The AI Stock Windfall Has a Second Act, and It's Written by the FTB
I have a number of clients in construction and real estate who had a genuinely good year, and I've noticed a pattern in how some of them are deploying the proceeds. The cash that used to go into a second truck or a down payment on a fourplex is going into AI stocks. Nvidia, Palantir, a basket of semiconductor names, a fund with "AI" in the title. The reasoning is understandable. The sector is moving, the story is compelling, and nobody wants to be the person who watched it happen.
I'm not here to tell you whether that's a good investment. That's not my lane, and anyone who tells you with certainty where these names are headed is guessing with confidence. What I will tell you is the part the breathless market coverage leaves out entirely: the day you sell, you have a second transaction, and the other party is the government. Two of them, actually, if you live where my clients live.
Let me walk through what that second act actually looks like, because the difference between handling it well and handling it carelessly is, on a meaningful gain, a five-figure number.
The clock matters more than the call
Here's the rule that quietly governs everything. Under IRC §1222, whether your profit is a short-term or long-term capital gain turns on one fact: did you hold the asset for more than one year before selling. Not "about a year." More than a year, measured to the day.
Sell at the one-year mark or sooner, and your gain is short-term, taxed at your ordinary income rates under §1, the same schedule that applies to your wages and your business income. Hold for more than a year, and the gain becomes long-term, taxed under the preferential §1(h) rates of 0%, 15%, or 20% depending on your taxable income [verify 2026 breakpoints against primary source].
That spread is the whole game. The top ordinary federal rate is 37% [verify 2026]. The top long-term rate is 20%. On a $100,000 gain, the decision to hold for thirteen months instead of eleven can be the difference between roughly $37,000 and $20,000 in federal tax on the exact same trade. Same stock, same profit, same investor. Seventeen thousand dollars, decided by the calendar.
In a fast-moving sector, the temptation is to trade actively, to rotate out of one name and into the next. I understand the impulse. But every one of those round trips inside a year is a short-term gain, and the tax code is, frankly, not subtle about how it feels regarding short-term trading. It taxes it like a paycheck. The code rewards patience and penalizes churn, and that's a feature, not a bug.
The wash sale window is wider than you think
This is the one I see misstated constantly, including in the consumer-facing article that prompted a client to email me this week.
If you sell a stock at a loss and buy it back too soon, IRC §1091 disallows the loss. The piece that trips people up is the timing. It's not a 30-day rule. It's a 61-day window: 30 days before the sale, the day of, and 30 days after. Buy a substantially identical security anywhere inside that span and the loss is deferred, folded into the basis of the replacement shares rather than deducted now.
I emphasize the "before" half because that's where people get caught. An investor reads "within 30 days," sells at a loss on the 20th, and assumes a purchase on the 5th is irrelevant because it came earlier. It is not irrelevant. It's squarely inside the window, and the loss is gone for the year. With volatile AI names, where people are dollar-cost-averaging in and tax-loss harvesting out in the same quarter, this comes up far more than you'd expect.
The surtax the software forgets
Here's a line item that almost never appears in the popular coverage: the Net Investment Income Tax under IRC §1411. It's an additional 3.8% on net investment income, including your capital gains, once your modified adjusted gross income clears $250,000 for married-filing-jointly or $200,000 for single filers.
Note that those thresholds are statutory and have never been indexed for inflation since the tax took effect. They don't move. So every year that incomes drift up, more people quietly cross into NIIT territory without realizing the line exists. For the business owner who had a strong year and a sizable AI gain, the real top federal rate on that gain isn't 20%. It's 23.8%. If you're modeling the after-tax result without that 3.8%, your number is wrong before you even leave the federal return.
And then there's California
This is the part I most want my South Bay readers to absorb, because it's the part your tax software is least likely to dramatize for you.
California does not have a preferential capital gains rate. None. Under the R&TC, a capital gain is just income, taxed on the same graduated schedule as everything else, up to a top marginal rate of 13.3% [verify 2026 brackets and thresholds] once you factor in the 1% Mental Health Services Tax that R&TC §17043 imposes on taxable income over $1 million.
Sit with what that means. The entire federal short-term-versus-long-term drama, the thing the headlines fixate on, the FTB doesn't participate in. Hold for eleven months or thirteen, California taxes the gain the same way either way. So when a Torrance or Manhattan Beach client tells me they "waited for the long-term rate," I gently point out they optimized the federal layer and left the state layer exactly where it was. The holding period is real and worth respecting, but it's a federal lever, not a California one.
Stack it up and the marginal reality for a high-earning California investor on a long-term AI gain looks like 20% federal, plus 3.8% NIIT, plus up to 13.3% state. That's a combined rate north of 37% on a gain the financial press told you was taxed "at the low long-term rate." The low rate is a federal headline. Your actual bill is a California story.
Losses are a tool, within limits
The flip side is real and usable. If some of your AI positions are underwater while others are up, you can harvest the losses to offset the gains. This is legitimate, ordinary tax planning, and I encourage it when the positions and the timing line up.
Just know the guardrails. Under §1211(b), if your capital losses exceed your capital gains, you can only use $3,000 of the excess against ordinary income in a given year. The rest doesn't vanish, it carries forward indefinitely under §1212(b) until it's used up. So harvesting is powerful against gains, but as a way to shelter your wage or business income, it's a slow drip, not a faucet.
Why I think this belongs in a planning conversation, not an April conversation
Here's my actual opinion, since you're reading a CPA's blog and not a press release.
The entire consumer tax-prep industry is built on a comforting fiction: that taxes are a thing you report in the spring, a clerical event, a form you feed numbers into. For a W-2 employee with a standard deduction, fine. But for a business owner with real investment activity in a high-tax state, that model is actively expensive. By the time you're sitting at the software in March, every decision that mattered has already happened. The holding period is fixed. The wash sale either occurred or it didn't. The gain is short-term or long-term and no amount of careful data entry changes it.
The decisions that move the number are made before you sell, often months before, and they're made with your whole picture in view: your entity structure, your business income for the year, your other gains and losses, whether you're better served realizing in this year or the next. That's not something a wizard interface asks you. It's a conversation. And in California especially, it's a conversation worth having with someone who treats the state layer as a headline rather than an afterthought.
I'd rather talk to a client in October about a sale they're contemplating than in April about one they already made. One of those conversations can change the outcome. The other one is just me explaining it.
Let's look at your actual picture
If you're holding AI stocks, ETFs, or funds and you've had a strong year in your business, the smart move is to look at the whole thing together before you sell anything, not after. An advisory consultation is exactly the venue for that: we map your holding periods, your gain and loss positions, the NIIT exposure, and the California overlay, and we build a plan that fits your entity and your year rather than reacting to a 1099 you can no longer do anything about.
Schedule an advisory consultation and let's put real numbers to it while you still have every lever available.
If we're reviewing brokerage statements or anything else with account numbers or personal financial detail, please send those through our secure client portal rather than email. It keeps your information protected end to end.
This article is general information and reflects my perspective as of the date of publication. Tax figures noted as inflation-adjusted should be confirmed against current-year primary sources. Reading it does not create a client or advisory relationship, and it is not a substitute for advice tailored to your specific situation. For guidance on your own circumstances, please consult a qualified professional.