Where things stand, and what keeping Microsoft actually looks like

Start here

What's actually held, where

"OWO" and "VUE" are typos for two ordinary index funds

The IRA runs a target-allocation model, and the targets add up to exactly 100% — so each unknown ticker occupies a specific slot in a deliberate design. That, plus the price per share implied by the file, identifies both with high confidence. Still worth a two-second confirmation on the custodian's website.

Both are plain, low-cost, broadly diversified funds — not stale or mystery holdings. And VEU at 13% means the IRA has meaningful international exposure, which is one more thing pulling in a different direction from US mega-cap tech.

Because this IRA is inherited: one thing for a CPA to confirm is the required-distribution schedule. Most non-spouse inheritances after 2019 must be fully drawn down within 10 years; spousal inheritances follow different rules. If this account is on a forced drawdown clock, that's a reason for it to carry a bit less equity risk than the model currently holds — a planning question, not an emergency.

Keeping Microsoft — what that actually looks like

Twenty years is a long time to hold something. Not selling is a reasonable position, and there are concrete financial reasons — not just sentiment — that it may be the right one.

Why not selling can be the smarter move

Microsoft traded around $25–30 in 2006. At ~$511 today, roughly 95% of the position's value is unrealized gain — on the order of $145,000 of the $153,000. Whether that matters depends entirely on one fact we don't know yet: what kind of account is the third account?

• If it's a regular (taxable) brokerage account: selling the whole position would mean a capital-gains tax bill in the rough neighborhood of $20,000–$35,000 depending on bracket and state. And there's a second, bigger point that advisors routinely make to older investors holding a decades-old winner: assets held until death pass to heirs with a "stepped-up" cost basis — the entire gain is never taxed by anyone. That is one of the strongest arguments in personal finance for hedging a position you're nervous about rather than selling it. A CPA should confirm how it applies here.
• If it's an IRA or 401(k): none of that applies. Sales inside the account trigger no tax, and there's no step-up. Then trimming is easy and the only reason not to is sentiment — which is still a real reason, but a different conversation.

The honest version of the Microsoft-specific risk

Microsoft is not a speculative AI startup. It's roughly a $280 billion-revenue company generating on the order of $70 billion a year in free cash flow, with more cash than debt, whose core businesses (Office/Microsoft 365, Windows, the non-AI parts of Azure, LinkedIn, gaming) would keep generating cash even if the AI build-out turns out to be overdone. In an AI bust, Microsoft's risk is "they overspent on data centers and the stock re-rates lower for a while" — not "the company fails."

The honest historical worst case is worth stating plainly: after the 2000 bubble, Microsoft fell about 60% from peak to trough and didn't get back above its 1999 high until 2016 — roughly 14 years — even though the business grew the entire time. That's the realistic bad scenario for a mega-cap leader after a bubble: a deep drawdown and a long flat stretch, not a wipeout. For him, a 50% Microsoft drawdown would be about −$77,000, or ~10% of the total portfolio. Painful. Not catastrophic.

The whole-portfolio stress test

A deliberately harsh "AI unwind" scenario — Microsoft −50%, the growth ETF −40%, the S&P fund −35%, Cisco/HP −30%, and everything else flat:

For scale: an ordinary 60/40 stock/bond portfolio lost about 16–17% in 2022 and 25–30% in 2008. So his severe AI-crash scenario costs roughly a normal bad year — because three-quarters of the money isn't in the trade. And that's before any offset from the dividend/value/energy names, which historically did well in exactly that kind of rotation. This is the number that should ease the fear: the worst plausible outcome is survivable, and it's already mostly diversified away.

Things that reduce the risk without selling Microsoft — ranked by how painless they are

1. Rebalance inside the IRA.no tax · no complexity
   IUSG ($37,844) is the most AI-concentrated thing he owns after Microsoft itself — roughly 37–40% of it is Nvidia, Microsoft, Apple, Alphabet, and Broadcom. Moving some of it into VTV (value — already his largest IRA holding) or VEU (international) cuts implied AI exposure by up to ~$15,000, and because it's inside an IRA there's zero tax consequence. Even taking IUSG's target from 14% to 7% and putting the difference in VTV/VEU is a meaningful, painless de-risk. The easiest single move on this list.
2. Put the idle $33,595 at Merrill to work defensively.no tax
   It's sitting in a bank-deposit sweep. Short or intermediate Treasuries, or a short-term bond fund, would build the defensive cushion from ~14% to ~18% of the portfolio without touching a single stock. (TIPS are the specific instrument if his worry is "debt-financed capex eventually means inflation.")
3. Stop reinvesting into the AI-heavy sleeve.no tax
   If dividends or contributions auto-reinvest into Microsoft or IUSG, redirect them to bonds, value, or international. Concentration drifts down over time without ever selling anything.
4. Trim a slice, not the position.small tax if taxable
   Selling 20% of Microsoft (60 shares, ~$31,000) drops it from ~19% of the portfolio to ~15%. In a taxable account that realizes roughly $29,000 of gain — maybe $4,000–7,000 in tax; in an IRA, nothing. A middle path that respects twenty years of holding.
5. Protective puts on Microsoft — the actual "insurance."annual premium
   300 shares = 3 put contracts. Buy puts struck ~15–20% below today's price, 12–18 months out (a "LEAPS" put just means the long-dated kind, so it doesn't need re-buying every month). If Microsoft crashes, the puts pay off roughly the drop below the strike, offsetting most of the loss. If it doesn't crash, the premium is gone — same as any insurance. Rough cost for long-dated out-of-the-money puts on a mega-cap: very roughly 3–6% of the position's value per year, so ~$5,000–9,000 a year on $153,000 — only a live quote is a real number. Puts never force a sale of the shares.
6. A collar — cheaper insurance, with a catch.near-zero premium, but see catch
   Sell a call ~15–20% above today's price to pay for the put; net cost can be near zero. The catch for someone avoiding a taxable sale: if Microsoft rallies above the call strike, the shares can be called away — a forced sale, which is exactly the tax event he's trying to avoid. It's manageable (roll or buy back the call before expiration), but collars need an advisor who will actually watch them, and they cap the upside for the collar period.
7. Hedge the market instead of the stock.annual premium, usually cheaper
   SPY puts are the most liquid and cheapest per dollar of protection. One long-dated SPY put position hedges the scenario he actually fears — "the AI unwind drags the whole market down" — across Microsoft, IUSG, and SPY at once. Imperfect (SPY won't fall exactly like Microsoft), but simpler and often cheaper than stock-specific puts.

On timing the crash

Nobody can. Alan Greenspan's "irrational exuberance" warning came in December 1996; the Nasdaq more than doubled after that before it finally topped in March 2000. People who sold on the warning missed three years of gains and still had to guess when to get back in. The whole point of diversifying and hedging is that you don't have to predict — you just make sure the bad scenario is survivable, which, per the table above, it already is.

The AI-capex concentration

IUSG carries roughly 37–40% of its weight in Nvidia, Microsoft, Apple, Alphabet, and Broadcom; SPY about 28–30%. Cisco and HP are networking/hardware names with some AI-buildout narrative attached, but they're not the same bet as hyperscaler capex — included for completeness, not as equivalent risk.

Why the worry has substance right now: current commentary (Jefferies' Chris Wood, among others) flags that hyperscaler AI capex is increasingly debt-financed — six major players are on pace for roughly $916 billion of combined capex over the next 12 months, rising toward $1.2 trillion the year after; Amazon, Alphabet, and Microsoft are each spending over 100% of their cloud revenue on capex this year; and capex-to-revenue has crossed ~22% against a historical 11–16% norm, a pattern some analysts compare to the run-up before the 2001 telecom bust. None of that makes a crash certain or imminent. It does mean the concern is real and widely shared — and it lines up with exactly the exposure this portfolio carries.

The biotech names, one by one

Tax-loss harvesting on EDIT/SPRO is a legitimate small housekeeping item even though the dollar risk from here is trivial — realizing the loss can offset gains elsewhere, and there's plenty of embedded gain in the same Merrill account to offset. A CPA question, not an investment one.

Open questions and caveats

• What kind of account is the third account (the Microsoft one)? Taxable brokerage vs. IRA/401(k) changes the Microsoft answer more than anything else in this document.
• OWO → IWO and VUE → VEU are high-confidence identifications from price and allocation logic, but a glance at the holding's full name on the custodian's site settles it for certain.
• Inherited IRA: confirm the required-distribution timeline with a CPA (10-year rule vs. spousal rules).
• Third-account position sizes are estimates, back-calculated from rounded day's-change figures. Microsoft, MIR, TTI, AIRG, RVMDW, MS, and RVMD divided out to clean, high-confidence numbers; GERN, NAK, IOVA, HPQ, CSCO, and LUV are rougher. None are positions where the estimate error changes the conclusions.
• The three files are three different snapshot dates (09/03, 09/17, 09/25) — a composite picture, not one moment. The third-account file shows no cash balance, so its true total is probably somewhat above the ~$219K estimate.
• Option costs above are ballpark ranges from typical pricing on mega-cap stocks — only a live quote from the broker is a real number.