Build your own ETF: create a custom ETF portfolio you weight yourself.
You cannot list a real fund without six figures and a board. You can build exactly what one holds. Define the members, set the weights, test the construction over real history, then track it like an index.
Educational only · Never places a trade
In short
An individual investor cannot create a real, exchange-traded ETF. Launching one through a white-label platform runs roughly $50,000 to $150,000 upfront, costs about $250,000 to $500,000 a year to operate, and needs $1 million to $2 million of seed capital before a single share trades. What people mean by building your own ETF is a custom basket: a defined list of stocks or ETFs, with stated weights and a rebalancing rule, held directly in your own brokerage account. That version has no expense ratio, gives you individual tax lots, and lets you use any weighting method you like. The part worth doing carefully is the construction, because a basket is only as good as the rule that decides what goes in it and how much of each you hold.
Last updated July 2026
Can you create your own ETF?
What launching an actual ETF costs
Worth putting the real figures up front, because most articles on this topic skip them and leave people thinking a fund launch is a paperwork exercise. These are industry estimates and they vary by structure, so treat them as ranges.
| Line item | Typical range | What you are paying for |
|---|---|---|
| Upfront launch cost, white-label | Roughly $50,000 to $150,000 | Registration statement, SEC filings, board preparation and legal work, done on someone else's trust structure. |
| Upfront launch cost, independent | Can exceed $500,000 | Your own trust, your own board, your own compliance function. Almost nobody does this for a first fund. |
| Fixed operating cost | Roughly $250,000 to $500,000 a year | Audit, administration, custody, listing, index licensing and compliance. It is owed whether the fund gathers assets or not. |
| Seed capital to list | $1 million to $2 million | What it takes to create the first shares and open for trading. It is the floor, not a viable size. |
| Assets before allocators engage | Commonly $10 million to $25 million | Many platforms and advisors will not buy a fund below this, which is why so many small ETFs close. |
The economics explain why the ETF shelf looks the way it does. A fund carrying $300,000 a year of fixed cost needs a lot of assets before a 0.50% fee covers it, which is roughly $60 million. Below that the sponsor is subsidizing the fund out of pocket, and most sponsors will only do that for so long. That is the mechanism behind the thematic fund closure rate: it is rarely that the theme was wrong, it is that the fund never gathered enough assets to pay its own bills.
None of which matters if what you actually want is the portfolio rather than the product. A fund wrapper exists so that strangers can buy your idea in one trade. If the only person who needs to own it is you, the wrapper is pure overhead, and everything below is about building the holdings directly instead.
Custom ETF builder options
Six ways to build your own ETF, and what each one really is
An honest comparison, including where each option beats ours. Pricing checked against each provider's own page in July 2026, and it does change, so confirm before you commit.
| Route | What you get | Cost | Minimum | Rebalancing | Best for |
|---|---|---|---|---|---|
| Launch a real ETF | A ticker anyone can buy | Six figures upfront plus six figures a year | Seed capital of $1M to $2M | You set the methodology | Asset managers, not individuals |
| Fidelity Basket Portfolios | A named basket of 2 to 50 stocks or ETFs you trade as one unit | $4.99 a month after a 30-day trial | $1 per security, no account minimum | You trigger it; the basket rebalances to your targets | The closest retail thing to a personal ETF |
| M1 pies | A target allocation that auto-rebalances new money into it | Commission-free; $3 a month platform fee unless you hold $10,000 or more | Fractional, low | Automatic on contributions | Set-and-forget contribution investing |
| Fractional shares, DIY | Individual positions you size yourself | Commission-free at Schwab and Robinhood | $1 per order | Entirely manual | Small baskets you are willing to hand-manage |
| Managed direct indexing | A separately managed account holding index constituents directly | 0.09% to 0.40% a year | $5,000 to $100,000 depending on provider | Automatic, with tax-loss harvesting | Larger taxable accounts wanting tax management |
| Design and test it first (Indexes) | A weighted index you define, backtest and track | From $12 a month | None; nothing is held or traded | You set the schedule and see its effect | Deciding the methodology before you commit money |
Broker basket tools
Fidelity Basket Portfolios is the closest retail equivalent to a personal ETF: 2 to 50 positions in one named basket, bought and rebalanced as a single action, $4.99 a month after a 30-day trial, $1 minimum per security and fractional shares to three decimals. M1 approaches it from the other side, letting you set target percentages and routing every new contribution toward them automatically. Neither one tests the construction for you.
Plain fractional shares
Schwab retired the Stock Slices branding and now offers fractional investing in most US-listed stocks and ETFs from $1, commission-free online. Robinhood accepts fractional orders of $1 or more in most stocks trading above $1 with a market capitalization over $25 million. Both are enough to build a weighted basket by hand. Neither gives you a basket-level rebalance button, so the arithmetic is yours.
Managed direct indexing
If what you want is an existing index held directly, with losses harvested automatically, that is a different product with a different price. Wealthfront starts at 0.09% with a $5,000 minimum on its S&P 500 offering, Fidelity's managed FidFolios run 0.40%, and Schwab Personalized Indexing starts at 0.40% with a $100,000 minimum. See direct indexing for the full comparison.
Being direct about where we sit in that table. Indexes does not hold your money, does not connect to a brokerage and never places a trade. You will still buy the shares wherever you already have an account. What we do is the part none of the brokers do: let you define the basket and its weights, run the exact combination over real market history, compare weighting methods against each other, and keep the result as a tracked index measured against the S&P 500, the Nasdaq 100 or BTC. The order that tends to work is design and test here, then execute at your broker.
How to build your own ETF portfolio
Build the basket in four decisions
Write the membership rule
Before any tickers, write the sentence that decides what belongs. "US-listed companies that make grid hardware" is a rule. "Stocks I like" is not. The rule is what makes the basket an index rather than a watchlist, and it is what tells you later whether a name still belongs.
Choose a weighting method
Cap weighting mirrors the market and concentrates in the biggest names. Equal weighting expresses the rule evenly and tilts smaller. Capped weighting stops any one holding from taking over. Same members, very different portfolios, and this decision moves returns more than most stock picks do.
Backtest the exact combination
Run the weighted basket over historical prices and read the drawdown alongside the return. Then run it again with a different weighting and compare. This is the step people skip, and it is the only one that tells you how the construction behaves when the theme goes out of favor.
Set the rebalancing schedule first
Decide quarterly, semiannual or annual before you buy, not after a position runs. Real indexes publish their schedule in advance for the same reason: a rule you set when nothing is at stake is the only rule you will actually follow when something is.
Step two is where custom baskets are won and lost. Take the same 30 companies and weight them by market capitalization, and a handful of megacaps will drive most of the result; weight them equally and you have a different risk profile, a smaller-company tilt and a real rebalancing cost. Neither is correct in the abstract. The comparison is worked through in equal weight index, and the mechanics of applying a method live in index construction.
On step four, the schedule matters more than the frequency. Rebalancing more often is not better; it raises turnover and, in a taxable account, realizes gains you did not need to realize. Established indexes mostly reset quarterly or annually, and portfolio rebalancing covers how to pick a cadence and what the evidence says about how much it actually earns you.
One caveat that applies to every backtest, including ours. It is hypothetical historical performance: it shows what a set of weights would have done over a period that already happened, which is useful for understanding how a construction behaves and useless as a prediction. Past performance is no guarantee of future results, and everything here is educational content rather than investment advice.
Custom ETF vs buying an ETF
When building beats buying, and when it does not
Build it yourself when
- No existing fund holds what you want, or the closest one holds forty names you would not choose.
- You want to exclude specific companies, whether for a values reason or because you already own that exposure through your employer.
- You want equal or capped weighting instead of the cap weighting nearly every fund uses.
- You want individual tax lots, so you can harvest a loss on one holding without selling the whole position.
- The theme is narrow enough that any fund covering it charges well above 0.50%.
Just buy the fund when
- A broad-market fund at 0.03% to 0.10% already holds essentially what you described. You will not beat that on cost or effort.
- You want the basket in a retirement account where the tax-lot advantage is worth nothing.
- You would need more than 40 or 50 holdings to express the idea properly.
- You know you will not do the rebalancing. An unmaintained basket drifts into whatever its winners became.
- The exposure involves anything you cannot buy directly as a retail investor.
The clean test: name one specific thing the closest existing fund gets wrong for you. If the answer comes quickly and concretely, a custom basket is the right call and the difference will show up in the backtest. If it takes a while, buy the fund and spend the time somewhere it pays better. The full head-to-head is in custom index vs ETF.
Who this fits
Who ends up building a custom ETF
Theme investors no fund serves
The thesis is specific and the nearest ETF is a loose approximation of it holding sixty names. See thematic investing for why the fund versions so often disappoint.
Anyone excluding a holding
Employer stock, a sector you already have too much of, or a company you will not own. A fund cannot leave one name out for you; a basket is nothing but that choice.
Equal-weight believers
Almost every fund is cap weighted, which now means heavy concentration in a few megacaps. Building the basket is the straightforward way to weight it differently.
People who want tax lots
Holding constituents directly means a loss in one name is harvestable while the rest keeps running. That is the whole idea behind direct indexing tax loss harvesting.
Crypto and equity in one basket
No ETF weights tokens and stocks together the way you would. Weighting both inside one index is a specific gap, covered on the crypto index page.
Anyone testing before buying
The cheapest version of a bad basket is the one you modeled and rejected. Backtest the portfolio before any money moves.
Questions
Building your own ETF, answered
Can I build my own ETF?
Not in the literal sense. An ETF is a registered investment company with a sponsor, a board, a custodian, an authorized participant and an exchange listing, and standing one up costs roughly $50,000 to $150,000 upfront through a white-label platform before any ongoing costs. What you can do, and what almost everyone means by the question, is build the same thing an ETF holds: a defined basket of securities with stated weights and a rebalancing rule, held in your own brokerage account.
Can you create your own ETF as an individual investor?
You can create the portfolio, not the fund. The distinction matters because the fund wrapper is what gives an ETF a ticker, intraday liquidity for other people, and its particular tax treatment. A personal basket gives you the same holdings and the same weights, with direct ownership of every share, individual tax lots you control, and no expense ratio. What it does not give you is a security anyone else can buy.
How much does it cost to start your own ETF?
Industry estimates put a white-label launch at roughly $50,000 to $150,000 upfront, with an independent launch running past $500,000. Fixed operating costs run about $250,000 to $500,000 a year regardless of assets, and a fund typically needs $1 million to $2 million in seed capital simply to list. Most allocators will not consider a fund until it holds $10 million to $25 million, which is why small launches so often close within a few years.
How do I make my own ETF portfolio?
Four steps. Write down the rule that decides membership, so the basket has a reason to exist beyond a list of favorites. Pick a weighting method and apply it consistently. Backtest the weighted combination over real history and read the drawdowns, not only the returns. Then set a rebalancing schedule before you buy anything, because the schedule is what keeps the basket matching the rule as prices move.
How many stocks should a custom ETF hold?
Between 15 and 30 for a broad basket, and 8 to 15 for a focused theme. The academic work on diversification shows most of the benefit arriving well before 30 names, with each addition after that buying very little. Fidelity Basket Portfolios caps a basket at 50 positions, which is a reasonable practical ceiling: past that you are running an index without index infrastructure.
Can you make your own ETF on Robinhood?
Robinhood has no basket product, so you build the position manually. Fractional orders must be at least $1, and eligibility covers most stocks trading above $1 per share with a market capitalization over $25 million, listed on national exchanges rather than over the counter. That is enough to construct a weighted basket from scratch. What it does not include is a tool that rebalances the whole set back to your targets, so you do the arithmetic yourself.
How do I build my own ETF at Fidelity?
Fidelity Basket Portfolios is the feature built for exactly this. You create a basket of 2 to 50 stocks or ETFs, assign each one a weight, and buy or rebalance the whole basket in a single action. It costs $4.99 a month after a 30-day free trial, has no account minimum, and supports fractional shares out to three decimal places as long as each order is at least $1.
Is it better to build your own ETF or just buy one?
Buy the ETF if a fund already holds close to what you want, because at 0.03% to 0.20% you cannot beat that on cost or convenience. Build the basket when no fund matches: you want a theme expressed differently, you want to exclude specific companies, you want equal weighting instead of cap weighting, or you want individual tax lots. The honest test is whether you can name something a real fund gets wrong for you. If not, buy the fund.
Does a custom ETF have an expense ratio?
No. A basket you hold directly has no expense ratio, because there is no fund charging one. Your costs are whatever your broker charges, which at the major US brokers is commission-free for stock trades, plus any basket subscription such as the $4.99 a month Fidelity charges. The trade is that you absorb the work an expense ratio pays for: the rebalancing, the corporate actions and the record keeping.
What is the difference between a custom ETF and direct indexing?
Scale and who does the work. Direct indexing is a managed service that buys the constituents of an established index in a separately managed account and harvests losses automatically, priced at 0.09% to 0.40% a year with minimums from $5,000 to $100,000. A custom basket is your own list, your own weights, and no manager. Direct indexing tracks somebody else's index efficiently; a custom basket exists because you did not want to track theirs.
A broker-by-broker walkthrough of the mechanics, with what each platform does and does not automate, is in building your own ETF at Fidelity, Schwab, Robinhood and M1.
Design the basket before you buy it
Pick the members, set the weights, backtest the exact combination over real market history, and track it as a named index against the S&P 500 or BTC. No account to link, no minimum, no trades. Educational and informational only.