Market Data Licensing: What Traders Need to Know

Here is a small mystery of modern markets. You can pull up a live share price on your phone in about two seconds. You can also hand a stranger a paper bag of cash for a flat, and nobody files paperwork. But the moment you build a product that shows that same share price to other people, you enter one of the most carefully policed licensing regimes in finance.

That is not a complaint. It is a map. And once you have the map, the terrain stops feeling random.

Why market data has owners at all

Exchanges are not libraries. They are businesses that match buyers and sellers, and the prices produced by that matching are, in effect, a product. To protect the integrity of that product — and to make sure everyone is looking at the same picture — regulators built rules around how it is distributed.

In the United States, the spine of those rules is Regulation NMS, short for National Market System. It created a structure where exchanges send their quotes and trades to central "securities information processors," usually just called SIPs. The SIPs stitch everything together into a single official view: the national best bid and offer, mercifully shortened to NBBO. That is the number you see quoted on a broker's screen.

Rule 603 of Regulation NMS then does something quietly important. It says information must be distributed on terms that are "fair and reasonable," and not "unreasonably discriminatory." It also contains the Vendor Display Rule, which requires broker-dealers to show customers a consolidated display — the NBBO, with prices, sizes and venue identities — when they are giving those customers quote information.

One nuance worth holding onto: a common industry practice is to display that consolidated quote only where the customer actually places an order, and to use cheaper indicative feeds elsewhere for a smooth-looking experience. Compliance and user experience, in other words, can be answered by different pipes.

The tier ladder, and why cost tracks precision

Most confusion around data licences comes from assuming it is all-or-nothing. It is not. U.S. equity data stacks into tiers, and cost scales with precision rather than usefulness.

At the bottom sits historical data — pricing, trades and volume at least one full trading day old. From an exchange-fee perspective it is the cheapest rung: no exchange redistribution fees, no reporting, no audits. It is not literally unowned, though. Exchanges still hold the data; they simply do not levy fees for it. What you pay is your vendor for access, storage and delivery.

Next comes delayed consolidated data, typically fifteen minutes behind. This still comes from the SIPs and supports intraday candlesticks, technical charts and portfolio views. It is dramatically cheaper than the real-time version — illustrative pricing runs to a modest monthly base fee plus a small annual administrative charge, with no per-user display fees.

Then there is real-time consolidated data, and this is where eyebrows ascend. Once you redistribute it externally, costs routinely reach five figures per month, before per-user charges, reporting obligations and compliance overhead. The number scales as your audience grows.

Above or beside that sit two escape hatches. Single-exchange feeds, such as IEX TOPS, give genuine real-time bid, ask and last-sale prices from one venue for a few hundred dollars a month, with no per-user fees or professional classifications — they track the broader market closely, but they are not the official NBBO. And derived data feeds calculate prices from market inputs rather than redistributing raw exchange quotes, so they typically trigger no exchange display licence at all; your terms come from the vendor alone.

The practical sequence many builders follow — history first, derived pricing next, then delayed intraday, then a single-exchange feed, and only much later consolidated real-time — aligns spending with product maturity. Think of it as buying a bicycle before a delivery van. Both move things. Only one of them makes sense on day one.

What you are actually paying for

Exchange fee schedules are a vocabulary test, so here is the glossary in plain words.

An access fee is a flat monthly charge for connecting to a feed, regardless of how many people use it. A user fee is charged per display — per screen showing the data — and varies by whether the user is classified as professional or non-professional. A non-display fee covers data feeding computers rather than eyes: execution algorithms, smart order routers, market-making engines. And a redistribution fee applies when you move data to a system, person or business other than the original purchaser.

That last one is the trap. Showing a price to your users is redistribution. Vendors then add their own charges on top — support, hardware, exchange reporting, port connectivity — which is why the same feed can carry very different total costs depending on who you buy it through.

Classification deserves care. Accessing data through a business entity generally makes you professional regardless of entity size. Trading or researching for others makes you professional. Holding a regulatory registration — SEC, FINRA, CFTC, NFA, the FCA, ESMA-member regulators, ASIC and equivalents — also makes you professional, even if the licence is dormant. But classification follows the use, not the person: a hedge fund quant running only a personal account on the side may still be non-professional for that personal activity. These determinations rest with the exchanges, and vendors act as vendor of record enforcing them.

The paperwork has a calendar

Licensing is not a checkbox, it is a queue. Non-professional users who are not redistributing can often sign an attestation and get access quickly. Professional users and redistributors are typically introduced to the exchange to obtain a formal licence, and that conversation cannot be automated. Vendors advise starting three to six months ahead. If a launch date is fixed and licensing is not started, the launch date is a wish.

The regulatory picture is also moving. ESMA's review of MiFID II found that the framework had not yet delivered its objective of reducing venue data costs, and recommended a real-time EU consolidated tape for equities, alongside supervisory guidance on what "reasonable commercial basis" means. In the U.S., the SEC adopted amendments to Regulation NMS covering tick sizes, access fee caps and odd-lot transparency, with compliance dates phased across 2025 and 2026. Separately, the SEC has proposed rescinding the Order Protection Rule and the ban on locked and crossed markets, with comments due in August 2026 and temporary relief extending access-fee and tick-increment compliance dates to late 2027. These are proposals and phased timelines, not settled outcomes, and they differ by jurisdiction. Anyone budgeting years ahead should watch them rather than assume.

What tends to go wrong

The recurring mistakes are rarely technical. Builders see consolidated real-time pricing, conclude the whole idea is impossible, and walk away — when in truth very few early products need SIP-level accuracy. Others assume "real-time" and "consolidated" are synonyms, when single-exchange and derived feeds deliver live-feeling prices without the licence weight. Many agonise over audits and reporting before a single user exists. And almost everyone overlooks derived feeds entirely.

There is also a structural option worth knowing: a bring-your-own-key model, where each user connects with their own entitlement and data never passes through your servers. It can sidestep redistribution status, but it demands a specific architecture — a backend that fetches on the user's behalf breaks the model. Note that vendors are not unanimous; some discourage the approach over onboarding friction and licensing complexity.

None of this is investment advice, and none of it is legal advice. Fee schedules change, exchanges apply discretion, and the rules genuinely differ between the U.S., Europe and elsewhere. Verify current pricing and requirements directly, and take professional advice on your own structure.

What you can rely on is the shape of the thing. Data has an owner, access has a price, and precision is what you are buying. Decide how much precision your users actually need, and the rest of the map draws itself.