Portfolio
This is a sample client profile, illustrating how a hypothetical retiree might structure $2,000,000 across four spending “buckets” based on when the money will actually be needed. The numbers, timelines, and account types below are intentionally specific, since vague illustrations are hard to learn from, but the person behind them is not real. Nothing here is personalized financial, tax, retirement, or legal advice, and nothing in it should be applied directly to your own situation without working through it with a qualified, licensed advisor who knows your actual circumstances.
The core idea is simple, even though the details get complicated: money you’ll need soon should be kept completely safe, even if that means earning less on it. Money you won’t need for many years can afford to take more risk, because it has time to recover if it dips.
Total amount in this sample profile:** $2,000,000, with spending modeled at 4% ($80,000/year).
Disclaimer: This is a hypothetical, illustrative example built for educational purposes only. It does not represent advice for any actual person, including the author, and should not be relied upon as personalized financial, investment, tax, legal, or retirement planning advice. Tax rules (including MAGI, ACA subsidies, IRMAA, and RMDs), account contribution limits, and product availability change over time and vary by individual circumstances. Past performance and current rates referenced are not guarantees of future results. Consult a licensed financial advisor, tax professional, and/or attorney before making any financial decisions.*
The Big Picture: How the $2,000,000 Is Split Up
Think of this as one large pie cut into different-sized slices, based on safety versus growth:
48% in stocks (28% U.S. companies, 20% international companies)
15% in bonds (a type of loan to governments or companies that pays steady interest)
17.5% in “alternative” assets (12.5% gold and metals, 5% in raw materials like copper and oil)
19.5% in cash and cash-like savings (19% in short-term government savings, 0.5% in a regular checking account)
The bigger idea: about half the money is in stocks (which can grow a lot but also bounce around), while the rest is split between safer options. The exact ratio matters less than the underlying logic, which the four buckets below explain.
Bucket 1: Year 1 Spending Money
Amount: $100,000, almost entirely safe, easy-to-access cash
In this sample profile, this bucket covers spending for Year 1 of retirement. It’s set up so that no matter what happens in the stock market, this money is never at risk of losing value right when it’s needed.
Why this matters: Imagine if all your spending money for the next year was invested in the stock market, and the market suddenly dropped 20% right when you needed to pay your bills. You’d be forced to sell at a loss just to cover everyday expenses. This bucket exists specifically to prevent that from ever happening.
How it’s funded: There’s an existing automated savings tool (in this case, a “bond ladder”) that pays out a steady monthly stream of money, roughly $6,000 to $7,000. Rather than letting that money pile up, it’s being spent down on purpose to cover this year’s needs. A separate float, roughly one month’s worth of expenses, sits in a regular checking account for quick, everyday access and gets refilled monthly.
Backup options: A few different ultra-safe savings tools (government-backed funds, money market accounts, or high-yield savings platforms) can hold any leftover cash from this bucket that isn’t needed yet that month.
Bucket 2: The Bridge Years Before Medicare
Amount: $320,000, covering Years 1 through 3 of retirement
In this example, before the sample client turns 65 and becomes eligible for Medicare, they typically need to buy their own health insurance, often through a government marketplace. This bucket covers both everyday living expenses and the cost of that health insurance for the three years leading up to Medicare eligibility.
How it’s funded: The same automated savings tool from Bucket 1 keeps paying out and rolling forward into this bucket too, once Bucket 1’s needs are covered. Once that funding source eventually runs dry, a flexible, short-term government savings fund (SGOV, used throughout these pieces) takes over to cover whatever’s left.
Why a flexible option here specifically: Since interest rates are currently more likely to rise than fall (explained in the Fed Rates piece), keeping this money in something that automatically captures higher rates makes more sense right now than locking it into a fixed rate.
Account type matters here, and there’s more than one good option:
A regular taxable account works, but selling investments for gains there counts toward taxable income (MAGI). This is exactly the number that determines an ACA subsidy. So gains here can work against you a little.
A Roth IRA can actually work better. Withdrawing money originally contributed comes out completely tax-free and doesn’t count toward MAGI at all (and once the account holder is 59½ and the account is old enough, withdrawals of the growth become tax-free too). That makes a Roth a genuinely strong fit for a bucket where avoiding extra taxable income is the whole point.
An HSA (Health Savings Account) is arguably the best fit, but only for the health insurance portion of this bucket. Money withdrawn from an HSA for qualified medical expenses, which includes ACA marketplace premiums in some cases, is tax-free and never counts as income, with no waiting period and no contribution-versus-growth distinction to worry about. The catch: new contributions to an HSA stop once someone enrolls in Medicare, so this only helps if there’s already an HSA balance built up from earlier working years to draw down.
The common thread across all three options is to keep this bucket’s withdrawals from quietly inflating the income number that determines a health insurance subsidy.
Contributions
Contributing to an HSA: to contribute, the account holder must be enrolled in a qualifying high-deductible health insurance plan and not yet on Medicare. If the health insurance used during this stretch qualifies as a high-deductible plan, new HSA contributions are still allowed, not just spending down money put in years earlier.
Contributing to an IRA. The requirement is different and stricter: earned income (a paycheck or self-employment income) must be earned in the same tax year. If these bucket years are funded entirely by savings, with no job or freelance income coming in, new IRA contributions generally aren’t allowed that year, no matter how much money is otherwise available. Contributing to an IRA isn’t about having cash on hand; it’s tied to actually earning income in that specific year. (Roth IRAs also phase out at higher income levels, but that’s a less likely concern during a deliberately low-income bridge period like this one.)
A Traditional IRA contribution can actually lower MAGI for that year. The contribution is often tax-deductible (subtracted from taxable income), as long as the contributor isn’t also covered by a retirement plan through a job, or income is below the relevant limit if they are. That makes a Traditional contribution helpful, not harmful, for staying under an ACA subsidy income threshold in a year with earned income.
A Roth IRA contribution doesn’t change MAGI at all, in either direction. It’s money that’s already been taxed, so it’s invisible to that year’s income calculation.
The bigger asymmetry between the two account types isn’t really about contributing, it’s about withdrawing later, which is the actual reason Bucket 2 favors Roth over Traditional for money being pulled out during these ACA years. Traditional IRA withdrawals are treated as taxable income and increase MAGI. Roth withdrawals of original contributions don’t count as income at all. Contributing to either type in a given year is mostly a non-issue for MAGI; it’s pulling money back out where the real difference between the two shows up.
Bucket 3: The Bridge to Social Security
Amount: $400,000 target, covering Years 4 through 7 of retirement
In this example, once Medicare starts but before Social Security payments begin, this bucket covers living expenses plus Medicare’s own premium costs (Medicare isn’t entirely free; it has its own monthly costs, too).
A wrinkle worth knowing about: IRMAA. Medicare premiums aren’t a flat fee for everyone; they increase if taxable income exceeds certain thresholds, through a surcharge called IRMAA. The tricky part is the timing: Medicare looks at a tax return from two years earlier to decide the premium, so a year of unusually high taxable income, such as a large Roth conversion, can mean higher Medicare costs two years later. This is the same kind of income-management puzzle as the ACA subsidies in Bucket 2, just showing up a few years later and tied to a different program.
Why this bucket can take a little more risk: since this money won’t actually be needed for several years, it has time to grow a bit before it’s needed. The goal here is modest growth, just enough to keep pace with rising prices, not aggressive risk-taking.
What it’s invested in: a mix of approaches, since this bucket has a fairly predictable, fixed timeline (several years, a known spending need each year).
One option worth knowing about is a MYGA, a multi-year guaranteed annuity. This is a contract with an insurance carrier in which the holder locks in a fixed interest rate for a set number of years, in exchange for not being able to withdraw the money early without a penalty. For a bucket like this one, where the timing of the need is already known, that locked-in certainty can be appealing.
But timing actually matters here, and it connects directly to the Fed Rates piece: MYGA rates move with broader interest rates, not the Fed’s rate directly, and with the Fed’s own outlook currently leaning toward a possible rate hike rather than a cut, waiting to lock in a MYGA rate may get a better deal than locking one in today. The same logic explained there (rates are more likely to drift up than down right now) applies here too.
The rest of this bucket is typically a mix of medium-term bonds (funds like ABNDX, CGCP, or BOND) and dividend-paying stocks (funds like RSP or QUAL), a more cautious approach than money meant for decades down the road, but a bit more adventurous than the cash sitting in Buckets 1 and 2.
Account type: could go in a Traditional IRA, a Roth IRA, or a regular taxable account, depending on which offers the better tax outcome for this kind of slow, steady growth.
A timing collision worth flagging: RMDs. This bucket covers several years leading up to Social Security. Required minimum distributions from a Traditional IRA start at age 73, and depending on exact birth year, that age can land right in the middle of this bucket’s own timeline, or right at its tail end heading into Bucket 4.
If part of this bucket’s money sits in a Traditional IRA, the account holder may be forced to start withdrawing some of it on the IRS’s schedule, regardless of whether this bucket’s own spending plan actually needed that withdrawal yet. Worth checking a specific RMD start date against this bucket’s timeline directly, rather than assuming it only becomes relevant once Bucket 4 begins.
Bucket 4: Social Security and Beyond
The largest bucket, roughly $1,488,000, covering Year 8 of retirement onward
In this sample profile, once Social Security payments start (assumed at $5,000 a month, or $60,000 a year), they cover most, but not all, of the assumed $80,000 a year in spending needs, leaving a $20,000 yearly gap to fill from savings.
This bucket gets split into two parts:
Part A: Filling the income gap (~$500,000). This portion is sized to reliably cover that $20,000-a-year gap, essentially forever, using a common rule of thumb that a portfolio can typically sustain a 4% annual withdrawal rate without running out of money over a long retirement. Since this is still covering an essential, recurring cost, it’s invested cautiously, similar to Bucket 3.
Part B: Long-term growth (~$988,000). This is everything left over. Since this money has the longest time horizon of any bucket, and the least pressure to be available on a specific date, it can take on the most risk and growth potential. It’s where the bulk of stock investments, gold, and other longer-term holdings live. It can also double as money intended to eventually pass on to family (sometimes called “estate planning”).
Account type: The long-term growth portion makes the most sense in a Roth IRA, since that account type grows tax-free forever and has no RMDs, meaning the holder is never forced to withdraw money on a schedule they don’t control. The income-gap portion fits well in a regular Traditional IRA instead, since its required withdrawals (RMDs) line up naturally with the steady, recurring income this portion is already meant to provide.
What’s Actually Inside the Stock Portion
The 28% allocated to U.S. companies (and the international portion too) isn’t just “buy the whole stock market.” It’s broken into specific themes:
Industrial and infrastructure companies: Businesses involved in power grids, heavy machinery, factory automation, and supply chains, the physical backbone needed to build and run things like AI data centers and the broader energy transition. This might include something like grid equipment maker Eaton (ETN) or a broader fund like (XLI)
Energy companies: Oil and gas infrastructure, along with companies that build and maintain the power grid itself. This might include a pipeline operator such as Williams Companies (WMB) or a grid builder such as Quanta Services (PWR).
Broad, diversified U.S. stock positions: Instead of just buying the most popular handful of giant companies, some of these positions intentionally spread risk across many companies equally (a fund like RSP), or specifically favor financially healthy, stable businesses over riskier, trendier ones (a fund like QUAL
Healthcare companies: This spans two different ideas. Major pharmaceutical companies benefiting from weight-loss and diabetes drugs, such as Eli Lilly (LLY), is one. The other is counter-cyclical: as covered in the K-Shaped Economy piece, when lower-income households lose jobs or income, Medicaid enrollment tends to rise mechanically, regardless of how people feel about the broader economy. Companies involved in managing those Medicaid plans, such as Molina Healthcare (MOH), are one way to get exposure to that pattern, alongside a broader healthcare fund like XLV.
Water-related companies: This might seem like a strange category, but massive AI data centers use enormous amounts of water for cooling, so companies that supply water, water technology, and water infrastructure are tied to the AI boom in a less obvious way. This might include a water utility, such as American Water Works (AWK), or a water-focused fund, such as PHO.
Digital payment companies: Businesses that process electronic payments and are positioned to benefit regardless of how new forms of digital money develop, such as Visa (V).
AI computing companies: The companies that make the computer chips AI runs on, such as Nvidia (NVDA), plus companies involved in surveillance and data technology, such as Palantir (PLTR).
International stocks: Spread across regions like Latin America (a fund like ILF), other developed countries (EFA), and Japan (using a currency-hedged tool like DXJ, since the Japanese yen has been moving a lot, as covered in the Fed Rates piece).
Precious metals: Gold, often held through a fund like IAU, and silver, often held through a fund like SLV, both as direct holdings or through mining companies.
Raw materials and commodities: A mix of general commodities (a fund like PDBC), copper (essential for almost all electronics and construction, often held through a fund like COPX), and uranium (in growing demand to power energy-hungry AI data centers, often held through a fund like URNM).
AI Data Center Build Outs: Seeing the Whole Picture
The categories above each touch on AI data centers from a different angle: the grid and industrial companies, the water companies, the chip makers, the uranium and copper miners, but it’s worth stepping back and looking at the data center buildout as its own connected theme, since that’s genuinely one of the larger forces running through this entire portfolio.
Why does this count as one theme, not several unrelated ones? Building and running a modern AI data center requires four things at once:
1. Massive amounts of electricity (a nuclear or grid company like CEG)
2. Enormous quantities of water for cooling (a water utility like AWK)
3. Specialized computer chips (NVDA)
4. Raw materials like copper and uranium to build and power all of it (a fund like COPX)
A position spread across these four categories isn’t really four separate bets; it’s one bet on the same underlying buildout, viewed from four different angles in the supply chain.
The distinction worth understanding is between who’s selling the shovels and who’s spending the money. As explained in more detail in the Earnings Quality piece, there’s a meaningful difference between companies selling into this buildout (the “enablers,” like chip makers and grid equipment companies, whose demand is already proven and real) and the companies actually spending the money building data centers (the “adopters,” the big tech companies funding $700 to 725 billion in 2026 capex, where it’s still genuinely unclear how fast that spending converts into real profit). Most of the positions in this portfolio sit on the “enabler” side of that line, selling the picks and shovels rather than betting directly on any one company’s AI strategy paying off. That’s a deliberate choice, not a hedge against the bull case being wrong. The demand evidence cited in Earnings Quality is real: Citi raised its 2026-2030 AI revenue forecast to $3.3 trillion, and Microsoft alone has an $80 billion Azure backlog it can’t yet fulfill due to power constraints, not a lack of demand. The enabler-over-adopter tilt captures that demand directly, regardless of which specific AI company ultimately wins the spending race, rather than betting against the buildout itself.
The risk this theme carries is that the individual categories don’t fully capture on their own. If enterprise AI adoption disappoints and big tech companies start slowing their spending (the scenario described in Earnings Quality as a possible 2027 plateau rather than a sudden crash), it doesn’t just hit AI stocks directly, it ripples backward through this entire supply chain at once: less chip demand, less new grid construction, less new water infrastructure investment, less demand for the copper and uranium needed to build it all. Because this portfolio holds pieces across that whole chain, a real slowdown in AI capex would likely show up as a broad, simultaneous softening across several of these categories together, not just in one isolated position.
A geography risk worth knowing about, too. As covered in the Monetary Order piece, Northern Virginia alone hosts roughly 70% of all U.S. internet traffic, an unusually concentrated geographic footprint for critical infrastructure. Several of the data center operators and grid companies touched on in this portfolio have real exposure to that concentration, which is a different kind of risk than the financing or demand questions above; it’s closer to a single-region disruption risk (severe weather, power grid failure, local regulation) than a financial one.
What to actually watch, tying back to Earnings Quality: enterprise AI renewal and contract expansion rates (not new pilot announcements) and whether hyperscaler capex guidance starts to decelerate in 2027 as one possible resolution to the usage gap. Either of those would be the earliest sign that this theme is cooling off, well before it would show up directly in any single stock’s earnings.
“Handle With Care”: Things This Plan Specifically Avoids or Limits
This section is basically a list of common, popular investments that this particular plan deliberately avoids or limits, along with the reasoning. Worth understanding even if you never invest a dollar, because the *reasoning* teaches you how to think about risk.
Plain S&P 500 index funds (funds like ITOT that just buy the 500 largest U.S. companies): These aren’t bad investments; in fact, they’re usually a perfectly reasonable choice for most people. By one common measure of how “expensive” the stock market is compared to company profits, the market is trading near one of its priciest points in history. That’s a real, separate question from whether earnings growth justifies the price. Reasonable analysts disagree here: some, like RIA strategist Brian Vendig, point to strong projected earnings and see no near-term problem. The Earnings Quality piece in this Monitor backs that up to a real degree. Q1 2026 beats were genuine, Citi raised both its capex and revenue forecasts, and Microsoft’s $80 billion unfulfilled Azure backlog points to real, unmet demand rather than overbuilding. Stifel takes the more cautious side of that same debate, warning that valuations “don’t matter until they’re the only thing that matters” and recommending the same defensive tilt, healthcare, staples, and gold, that this plan already leans into. Reasonable, well-resourced research desks sit on both sides of this question right now. Strong earnings and a high entry price aren’t contradictory, though; they can both be true at once. Avoiding a broad index position here isn’t a bet that the bulls are wrong about earnings. It’s a bet that paying today’s price for that earnings strength isn’t the best risk-adjusted way to get the exposure, since the bar for outperformance from here is already higher. On top of that, just the 10 largest companies make up an unusually large share of the entire index, so buying “the whole market” right now means betting heavily on a small handful of giant companies continuing to dominate.
The classic “balanced 60/40” approach (60% stocks, 40% bonds, a fund like AMECX): This traditional strategy assumes that when stocks fall, bonds usually rise to balance things out. As covered in the Fault Lines piece, that relationship has been breaking down during periods of high inflation, both can fall together, which defeats the whole purpose of the balance. It’s worth being fair to the other side here: J.P. Morgan’s own 2026 research still calls the 60/40 portfolio “resilient” with “solid potential” over a 10-15 year horizon, while separately warning in May 2026 that sticky inflation poses a “silent risk to wealth” best addressed by adding real assets and alternatives alongside the 60/40 core, not by abandoning it outright. This plan takes the more cautious side of that same debate, leaning further into real assets than J.P. Morgan’s baseline recommendation, not because the firm is wrong, but because it weights inflation risk more heavily.
Broad clean energy funds (like PBD or ICLN): It’s easy to confuse speculative, trendy clean-energy stocks with the actual physical infrastructure needed to build the new power grid. This plan specifically avoids broad clean-energy fund baskets, since many of them are loaded with companies that spend a lot of money but haven’t become profitable yet, and instead prefers companies more directly tied to the physical grid itself, like Eaton (ETN), Quanta Services (PWR), or GE Vernova (GEV).
High-yield “junk” bonds (funds like HYG or JNK, loans to riskier companies that pay higher interest because there’s more risk of not getting paid back): Earning 7-8% interest sounds appealing, but if the lower-income part of the economy (covered in the K-Shaped Economy piece) continues to weaken, a lot of these riskier companies could struggle to repay their debts. In a real economic squeeze, these bonds tend to lose their “safety” benefit and fall in value alongside stocks, rather than cushioning the blow as they are supposed to.
Actively managed large-company growth funds (like AGTHX): Similar concern to the plain index funds above, a focus on big, expensive growth companies when the overall market is already historically pricey.
Long-term government bonds (a fund like TLT): as covered extensively in the Fed Rates piece, betting heavily on long-term bonds right now means betting that interest rates will fall soon. The current signals point the other way.
Stores that depend heavily on lower-income shoppers, like Dollar General or Dollar Tree: discount retailers might seem like a safe, recession-proof bet, but if their core customers are already financially stretched (as covered in the K-Shaped Economy piece), these companies may actually struggle more than people expect, not less.
Broad consumer discretionary stocks (companies selling non-essential goods): With paychecks not keeping up with prices and government benefit cuts coming, this category of “nice-to-have” purchases is considered especially exposed right now.
High-risk, high-growth innovation-focused funds (like ARK Innovation, ticker ARKK): These tend to be the most sensitive to interest rate changes, since so much of their expected value depends on profits far in the future, exactly the dynamic explained in the Fed Rates piece about why rate hikes hit fast-growing companies hardest.
Broad real estate investment funds, especially those tied to office buildings (a fund like VNQ): As covered in the Fault Lines piece, office buildings specifically are under serious stress right now. This plan prefers obtaining “real asset” protection through physical commodities and precious metals rather than real estate funds that may be loaded with struggling office properties.
The Big Takeaway
Strip away all the account types and ticker symbols, and the core lesson of this entire piece is something anyone can apply, even with a much smaller amount of money: figure out roughly when you’ll need each portion of your savings, and match how much risk you take to how soon you’ll need it. Money needed soon stays safe and boring. Money you won’t touch for many years can afford to ride out some bumps along the way, because it has time on its side.
Reminder: the figures, account placements, and tax strategies above are illustrative and based on a hypothetical sample client. They are not a recommendation, are not personalized to your situation, and should not be used as a substitute for advice from a qualified financial advisor, CPA, or attorney who can review your actual income, accounts, goals, and tax picture.