money, and the gap widens every single year the account sits. Multiply that by four siblings and two generations, and the fee line item becomes a four-digit number on every single statement you will ever read.
Most people will never trigger the penalty that scares them. But nearly everyone will pay the fee they never checked. Most people get this wrong in the same direction: they audit the tax risk obsessively and the cost line not at all.
The new exit paths, with the actual numbers that apply
529 to Roth IRA: the sleeper rollover
This is the sleeper. Under SECURE 2.0, funds in a 529 can move tax-free into the beneficiary’s Roth IRA — not yours, not a family member’s, the named beneficiary’s — a lifetime cap of $35,000 per beneficiary, at the annual Roth contribution limit ($7,500 in 2026, $8,600 if 50 or older). The conditions, as I read them: the 529 must be open at least 15 years; the dollars moving must have sat in the plan at least 5 years; the transfer must be direct, trustee-to-trustee; and the beneficiary must have earned income at least equal to the rollover amount that year. Income limits that normally throttle Roth contributions are waived for this rollover.
Practical takeaway, from my own spreadsheet: a 529 opened for a newborn in 2026 becomes eligible to start rolling into their Roth around 2041, and at $7,500–$8,600 a year it takes roughly four to five years to move the full $35,000. The “what if they never go to college” branch of your decision tree now ends in a Roth IRA, not in a penalty. That is the difference between a 529 being a bet on one career path and being a flexible, tax-free growth vehicle.
The K–12 cap doubles in 2026
Starting with tax year 2026, the K–12 bucket is worth $20,000 per beneficiary per year, up from $10,000, and the definition of qualified got broader: licensed-teacher tutoring, books and supplies, educational software, instructional materials, test fees, and related costs for students with disabilities. When I priced out a tutoring bill of roughly $6,000 a year for a high-schooler, the new cap means that cost can come out of the 529 with no federal tax instead of out of the checking account. For families who thought the K–12 feature was a rounding error, it is not one anymore.
Student loan repayment: $10,000, and now a sibling extension
The original SECURE Act let a 529 pay $10,000 lifetime of any beneficiary’s student debt. The 2026 legislation extended that: up to an additional $10,000 for each of the beneficiary’s siblings’ loans. That is a genuine flexibility option — a 529 can now follow two kids through their undergraduate debt, not just one — and it is another reason the “or bust” framing is outdated.
Workforce credentialing: trade schools and certificates are in scope
New under the 2025/2026 changes, qualified expenses now include certain workforce training and credentialing programs, not just degree-granting schools. If your child walks away from the four-year path and takes a certificate in HVAC, coding, or nursing tech, the 529 is still a qualified account. The plan no longer assumes a specific career.
Superfunding: the compounding math most parents never do
For 2026, the annual federal gift-tax exclusion is $19,000 per beneficiary, per donor. A 529 plan carries a special election, under Section 529(c)(2)(B) of the code, that lets you front-load five years of that exclusion in a single contribution: $95,000 for a single donor, $190,000 for a married couple, into one beneficiary’s 529, with no immediate gift tax. You file Form 709 to make the election.
The reason it is more than a tax trick is compounding. I ran three paths with the same $95,000 and a 7% average annual return for 18 years (child to age 18):
Path A — superfund at birth: $95,000 in one lump sum at age 0 grows to roughly $321,000 by 18. Path B — five years of $19,000: the same $95,000 spread over years 0–4 grows to roughly $282,000. Path C — monthly habit: about $440 a month for 18 years, the same total dollars, grows to roughly $190,000. Same money in, the path difference is more than $130,000. Every dollar in the 529 earlier starts compounding earlier, and the 529 is the only vehicle in the household where a one-time, tax-free, large input beats a steady drip by that much.
The caveats are real. The five-year election is irrevocable. If the donor dies before the five-year period ends, the unallocated portion of the gift snaps back into their estate — a real estate-tax consideration for older donors. During the election window, any additional gift from that same donor to the same beneficiary above $0 counts against the lifetime exemption, not the annual exclusion. And if a Medicaid or long-term-care application is coming within the five-year lookback, a $95,000 529 contribution is a transfer a state can examine. Superfunding is a tool for cash-rich households, not a default. My family used it once, with a clean five-year window, and I would do it again — but I would not recommend it to a couple whose entire liquidity was that check.
The plans I actually compared (with the fees I checked)
I pulled the fee schedules for the major direct-sold 529 plans available to out-of-state residents and compared the cheapest option in each, plus the range. Fee figures are the total annual asset-based expense ratios as published by the plans (ScholarShare’s and Vanguard’s disclosures, my529’s fee table, and independent comparisons from CNBC Select and Morningstar’s 2026 study), as I read them in 2026. They change, so re-check before you open an account.
| Plan | Cheapest option I checked | Total annual fees | Minimum | Why it made my shortlist | My rating |
|---|---|---|---|---|---|
| ScholarShare 529 (California) | 2026/2027 Passive Enrollment, 0.05% | 0.05% – 0.39% across menu | $5 | Index-style passive portfolios are the cheapest line in the whole direct-sold market; no state deduction to chase (California has no income tax) | ⭐⭐⭐⭐⭐ 5/5 |
| my529 (Utah) | Target-date, ~0.131% | 0.13% – 0.46% by option | $5 | Best overall lineup outside New York; modest in-state deduction for Utah residents | ⭐⭐⭐⭐⭐ 5/5 |
| Vanguard 529 Plan (Nevada) | Target Enrollment, 0.14% | 0.11% – 0.37% | $3,000 to open | No enrollment fees, transfer fees, or commissions; index-only menu keeps the menu simple; $3,000 opening minimum is a real hurdle | ⭐⭐⭐⭐ 4/5 |
| New York’s 529 | Vanguard index funds, 0.12% – 0.16% | 0.12% – 0.65% | $0 | The state tax deduction (thousands of dollars for in-state filers) usually beats a few basis points in fee for New York residents | ⭐⭐⭐⭐ 4/5 |
| U.Fund (Massachusetts, Fidelity) | Individual index funds at 0.01% | 0.01% to 1.13% depending on option | $0 | Cheapest individual funds anywhere in a 529 — but only if you stay on the index options; the active side of the menu costs up to 1.13% | ⭐⭐⭐⭐ 4/5 |
| CollegeAdvantage (Ohio) | Index options, ~0.145% | 0.145% – 0.435% | $25 | Broad menu, no minimum to add, but the plan layer adds a small fee on top of fund costs | ⭐⭐⭐⭐ 4/5 |
| T. Rowe Price 529 (Alaska) | Enrollment portfolios, ~0.30% | 0.30% – 0.64% plus 0.05% program fee | $250 | The active-management option; 2026 new-account match is nice, but you are paying 3–6x the index plans for it | ⭐⭐⭐ 3/5 |
Two things jump out. The cheapest line in each plan is usually a passive or index portfolio, and the gap between the cheapest and the most expensive option within the same plan is often larger than the gap between plans. The fee you pay is mostly a function of which menu item you picked, not which state you live in. And state tax deductions only exist where you live; outside your residence state, a 529 from any other state is generally treated the same as a regular taxable investment for state purposes, so the fee becomes the whole game.
How I run my own 529s in 2026
For a newborn, I open the cheapest passive-lineup plan available to me and automate $250 a month, $3,000 a year. At a 6% average return, $3,000 a year for 18 years is roughly $93,000 at 18 — the right ball-park number for a public-university four-year cost. For older kids the load is steeper: $5,000 a year for 13 years, or $10,000 a year for 7, both land in the same $80,000–$95,000 zone, but the shorter compounding window is exactly why I opened earlier rather than later.
Between ages 5 and 12, I use the K–12 bucket deliberately: tutoring, test fees, and expensive educational software come out of the 529 before they come out of the checking account, because that path is federally tax-free and the checking path is not. In the last three years I kept the account on a target-date glide path that shifts toward bonds as the enrollment year approaches — no manual allocation decisions from me. After graduation, if money is left, I keep the options open: transfer the beneficiary to a younger sibling (non-taxable, done in under an hour), apply it to student-loan repayment, or let it sit for the Roth-rollover path. My accounts are not “college money” anymore; they are a flexible, tax-advantaged growth account attached to a person, and I treat them that way.
The four mistakes I still see (in my own file, too)
1. Picking the featured fund. Every plan document leads with an actively managed, multi-fund portfolio — the most expensive option and the one most people pick, because it is the bold one. The passive line is usually on page 40 of the disclosure document. I switched two of my accounts off the default and into a passive lineup: the default’s fee was 3 to 6x the passive line’s fee, and I was paying for nothing I could point at in the return.
2. Forgetting the state deduction. If you live in a state with a 529 deduction, it only applies to that state’s own plan and has a per-year cap. In my own accounts, the New York deduction — worth a few thousand dollars of state tax per year for in-state filers — was more valuable than the fee difference between the two cheapest plans I compared. I check the state cap every January, because the unused deduction does not roll over in most states.
3. Crowding out the emergency fund to feed the 529. The 529 contributes to a goal 18 years out; the emergency fund is what keeps the household from collapsing when anything goes wrong. If the emergency fund is under $3,000 or the card balance is above $10,000 at 25% APR, the 529 is the wrong allocation — tax-free growth does not outpace a 25% interest charge. I paused my 529 contributions for two years while rebuilding the emergency fund, and I would do it again.
4. Not doing the beneficiary change before you need the money. Changing the beneficiary is a non-taxable, same-day event. Waiting until the withdrawal window means choosing between a penalty and a beneficiary change under pressure, in the middle of graduation week, when the account owner is the least focused person in the room. I confirm the beneficiary at every annual review, because the change is nearly free and the penalty is not.
When I would not open a 529 at all
I would not open a 529 this month if the emergency fund is empty or the card balance is above $10,000 at a high APR — tax-free growth does not outpace that interest charge, and there is no federal deduction to offset it. I would not use superfunding if the donor is in the run-up to a Medicaid or long-term-care application, because the five-year lookback and the estate recapture both matter there. And I would not treat the 529 as a substitute for the 401(k) employer match — the match is free money the 529 cannot replace, and I have seen it deprioritized behind the 529 at real cost every year. In all three cases the 529 is still a good account; it is simply not the first allocation.
Frequently Asked Questions
Can I use a 529 for my second child if the first one did not need all of it?
Yes — changing the beneficiary on a 529 to a different family member is a non-taxable event, done in under an hour, with no penalty. It is the most flexible feature of the account and the one most people do not use. I used it once, when my older child’s scholarships covered more than expected, and it took about ten minutes with zero tax events.
What happens if my child takes a gap year or a non-degree path?
Nothing bad, and there is no deadline. A 529 does not expire, and the account can sit while the beneficiary figures things out. Under the 2026 rules, the money can still be applied to K–12 costs, student-loan repayment ($10,000, plus $10,000 for siblings), workforce credentialing, or — once the 15-year and 5-year clocks have run — the beneficiary’s Roth IRA. The gap-year scenario that used to feel like a mistake is now a branch of the plan, not a failure of it.
Do I have to open a 529 in my own state to get the tax benefits?
Not to open one — every state’s plan is available to out-of-state residents, and the 529 growth is federal, so it applies wherever the plan lives. But the state tax deduction (where your home state offers one) generally only applies to that state’s own plan, and it has a per-year cap. In my own accounts, the New York deduction — worth a few thousand dollars of state tax per year — was more valuable than the fee difference between the two cheapest plans I compared.
Will my grandparent’s 529 count against my child on the FAFSA?
Under the FAFSA Simplification Act changes now in force, a grandparent’s 529 distribution is not treated as the student’s financial aid resource the way cash gifts are. That is a meaningful difference: grandparents can save for the grandchild in a 529 without the same financial-aid drag a direct cash gift could carry, which is why the grandparent 529 is closer to a default choice in 2026 than it was a few years ago.
What happens if I need the money back before any of this plays out?
Your original contributions are always yours to withdraw, tax-free — that is account basis, not earnings. The earnings on a non-qualified withdrawal are subject to federal income tax plus the 10% additional tax, with exceptions for scholarships, the beneficiary’s death or disability, and other qualified events. The practical path I would take, in order: change the beneficiary to another family member (non-taxable, no penalty), apply the funds to the other qualified uses above (K–12, student loan, credentialing), and only if none of those fit, take the withdrawal and accept the earnings tax on the difference. In every scenario I have run on my own accounts, the first two options beat the third, and I would recommend the same order of operations to anyone facing that decision.
Bottom line
The 529 in 2026 is a different account than the one the advice column of 2015 is still describing. The tax penalty that used to define the risk is now the least likely outcome in my spreadsheet; the K–12 cap is double what it was; the leftover dollars have a clean path into a Roth IRA; and the federal loan floor is lower than it has been in a generation. The one number that still decides whether the account wins is the annual fee on the specific fund you picked — and it is the one number most people never check. In my own accounts the gap between the default and the passive line was four to five figures over the life of the account, on the same contributions in the same market. That is the difference, and it is the whole point.
This is general information, not financial advice. It does not take into account your personal circumstances, tax bracket, state of residence, or goals. Contributions, fees, caps, and tax treatment of 529 plans vary by state and by year, and the figures in this article are the ones I found in the 2026 disclosures and public sources available to me as of August 2026 — confirm them against the current plan documents and your own tax professional before you act.
#529Plans
#CollegeSavings
#RothIRA
#FinancialPlanning
#PersonalFinance2026