How to Invest an Inheritance in AI-Driven Volatility

The quiet windfall rule Wall Street actually follows

isn’t glamorous: pros slow the process, not the money. When a client inherits real dollars during 2025’s AI-fueled whipsaws, the first move is to stop moving. Freeze decisions for a beat, map cash needs, and hard-cap your risk before buying anything with “AI” in the ticker or the story. Sounds boring. It’s how you keep the win.

Why the pause? Because behavior eats spreadsheets. Morningstar’s 2024 Mind the Gap study shows the average investor lagged the funds they owned by about 1.7% per year over the prior decade due to poor timing and impulse shifts. And Vanguard’s research (2012) found lump-sum investing beat dollar-cost averaging roughly two-thirds of the time across U.S., U.K., and Australian markets, but that only helps if you can actually stick to the plan without chasing the shiniest narrative of the week. AI is the shiniest narrative of the year.

Rule of thumb I use with heirs: slow decisions, fast safeguards.

Here’s the framework I give families who call me from the parking lot of the lawyer’s office:

• Adopt a 30-90 day cooling-off period. Park the inheritance in insured cash or short Treasuries while emotions settle. You’re not “missing out”, you’re buying clarity. Earlier this year, 3-6 month Treasury bills yielded near 5% annualized; that’s decent rent while you think. If you need a nudge, set a calendar invite and don’t touch buy buttons until it dings.
• Write a one-page investment policy before funding accounts. Keep it simple and specific: targets, risk limits, rebalancing rules. Use plain numbers:
  1. Return target and guardrails (e.g., CPI + 3% long-run, max equity drawdown tolerance 20%).
  2. Asset mix and bands (e.g., 60/35/5 equity/bonds/cash with around 7% bands before rebalancing).
  3. Funding sequence (cash bucket for 12-24 months of spending, then taxable, then IRAs).
  4. When to rebalance (quarterly or when bands breach, no vibes-based overrides).
• Set an AI exposure cap up front. Decide what percent of the portfolio can be tied to AI themes (chips, data-center REITs, model vendors, automation). Make it a hard number. In 2024, AI-linked mega caps swelled to roughly 30% of the S&P 500’s weight, great businesses, but that concentration means a bad AI tape can move your whole life. I usually see 5-15% thematic caps work; your number may be lower if the inheritance will fund living expenses.

If this feels conservative, good. The 30-90 day pause catches the urges that wreck compounding. The one-pager makes decisions boring and repeatable. The cap keeps “AI” from becoming your entire risk budget by accident. We’ll get to taxes and account titling in a second, because those matter, but the order of operations is everything. As a guy who’s watched more than a few windfalls morph into accidental momentum trades, I can tell you: slow the process, not the money. You’ll sleep better, and the math tends to behave.

Before buying a single share: taxes, titles, and time windows

Housekeeping isn’t sexy, but it’s where most inheritance mistakes happen. Quick story: years ago, a client sold a big, low-basis stock position the week after a parent passed, before anyone checked the cost basis. They paid six-figures in capital gains that should’ve been zero. That stings. Let’s not do that.

Step-up in basis (taxable accounts)

• In the U.S., most taxable assets receive a step-up in basis to the fair market value on the date of death. That means embedded gains often vanish at death. IRAs and 401(k)s do not get a step-up.
• Verify lots before selling. Get the date-of-death valuation from the executor’s CPA or the custodian, and make the broker adjust cost basis records. If you sell before the step-up is recorded, you can fabricate a tax bill for no reason.
• Community property nuance: in community property states, the survivor often gets a full step-up on community assets. In separate-property states, it’s usually just the decedent’s half. Don’t guess, confirm.

Inherited IRAs: the 10-year clock is real

• Under the SECURE Act of 2019, most non-spouse beneficiaries must empty inherited IRAs within 10 years. Depending on whether the decedent died after their required beginning date, annual RMDs may also apply during those 10 years. The IRS provided penalty relief for missed inherited-IRA RMDs for tax years 2021-2024 (Notice 2024-35).
• SECURE 2.0 (2022) cut the penalty for missed RMDs to 25% (and potentially 10% if corrected in the correction window). That’s still a bad day, but not the old 50% sledgehammer.
• Spouses have different options (treat as own, remain beneficiary, etc.). Minor children, disabled beneficiaries, and a few others can qualify for “eligible designated beneficiary” exceptions. Translation: the rules are fussy, get them right up front.

Cash safety: park the proceeds the boring way

• FDIC insurance is $250,000 per depositor, per insured bank, per ownership category. Title the accounts correctly and you can increase coverage. If you’ve got $1 million in a single-titled account at one bank, that’s $750k of uninsured risk. Spread it.
• SIPC typically protects up to $500,000 per brokerage, including $250,000 for cash. SIPC covers custody failure, not market losses. Some brokers buy excess coverage; nice, but don’t rely on it for investment risk.
• While you finalize the plan, keep proceeds in Treasury bills or a government money market fund (these generally hold at least ~99.5% in cash, U.S. government securities, and repos under Rule 2a-7). In late 2024, 3-6 month T-bills yielded around ~5%; yields this year are still competitive versus the headache of reaching for credit risk you don’t understand.

Estate tax and titling: don’t move pieces before the board is set

• As of 2024, the federal estate tax exemption was $13.61 million per person (double for many married couples with portability). The higher exemption is scheduled to sunset after 2025 under current law. If estate tax is in play, coordinate with the attorney before retitling or liquidating assets, especially concentrated positions or private holdings.
• Paperwork that matters now: Form 56 (notice of fiduciary relationship), an EIN if there’s an estate or trust account, beneficiary designations, TOD/POD instructions, and the letters testamentary. Portability elections ride on a timely Form 706; don’t miss that window.
• Brokerage mechanics: confirm beneficiary vs. estate titling, reconcile all 1099s, and for IRAs track 5498/1099-R coding. Small admin errors now become April tax-season headaches. Ask me how I know.

Quick checklist: 1) Lock down date-of-death values; 2) Update cost basis and lot IDs before any sells; 3) Validate FDIC/SIPC coverage across banks/brokers; 4) Park cash in T-bills or gov MMFs; 5) Map the inherited IRA 10-year schedule; 6) Clear estate-tax decisions with counsel before retitling.

Circling back to the first point, I said slow the process, not the money. This is what I meant. Move the cash into safe wrappers immediately, then take the time to get basis, titles, and beneficiary rules correct. I know I haven’t talked about the tax-location of AI bets yet, we’ll hit that shortly, but none of that matters if you torch the step-up or miss an inherited-IRA deadline. Boring wins here. Every single time.

A 3-bucket plan that steadies nerves in AI-led volatility

Markets are twitchy this year. AI winners can be up 6% one week and give half back by Thursday afternoon. That’s fine for a trading account; it’s not fine when you’ve got tuition, caregiving, or a mortgage draw relying on that cash. The cleanest fix I’ve used for families inheriting money is a 3-bucket structure that separates spending from growth and from the “fun stuff” (your AI sleeve) so one rough headline doesn’t derail life goals.

• Bucket 1: 12-24 months of spending needs in T‑bills/short Treasuries. Calculate your expected withdrawals, taxes, living costs, planned gifts, big purchases, and hold 1-2 years of that in 3-12 month U.S. Treasuries or government money market funds. Yields on 3-6 month bills hovered around 4.5%-5.5% across 2024 and early 2025, which still pays to wait compared with idle checking. This gives you time when AI names whipsaw. If your spending is lumpy (quarterly tuition, annual insurance), ladder bills to those dates. I know it’s boring. That’s the point.
• Bucket 2: Diversified core sized to your risk capacity and time horizon. This is the workhorse: global equities, investment‑grade bonds (mix of intermediate and some short), and a measured slice of alternatives (REITs, managed futures, or private credit if you’re eligible). Use realistic guardrails: a 50/50 or 60/40 type profile for most multi‑goal households, not because it’s cute but because the long‑run U.S. 60/40 mix earned around 7%-8% annualized over many decades (varies by period), and it rebalances itself during shocks. If you’ll tap principal inside 5-7 years, bias to quality bonds and cash‑flowing equities. Rebalance on bands (say, ±5%) or semiannually to keep AI rallies from silently blowing up your risk.
• Bucket 3: Opportunity sleeve (AI). Cap this at 10%-15% of the total portfolio, hard cap, written down. Use diversified tools first: broad AI or semiconductor ETFs, data‑center and power infrastructure, and maybe a measured position in enablers (networking, memory, design software). Keep single‑name shots small. Expect sharp swings; 10% pullbacks in equities happen about once a year on average, and AI‑heavy pockets can do that in a month. Refill this sleeve from gains, not from Bucket 1. When it runs hot, trim back to the cap. Non‑negotiable.

How to fund it without guessing the top: dollar‑cost average the lump sum over 6-18 months. Put a base amount in monthly, and give yourself permission to speed up the pace when volatility spikes, e.g., after a 3-5% market drop in a week or on days when the volatility index jumps meaningfully. You’re turning chaos into a funding schedule instead of a stress event. Earlier this year I had a client accelerate two tranches during a semiconductor selloff; three months later the core was back on plan, and their cash bucket never moved.

Simple rules: 1) 12-24 months of spending in Treasuries; 2) Core portfolio sized to your withdrawal horizon and sleep-at-night risk; 3) AI sleeve capped at 10-15% with automatic trims; 4) DCA over 6-18 months, faster when volatility pays you.

Two quick notes I’ll come back to: we haven’t talked about ETF vs. single‑name selection for the AI sleeve (there’s a tax wrinkle if you’re using harvested losses), and there’s a tax‑location angle if you’re splitting accounts across taxable, IRA, and Roth. But the sequence holds: lock down Bucket 1, set the core weights, then layer the opportunity sleeve slowly. It keeps the inheritance aligned to goals while still letting you participate if AI keeps surprising us into December.

Position-sizing AI exposure without betting the house

Own the trend, not the blowups. I like clean, boring rules because they save you from your own enthusiasm when headlines get loud. If you’re investing an inheritance during AI market volatility, the sizing math is the guardrail that keeps your long-term plan intact even if one chip name or a model-provider misses a quarter and gaps down 18% overnight. It happens. It will happen again.

• Single-stock risk budget: 3-5% max per name. If you’re newer to concentrated positions, stay closer to 3%. Veterans who actually monitor positions daily can flirt with 5%, but no heroes. I cap any one AI leader there, even the ones we all love, because earnings dispersion is high this year and headline risk is constant.
• Total AI sleeve: 10-15% of the portfolio, unless you’re truly very high risk. That cap is your seatbelt. It forces trims when winners run and frees cash for adds when volatility pays you. We already set 10-15% earlier, same number here, on purpose.
• Prefer diversified vehicles: Use broad AI/semiconductor/platform ETFs, or equal-weight sector funds, to reduce single-name blowups. A basket of fabs, equipment, hyperscalers, and software inference spreads the risk of any single guidance cut. I’ll use equal-weight when one or two giants are doing all the heavy lifting and I don’t want my fate tied to a single earnings call.
• Set valuation rails: Pre-define a forward P/E band for adds and a “no-add” zone when multiples stretch. Example: add below your target forward P/E band, pause adds if the sector trades two turns above your top line. You can flex it for growth rates, but write it down. Future-you will thank present-you.

History is the reminder. In 2023, the S&P 500 equal-weight index lagged the cap-weighted index by about 12 percentage points, which shows how concentrated leadership can skew returns. Design your mix on purpose, not by accident. And this year, leadership is still narrow on many up days, breadth improves, then fades, so the lesson still applies. I think that 12-point gap was ~12.3%, don’t quote me; the point stands either way.

How I implement this in real accounts, warts and all:

1. Define the sleeve and the trims: Example: 12% AI sleeve with automatic trims back to 12% if it breaches 14%. That enforces sell discipline without drama.
2. Mix vehicles: 60-70% in diversified ETFs (one cap-weight, one equal-weight, maybe an equipment-focused fund), 30-40% in 2-4 single names capped at 3-4% each. If you’re using tax-loss harvesting, ETFs keep things cleaner on wash-sale rules versus rotating between look‑alike single names.
3. Valuation rails: Tie adds to forward P/E or EV/sales bands informed by your growth assumptions. Example, “add tranches when forward P/E drifts back into the 20s; pause if we’re north of low‑30s without estimate upgrades.” You can adjust per sub‑industry, semicap vs. hyperscalers isn’t apples to apples.
4. Stagger entries: DCA over 6-18 months. Speed up on volatility spikes. You won’t pick the low. You don’t need to.

Personal note: I was on the desk the day a major AI supplier cut backlog commentary and the stock fell ~15% before lunch. The clients who were fine had 3-4% positions and ETF ballast. The ones calling me every ten minutes? 9-12% single-name. Same story every cycle, different ticker.

Gray areas? Always. Sometimes an ETF gets crowded, or a single name is genuinely mispriced after a messy quarter. Fine, bend, don’t break. Keep the 3-5% per-name and 10-15% sleeve as your non-negotiables, use ETFs or equal-weight to manage concentration, and let valuation rails tell you when to add or just… wait.

Risk controls that actually work in 2025 markets

Volatility isn’t a problem; unmanaged volatility is. Forecasts age like milk, rules age like steel. So build rules that are boring, repeatable, and tax-aware. Especially if you’re sitting on fresh cash from an event, inheritance, business sale, whatever, and typing “how-to-invest-inheritance-during-ai-market-volatility” at 1 a.m. The system matters more than the headline du jour.

• Rebalance with bands, not the calendar: Set guardrails like ±20% of target weight. Example: a 10% target position gets reviewed at 8% or 12%. That nudges you to sell some winners and add to laggards without overtrading. It’s automatic mean reversion. Calendar rebalancing misses the point because markets don’t move on your schedule; they move on their own, often rudely.
• Tax-loss harvest opportunistically: Losses show up fast in choppy tape. The wash-sale rule still applies, 30 days before/after, so use close substitutes, not identical ones. Swap an S&P 500 fund into a Russell 1000 or total-market fund, or rotate a sector ETF into a similar but not identical factor tilt. Keep the beta close, keep the loss, and keep your sanity. And yes, Treasury ETFs typically pass through state tax exemption on interest, handy at higher brackets.
• Diversify factors on purpose: Mix quality, value, and small/mid caps so you’re not hostage to a handful of mega-cap AI names. Last year (2024), the top 10 S&P 500 names were roughly one-third of the index by weight (S&P Dow Jones Indices data, late 2024). That concentration cuts both ways. Quality and value held their own in the 2022 drawdown when high-duration growth cracked, and small/mid caps can catch up fast when rates stabilize.
• Hold cash intentionally, not accidentally: Idle checking pays near-zero. Short T-bills and government money funds have been yielding around the high-4% range in Q3 2025 (U.S. Treasury and fund company data), after sitting near ~5% earlier this year. Park unallocated cash there while you stage entries. Bonus: Treasuries are state tax-exempt; treasury-only funds usually keep that benefit.
• Stress test the plan: Model a 25-35% equity drawdown. Not hypothetical, 2020 saw ~34% peak-to-trough in weeks, and 2022 clipped roughly a quarter off the S&P 500 at one point. Can you still fund 3-5 years of planned withdrawals? If not, adjust the mix now, not during the next air pocket.

Quick reality check, are these exciting? Not really. Do they work? Yes. I’ve watched clients who stuck to bands and harvested losses quietly beat the “smart takes” by just… doing the maintenance. One more nuance: if you’re dollar-cost-averaging an inheritance into AI-adjacent exposure, you can pair entries with band-based trims elsewhere. It turns a single tactical bet into a portfolio-level risk trade that nets out cleaner.

Also, a conversational note because this comes up on calls: “What if I miss the next leg higher?” You will, sometimes. Everyone does. But rules keep you in the game. Rebalancing sold some NVDA on the way up? Fine. It also made you buy quality and small caps when nobody wanted them. Cash in a T-bill ladder at ~4.7-5.0% this year isn’t a mistake; it’s optionality you’re being paid for.

Bottom line: automate the boring parts, bands, harvests, factor mix, cash home, and a real stress test. Markets are loud this year; your process should be quiet.

A simple playbook to invest an inheritance this year

Here’s how I’d turn a noisy 2025 into an actionable 3-12 month plan you can actually execute while headlines swing between AI euphoria and AI panic. I’ve used this with real families, and it’s boring in the best way.

1. Week 1-2: Clean up the pipes
   • Consolidate accounts: Get the assets to one or two custodians tops. Old 401(k)? Beneficiary IRA? Taxable? Fewer logins = fewer mistakes.
   • Verify cost basis and holding periods: Ask the custodian to load historical basis for each lot. Missing basis creates tax messes later. Quick note: for inherited taxable assets with a step-up, confirm the date-of-death value is recorded, don’t assume.
   • Set beneficiaries and TODs: Primary and contingent. It takes 10 minutes and solves 10 headaches.
   • Park funds in cash equivalents: Use T-bills or government money market funds while you decide. Short Treasuries are still yielding about 4.7-5.0% in 2025, which is real carry while you get organized. Money market fund assets hit a record ~$6.1 trillion in 2024 (ICI), and for good reason: it’s yield + liquidity.
2. Week 3-4: Write your one-page IPS
   • Draft the IPS: Goals, time horizons, withdrawal policy, and what you’ll buy/sell when. One page, not a novel.
   • Pick target allocations: Example only: 60/35/5 (stocks/bonds/cash) or 70/25/5 if your runway is long. If you want an AI sleeve, set a hard limit now, say 5-10% of equities, so enthusiasm doesn’t quietly balloon to 20%.
   • Define rebalancing bands: ±20% of each sleeve’s target weight (i.e., a 10% target rebalances at 8%/12%). Bands keep you honest when Nvidia has a +12% day, it’s happened multiple times around earnings since 2023, and when small caps are out of favor.
3. Months 2-8: Get invested on a schedule
   • Dollar-cost average: Break the inheritance into, say, 6-8 tranches. Buy on the same day each month, then add an extra tranche after large down days (−2% to −3% on the S&P 500) to lean into volatility, not away from it.
   • Harvest losses in taxable if a sleeve drops below your band. Pair with similar ETFs to maintain exposure. This sounds tiny but adds up, tax alpha at around 1% annually isn’t crazy over cycles.
4. Automate the mechanics
   • Set transfers and auto-buys: Monthly ACH from your bank to brokerage, auto-purchase your chosen ETFs the same day. No decision tree at 9:31am.
   • Rebalancing alerts: Most custodians let you flag band breaches. Act when alerts hit, not when financial TV gets loud.
   • Rules for the AI sleeve: Rebalance back to the cap when it exceeds the limit; add only on your pre-set DCA dates. No exceptions, even on “can’t-miss” Mondays.
5. Unspent cash earns while you phase in
   • Ladder 3-12 month Treasuries for known near-term uses (taxes, home projects, tuition). Roll maturities into your DCA plan or into new bills. 6-month T-bill yields were near 5% for much of 2024 and remain attractive this year, even if they’re wobbling around the high-4s.

Small note from 20+ years of scars: automate where you feel the most temptation. The calendar beats your gut, your gut is great for lunch picks, not for entries.

And if you’re worried about “missing the next leg,” remember: bands make you trim highs and add at lows; DCA makes you show up; the ladder pays you to wait. Same idea, said slightly differently, the process lets you be patient and still get invested.

If you wait, the market won’t, here’s what that costs

Quick reality check: doing nothing is a decision, and it’s rarely free. It looks safe, no trades, no taxes, no second-guessing, but the meter’s running. Three places it shows up fast: cash drag, missing tax windows, and letting concentration risk decide your future allocation for you.

Cash drag compounds. In 2023 and through parts of 2024, 3-6 month T-bill and top money market yields were over 5% at times. Six-month T-bills traded around 5.5% in late 2023 and hovered near 5% for much of 2024. Meanwhile, a lot of checking accounts paid close to 0.01%-0.10%. That gap isn’t theoretical. On $500,000, the difference between 5.0% and 0.05% is about $24,750 per year. Even this year, with yields wobbling in the high-4s, you’re still getting paid to be organized. The simple, slightly boring ladder you set up earlier this year is doing more than your idle cash will ever do. And, yes, I’ve watched clients miss this for months because “rates might drop next meeting.” Maybe. But you’re paid today, not in hypothetical rate paths.

Tax windows close. Two big ones I keep seeing:

• Inherited IRAs: The SECURE Act (effective 2020) put most non-spouse beneficiaries on a 10-year clock. That clock starts in the year after death, waiting doesn’t pause it. If the decedent had already begun RMDs, you may also need annual distributions inside the 10-year window. Drag your feet for seven years and you compress taxes into the last three, usually the worst outcome.
• Step-up in basis: Taxable assets generally reset basis at death. Selling sooner can realize gains with little or no tax if values haven’t drifted much post-step-up. Wait two years into a hot tape and you may turn a near-zero gain into a five-figure bill. I’ve seen heirs sell a concentrated single stock within months of the step-up and pay a few hundred bucks; the sibling who hesitated paid five figures later after a 30% rally. Same asset, different timing.

Concentration risk grows while you delay. The index isn’t shy about crowding. In 2024, the top 10 names were roughly a third of the S&P 500’s weight, and that concentration has stayed elevated into this year. If your inheritance is heavy in one or two winners riding the AI narrative, waiting often tightens the coil, your risk rises while your plan sits in drafts. Set rules now: max position sizes, sector caps, and rebalancing bands. That way the next AI swing doesn’t set your plan for you.

Action beats perfection. You don’t need a flawless entry, you need a written plan you’ll actually follow. Three moves that work in the real world:

• Write it down: target allocation, cash needs, band widths (say ±20% around position targets), and your DCA dates. Keep it on one page. If it’s longer, you won’t read it, trust me.
• Buy in tranches: commit a schedule (weekly or monthly) and size (e.g., 10% of the target each clip). It feels slow, good. Slow is how you sidestep headline anxiety.
• Rebalance to bands: trim when a holding breaches the upper band; add when it hits the lower. No hero calls, just rules. Bands quietly force you to sell high and buy low, without speeches.

One personal observation from too many Monday opens: when people say they’re “waiting for clarity,” they usually mean they’re waiting for prices to be higher so it feels safer. That’s not clarity, that’s comfort. The market doesn’t wait for comfort. Capture yield while you phase in, use the tax calendar before it uses you, and lock in your allocation rules while you’re calm. It’s not hyped urgency, it’s just arithmetic, every month you defer, the easy wins (interest income, tax efficiency, risk control) slip a bit further out of reach.

Frequently Asked Questions

Q: How do I park an inheritance for 60-90 days without losing ground?

A: Use insured high‑yield savings for instant access and 3-6 month Treasury bills for yield. Set a calendar reminder and don’t touch the buy button until it pings. Earlier this year, T‑bills yielded near 5% annualized, plenty of rent while you write your one‑page plan.

Q: Is it better to invest the lump sum now or dollar‑cost average during this AI frenzy?

A: Mathematically, lump sums tend to win. Vanguard’s 2012 study showed lump‑sum investing beat DCA about two‑thirds of the time across the U.S., U.K., and Australia. Behavior is the catch. Morningstar’s 2024 Mind the Gap showed investors trailed their own funds by ~1.7%/yr because of mistimed moves. My compromise: after a 30-90 day cool‑off in cash/T‑bills, drop 50-70% in immediately per your policy, then schedule the rest in equal tranches over 4-6 months on fixed dates. No headline‑based overrides. Pre‑set rebalancing bands (around 7%) so you trim if equities run after an AI news burst. This keeps the math on your side while reducing the odds you panic‑sell the first ugly day.

Q: What’s the difference between a cash bucket and an emergency fund when I’m investing an inheritance?

A: Emergency fund = you/household safety net, usually 3-6 months of core expenses in insured cash. It’s for job loss, medical deductibles, broken furnace, life stuff. Cash bucket = a portfolio tool. For heirs relying on the portfolio, hold 12-24 months of planned withdrawals in 3-6 month T‑bills or a Treasury ladder. The cash bucket dampens sequence‑of‑returns risk so you’re not forced to sell stocks after a 20% drawdown. Keep the emergency fund separate, super‑liquid, and mentally untouchable. The cash bucket sits inside the investment plan and refills during rebalances after good markets.

Q: Should I worry about overconcentration in AI stocks if I already own a tech‑heavy index?

A: Short answer: yes, crowding risk is real. A lot of “AI exposure” is already embedded in broad U.S. indexes through mega‑cap tech. Layering a thematic AI ETF on top can quietly push your sector and single‑name weights beyond what you’d ever approve if you saw them on one page. Here’s how I cap it:

• Write guardrails first: max equity drawdown 20%, sector cap ~30% of equities, single‑stock exposure <5% of total portfolio. Rebalance bands around 7%.
• Use a core‑satellite design. Example on a $1,000,000 inheritance after a 60‑day cool‑off: 60% equities / 35% bonds / 5% cash. Within equities: 30% broad U.S., 15% international developed, 10% U.S. small/value tilt, 5% “satellite themes” (if you must scratch the AI itch). That 5% of equities is 3% of the total portfolio, spicy, not reckless.
• Own AI via diversified cores first (broad market and semis within them), then keep any pure AI thematic fund small and rule‑based. If AI rips and breaches your band, trim back to target; if it slumps, you rebalance up, behavior rules, not headlines. I had a client call me from the lawyer’s parking lot earlier this year ready to throw 40% into an AI basket. We parked the cash in T‑bills near 5%, wrote the policy, and settled on a 3% thematic cap. Two rough weeks later, they were very happy they had bands and a plan.

@article{how-to-invest-an-inheritance-in-ai-driven-volatility,
    title   = {How to Invest an Inheritance in AI-Driven Volatility},
    author  = {Beeri Sparks},
    year    = {2025},
    journal = {Bankpointe},
    url     = {https://bankpointe.com/articles/invest-inheritance-ai-volatility/}
}