From signup to your first rebalance.
Five signup steps, then the custodian’s own account opening. After that, one rebalance and one letter each quarter.
Opening your account
Five steps, in the order you will see them.
01
Create your login.
Email, password and a mobile number for sign-in codes. Oakshire never sees your brokerage login or your bank details; those go to the custodian directly.
02
Answer the questionnaire.
Identity, income, net worth, objective, time horizon, risk tolerance and experience. If the answers say this portfolio does not fit you, signup stops there.
03
Read and sign.
Form ADV Parts 2A and 2B, Form CRS, the privacy notice and the advisory agreement, signed once by typing your name. Every document stays in your dashboard.
04
Choose a plan.
$15 a month, or $10 a month billed yearly. Cancel any time and the model stops rebalancing; your positions stay.
05
See the portfolio, then fund it.
The 20 names the model holds this quarter, before you deposit anything. Then the custodian's own account opening, and a first deposit of $100 or more.
What goes in, what comes out.
Only public data goes in
The model reads 124 inputs: prices and returns, quarterly financial statements, interest rates and inflation, and a few research files anyone can download. Academics first studied each one as a predictor of stock returns. It retrains every year, on past years only.
Twenty companies, one list
The model ranks every listed US company against its peers, keeps the top tenth, and holds the 20 largest, one per peer group. Everyone gets the same 20. Each quarter brings a rebalance and a letter: who left, who arrived, and why.
What you keep
The account stays yours.
The account is in your name.
The account is in your name at a US brokerage. Oakshire holds trade authorization and nothing else: it cannot withdraw, transfer or borrow against your money.
Cancel and nothing moves.
End the subscription and the model stops rebalancing. Every position stays exactly where it is, in your account, for you to keep, sell or move.
Every trade is visible.
The dashboard lists each purchase and sale with the reason. In a taxable account, sales can create capital gains or losses; the custodian issues the tax forms, and Oakshire does not give tax advice.
The model can be wrong for years at a time, and the account can lose money. Past results do not predict future ones. Oakshire acts once a quarter and never in between, so news in February waits until April.
The research
Public information is free to read and expensive to use.
Every listed company’s numbers are public: prices, quarterly statements and the economic data around them. Reading them costs nothing. Using them well takes time, skill and money, so most of that information goes unused. We built Oakshire around that gap.
Our model is a random forest, a kind of machine learning that learned its rules from data instead of rules a person wrote. Every year we retrain it on the years before, so it never sees the period it ranks.
It runs on 124 inputs. Each one was first studied in academic research as a predictor of stock returns, in research running from 1973 to 2025. We read that research and the papers behind it. We run the method the way the research describes it.
Founding list
Get the launch date first.
One email when subscriptions open, and the launch price for as long as you stay subscribed.