Ask a founder with a company worth $10m-plus how diversified they are, and they'll usually point at everything except the company: a pension, an index fund in a GIA, some cash earning a decent rate. The mental model is two buckets: the risky one they run, and the safe one compounding in the background. As long as the second bucket exists, diversification feels handled.
I see this constantly in wealth management, and not only from founders: advisers count the liquid portfolio as diversified by default because it holds hundreds of names. Almost nobody asks what those names are. Concentration risk gets checked on the business and ignored everywhere else. And the dangerous concentration isn't the stake you can see. It's the "safe" money sitting on the same bet as the business.
This Week in 30 Seconds
- Your "diversified" fund is loaded on your sector. Top 10 stocks are about 44% of the S&P 500, and the seven biggest names, all tech, are about a third of it. A tech founder holding the index is stacking the same exposure on top of the business.
- A private stake can't be hedged directly. Exchange funds, 351 conversions, collars and forwards all need publicly traded stock. Pre-exit, the lever you control is the liquid side.
- Two moves on the liquid side, both with costs. Sector options for founders who know them, with basis risk left over, or real ballast: safe, uncorrelated assets that give up returns while you wait. The balance depends on how soon the business turns liquid.
- On the Radar. 351 conversions, a $250bn secondaries year with wide discounts, "next NVIDIA" funds, and Beijing tightening capital exits.
Where the concentration risk lives
Start with the number everyone already worries about. Long Angle surveyed 233 investors, averaging $17m in net worth, and found that founders hold about 61% of their private and alternative portfolios in their own company. Employees with equity run 67% in their employer's stock. At this tier, you mostly are the portfolio.
That stake is also the part you can't move. A sale planned for 2027 slides to 2029 because a buyer withdraws or the market closes. Bessemer Trust suggests keeping any single company under 4-5% of net worth, and its list of single-stock wipeouts stretches from Kodak and Lehman through SVB and Signature in 2023 to Spirit Airlines and 23andMe in 2025. Most founders hold ten times that, with an exit date they don't control.
So far, so familiar: this is the risk every adviser already names. What gets missed is the second layer.
An index fund that doubles the bet
Concentration has three dimensions. How much sits in one company is only the first; the other two are sector and country, and that's where the safe bucket lets you down.
The S&P 500 is the default diversification instrument for most founders in this range. As of July 2026, the top 10 stocks make up about 44% of it, the highest share on record. NVIDIA alone accounts for around 7%, and the seven biggest names, all tech, add up to about a third of the index.
Run a software company and hold that index, and about a third of every "safe" dollar is in big tech, the same cycle your business lives in. The scenario to take seriously: AI could hurt your company's business model while the AI names are a third of your index fund. One event, both buckets. I made a version of this point in March, in When Everything Correlates, after one geopolitical weekend moved every asset class together. Sector overlap is the same problem, except it never goes away.
Country works the same way: business, pension, index money, and home often share a single jurisdiction, which is a concentration in its own right.
Why the stake itself can't be hedged
The obvious answer is to hedge the business itself. It doesn't work because the tools bankers use for concentrated positions (exchange funds, the newer 351 ETF conversions, collars, and prepaid forwards) all require publicly traded stock. Cache, one of the firms building modern exchange funds, lists the standard terms: a 7-year lock-up and a qualified-purchaser minimum of $5m. A private, pre-exit stake can't go in at all. Until there's a ticker, you can't hedge the stake.
So the only lever a pre-exit founder controls is the liquid side: the pension, the index money, the cash. That part can be repositioned this month.
Two honest moves on the liquid side
Founders who know options well sometimes deal with the overlap head-on. They offset the correlated exposure with options on a broad or technology index, paying a premium so that one AI shock can't hit both buckets at once. It works up to a point, and J.P. Morgan's guide to concentrated positions spells out why: a proxy hedge never fully removes the risk, because the proxy doesn't trade like your company. An index can hold its level while your business halves, and the premium is spent either way. A sector option hedges the sector. The company risk stays yours.
The second move is simpler. Once the overlap is visible, the liquid side stops trying to perform and does one job: staying uncorrelated with the business in very safe instruments, such as bonds or cash. This is the barbell from Barbell Wealth, taken one step further: the business already carries the equity upside, the illiquidity premium and the sector risk, so the other end only works if it's truly safe. For a tech founder, an index fund with a third of it in seven tech names doesn't qualify.
Before you add anything new
Neither move is free. The proxy hedge leaves basis risk and burns premium on protection that can miss the actual failure. Sitting fully in cash and bonds has a cost too: the index got this concentrated because those bets kept paying off (NVIDIA delivered its returns through two-thirds drawdowns, per Bessemer), and a founder who waits in cash for an exit that keeps sliding gives up years of returns. Bonds add rate and inflation risk of their own.
Morgan Housel put it best: getting wealthy takes optimism and risk-taking, staying wealthy takes humility and fear. A founder mid-build has to run both at once. How much stays safe and how much stays invested is a judgment call, and it hangs on one question: how liquid is the business really, and how soon.
One thing costs nothing, though. Take the whole balance sheet (business at a realistic valuation, pension, index funds, cash, property) and map what each line holds by sector and by country. Most founders who do this find the two buckets were never as separate as the statements made them look. Map the overlap before adding a single new position.
Everything here is free. Subscribing just tells me the content is useful — and helps me decide what to write next.
From where I sit
From inside the industry that's meant to catch this, I watch the same gap slip past. An adviser looks at the pension, the funds, and the cash and calls it diversified because, on its own, it is. The business is never on the statement in front of them, so the one position that dwarfs everything else stays outside the review. The question that would change the picture rarely gets asked: how much of your safe money is the same bet as your business?
The mapping takes one page. Every line of the balance sheet down the left: the business at what it would fetch today rather than the last term-sheet number, the pension, each fund inside it, the GIA, cash, property. Two columns on the right: sector and country. The pension line is where it gets interesting, because a default UK pension fund runs heavy on global equities, and global equities now run heavy on US technology. A founder can be "diversified" across nine accounts and still find that seven of them lean on the same handful of companies.
Then the two questions that matter. If the sector the business lives in repriced by half, what else on the page moves with it? If the country changed the rules (capital controls, taxes, exit frictions), how much of the page would be affected by that one system? I grew up watching the second question become less hypothetical, which is probably why I weigh it more heavily than most advisers would.
I'm not going to tell you what the right split is; that depends on things no newsletter can see. The map is the part I'd push for. One evening of work, and every decision after it gets sharper.
On the Radar
351 conversions solve concentration only after your stock is public.
Bernstein and Kitces have made the 351 ETF conversion the fashionable 2026 alternative to the old exchange fund: hand a concentrated marketable position into a new ETF, defer the gain, walk out diversified. It does nothing for a private stake, the diversification tests have to be pre-engineered, and you end up in a fund you don't steer. Worth knowing cold before a banker frames it as your concentration fix. Read more →
Secondaries are heading for $250bn, and discounts are wide again.
EquityZen's Q2 read has secondaries on track for roughly $250bn this year, with transaction count up about 20% on the quarter. The average discount widened to around 38% from about 8% in Q1, and EquityZen's own explanation is mixed: the recently listed names that tightened Q1 pricing have left the market, and older mega-unicorns now dominate the volume. For a stake with no ticker, a secondary is the closest thing to a hedge. A reason to read your company's tender policy while the market is this deep. Read more →
"Next NVIDIA" funds can rebuild the concentration you sold.
Jan Voss at Cape May Wealth looks at the AI-infrastructure and space funds now pitched to anyone who's sold a business. His line hasn't moved: he's no better placed than anyone to pick winners inside a theme, and the useful question is what concentration of AI a portfolio ends up carrying. Rolling one stake into a cluster of AI-adjacent names can feel like diversifying while quietly rebuilding the same single-bet risk in a new wrapper. Read more →
Beijing tightens the exits, and Hong Kong feels it first.
Mr Family Office reports Beijing's clampdown on cross-border capital flows already rippling through Hong Kong, now the world's largest offshore wealth centre. Jurisdiction concentration fails the same way a single stock does. A reader whose operating stake, custody and residency all sit inside one system is running a concentrated position, whatever the asset mix says. This week's reminder to map country the way you map sector. Read more →
New on the Site
Last week's piece looked at what AI changes in practice about running your own money at $5M–$100M: which family-office functions a founder can now run from a laptop, and which were never about software in the first place. If this memo has you mapping exposures, that piece covers the tooling half of the same job.
Read it: What AI Actually Changes About Running a Family Office at $5M–$100M
This was Capital Signals — weekly briefings on what's reshaping founder strategy on wealth.
Go deeper: Playbooks · Wealth Architecture · Investment Strategy · Business Building · Life Design
Found this useful? Forward it to a founder who's thinking about this stuff. Got a question or disagree with something? Get in touch.
New here? Subscribe for one email a week.
