Starting a prop firm used to take a trading platform, a risk team, a support desk, legal help, and months of build time. In 2026 a small team can launch one in weeks: rent a white-label trading engine, plug into a perpetual futures venue for prices, and use AI to write the rulebook, the website, the support answers, and the marketing. That's why a whole category of perps prop firms showed up almost overnight. The flip side is that launching is now easy, and paying traders is still hard. This guide covers both sides: what it takes to start one, where AI actually helps, and how traders can tell the real firms from the rest.
TL;DR
To start a prop firm today you need four things: a trading platform (usually white-label), a price source, a written ruleset with drawdown and payout terms, and enough money set aside to pay the traders who pass. AI tools like Claude have collapsed the cost of everything except that last one, which is why so many crypto and perps prop firms launched in the past year. If you're a trader, the useful question isn't "which firm has the best score" but "which firm can I prove pays." PerpsMatch (opens in a new tab) tracks that by following payout wallets on-chain. If you want to use AI for the business side of a firm (or any business), the free Claude Cowork course walks you through it.
What a prop firm actually is
A proprietary trading firm (prop firm) gives traders access to firm capital, usually after they pass a paid evaluation. You pay a fee, trade a challenge account under rules like a profit target and a maximum drawdown, and if you pass, you get a funded account and keep a share of the profits, often 80 to 90 percent.
The model started in forex and futures with firms like FTMO and Apex. Crypto traders adopted it next, and in the last year or two it moved onto perpetual futures (perps), the always-open leveraged contracts that trade on venues like Hyperliquid, Bybit, and Kraken. Many evaluation accounts are simulated: you trade against real market prices, but the firm pays you from its own funds rather than routing your orders to an exchange. That matters later when you're judging whether a firm can pay.
Why perps prop firms appeared overnight
A few things happened at once.
The plumbing became something you rent
White-label platforms now let you launch a fully branded prop firm on someone else's trading engine, risk system, and payout tooling. Most forex-style firms launching in 2026 run on a white-labeled Match-Trader, cTrader, or MetaTrader 5, and crypto-native options have appeared too. On the perps side, firms can price a simulated account off Hyperliquid or Bybit market data without building an exchange.
Exchanges started buying in
The category got a credibility boost when Kraken acquired Breakout, a crypto-native prop firm founded in 2023, completing the deal on September 1, 2025. It was the first major crypto exchange to own a prop firm outright. Other venues followed with programs of their own, and established futures firms launched perps products.
On-chain payouts made new firms easier to trust
Some new firms pay traders in USDC from public wallets or smart contracts, so anyone can check the payouts on a block explorer. Propr, for example, launched directly on Hyperliquid with every payout recorded on-chain and an API built for automated trading agents. A new firm with checkable payouts can earn trust faster than an old firm that only posts a screenshot.
AI removed the paperwork bottleneck
This is the part people underestimate. A prop firm is mostly documents: rules, terms, FAQs, payout policies, onboarding emails, support replies, comparison pages, and daily marketing. Two years ago that was a team of writers and a lawyer's time. Now a founder with Claude can draft all of it in a week and spend their real effort on the risk model.
Put those together and you get what we have now: PerpsMatch (opens in a new tab) tracks around 18 perps prop firms across 9 venues at the time of writing, many of them under two years old, with evaluations starting around $10.
How to start a prop firm: the steps
If you're seriously considering it, this is the shape of the work. None of it is legal or financial advice. Talk to a lawyer about how your jurisdiction treats evaluation fees and funded accounts before you take a customer's money.
| Step | What it involves | Where AI helps |
|---|---|---|
| 1. Pick a market and venue | Forex, futures, or crypto perps; which venue prices your accounts | Summarizing venue docs and fee schedules |
| 2. Choose your platform | White-label provider vs. building your own terminal | Comparing provider contracts and feature lists |
| 3. Design the rules | Profit targets, daily and max drawdown, leverage, consistency rules, payout schedule | Drafting the rulebook and stress-testing it for loopholes |
| 4. Model the risk | How many traders pass, how much you pay out, how much reserve you need | Building the spreadsheet model with you (you verify the math) |
| 5. Legal and payments | Entity, terms of service, KYC, payout rails (card, USDC) | First drafts of terms for your lawyer to review |
| 6. Website and support | Pricing pages, FAQ, help docs, support replies | Nearly all of the first drafts |
| 7. Payout transparency | Public wallets or a payout ledger traders can check | Writing the explainer and monthly payout reports |
Step 3 is where firms live or die
The rules are the product. Vague wording causes the arguments that end up on Trustpilot and X ("I hit the target, why was my payout denied?"). Use AI here as an adversary, not just a writer:
Here is our draft prop firm rulebook: [paste]. Act as a trader trying to get paid while breaking the spirit of these rules. List every loophole, ambiguous term, and rule that could be read two ways. Then act as a trader who was denied a payout and list every rule they could reasonably say was unclear. Don't rewrite anything yet, just list the problems.
Then fix the wording yourself and run it again.
Step 4 is where AI can't save you
AI can help you build a pass-rate and payout model, but it can't tell you what your real pass rate will be, and it will happily produce a confident-looking spreadsheet with a wrong assumption buried in it. Check every formula. The firms that close are almost always the ones that sold more evaluations than they could afford to pay out on.
Where AI helps traders, too
The same tools work from the other side of the table. If you're picking a firm, the hard part is reading 18 sets of terms and noticing the clause that changes everything: a consistency rule, a weekend holding ban, a payout cap, a drawdown that trails instead of staying static.
Here are the rules and payout terms for this prop firm, copied from their site: [paste]. Summarize in a table: profit target, daily drawdown, max drawdown (static or trailing), leverage, consistency rule, news or weekend restrictions, payout schedule, payout cap, and profit split. Then list anything unusual that could stop a passing trader from getting paid. Quote the exact sentence for each item.
Asking for the exact quote matters. It keeps the AI from inventing a rule and lets you check each line against the source in seconds.
How to tell which new prop firms actually pay
This is the question that matters most, and it's where most "best crypto prop firm" lists fall short. They hand out scores, but nobody outside a firm can see its balance sheet, and a five-star rating on a nine-month-old firm doesn't tell you much.
PerpsMatch's methodology (opens in a new tab) takes a different approach that I think is the right one for a market this young. They don't rate or rank firms. Instead they check four things:
- Payout proof. Where a firm publishes payout wallets or contracts, they follow them on the block explorer and total what was actually sent to traders. If a firm only posts a total on its site, it's labeled as the firm's own claim.
- Rules from the source. Prices, targets, drawdowns, and payout terms are checked against each firm's own docs, with the sources and last-checked date listed.
- Things to know before buying. Rule quirks like consistency rules and payout caps are listed as facts to weigh, not marks against the firm.
- Red flags. Only documented payout problems count: denied, withheld, or clawed-back payouts. Firms that keep denying valid payouts get delisted.
That's a useful checklist even if you never use the site. A firm with a long, checkable on-chain payout record has already passed the hardest test, and a firm that just launched hasn't had the chance to.
Where AI fails in trading
Worth saying plainly, because "AI for trading" gets sold as a shortcut to profits.
- AI doesn't predict prices. A chatbot has no edge in the market, and most have no live data at all. Anyone selling an "AI that passes your challenge" is selling you the evaluation fee twice.
- It invents details. Ask Claude or ChatGPT about a specific firm's rules from memory and it may give you outdated or made-up terms. Always paste the current terms in.
- It sounds certain when it's wrong. A clean risk model or rulebook draft can still be wrong. Verify numbers and have a lawyer read anything legal.
- It can't check payouts. Only the blockchain, the firm's records, or a third party following the wallets can do that.
Used for documents, analysis, and checklists, AI is a big time-saver. Used as a trading oracle, it's a way to lose money faster.
Frequently asked questions
How much does it cost to start a prop firm?
It depends mostly on your platform and your payout reserve. White-label platforms cut the build cost a lot, and AI cuts the cost of content and support, but the biggest number is the cash you hold back to pay funded traders. Budget for legal review on top of that.
Can you start a prop firm with AI?
AI can't run a prop firm for you, but it can handle most of the writing work: rulebooks, terms drafts, websites, FAQs, support replies, and marketing. The trading platform, risk management, legal setup, and payout funding still need real people and real money.
What is a perps prop firm?
A perps prop firm funds traders to trade perpetual futures, usually on crypto, after they pass an evaluation. Many price their accounts off venues like Hyperliquid or Bybit and pay out in USDC, sometimes from public wallets anyone can check.
How do I know if a crypto prop firm is legit?
Check whether it pays. The strongest evidence is payouts you can verify on-chain from a public wallet or contract. After that, look for a clear rulebook, a known company or backer, and recent payout reports from traders. Sites like PerpsMatch track this for perps firms.
Is a white-label prop firm a good idea?
For most new founders it's the fastest way to launch, because you rent the trading engine, risk system, and dashboard instead of building them. The tradeoff is less control and ongoing platform fees, and your firm can look very similar to every other firm on the same platform.
Learn to use AI for the business side
Whether you're running a prop firm, a trading education business, or something unrelated, the useful AI skills are the same: getting accurate first drafts, checking documents for problems, and building simple workflows you can repeat. The free Claude Cowork course teaches exactly that in plain English. For more on putting AI to work in a small business, see how to use AI for my business and AI agents for business.