In the modern U.S. equity market, a single stock does not trade in one place — it trades simultaneously across dozens of competing venues stitched together by regulation into a "National Market System." These venues fall into three buckets: lit exchanges (public, pre-trade-transparent order books like the NYSE and Nasdaq), Alternative Trading Systems / ATSs (broker-operated venues including ECNs and dark pools, regulated under Reg ATS rather than as exchanges), and wholesaler internalizers (market-making firms that fill retail orders against their own inventory off-exchange). The core tension of this structure is fragmentation vs. competition: spreading volume across many venues drives down explicit fees and tightens spreads, but it also splinters liquidity, complicates routing, and hides a large share of trading from the public quote.

The venue landscape

Lit exchanges. As of 2024 there were roughly two dozen registered national securities exchanges, though most are operated by three groups: Intercontinental Exchange (NYSE, NYSE American, NYSE Arca, NYSE Chicago, NYSE National), Nasdaq (Nasdaq, BX, PSX), and Cboe Global Markets (BZX, BYX, EDGX, EDGA). Smaller independents — IEX, MEMX, MIAX Pearl, LTSE — together trade a low single-to-high-single-digit share of volume (industry summaries put the five smallest exchanges under ~8% combined). Exchanges must publish their best bid/offer to the public (the "lit" book) and are bound by exchange-level fair-access and self-regulatory obligations.

ATSs (ECNs and dark pools). An ECN (Electronic Communication Network) is an automated order-matching system that lets buyers and sellers interact directly without an exchange-floor intermediary. Instinet (1969) was the first; Island (1996, Levine and Citron) brought low fees and the ITCH feed and became the day-trading workhorse before merging with Instinet in 2002. Most original ECNs (Island/INET, Archipelago, BRUT) were absorbed into Nasdaq and NYSE in the mid-2000s. The SEC's Regulation ATS (1998) created the framework these venues live under: an ATS registers as a broker-dealer rather than an exchange, with lighter transparency obligations. Dark pools are ATSs that do not display quotes pre-trade — designed so institutions can work large blocks without telegraphing intent.

Wholesalers / internalizers. Market-making firms (Citadel Securities, Virtu, G1 Execution Services/Susquehanna, Wolverine, Jane Street) execute retail orders in-house against their own book. These are not exchanges or ATSs; they are off-exchange trades reported to a FINRA Trade Reporting Facility (TRF).

How it's used in practice

Routing is governed primarily by Regulation NMS (2005). Its Order Protection Rule (Rule 611) forbids a venue from executing a trade at a price inferior to a protected (automated, top-of-book) quote displayed elsewhere — preventing "trade-throughs." Together the protected quotes form the NBBO (National Best Bid and Offer), the consolidated best price across all exchanges. A broker satisfies its best-execution duty by routing to capture (or improve on) the NBBO, considering price, speed, and fill likelihood.

Venue economics drive most routing decisions:

• Maker-taker (most lit exchanges): pays a rebate to liquidity providers (resting limit orders, "makers") and charges a fee to liquidity takers (marketable orders). Reg NMS (Rule 610) caps the access fee for stocks ≥ $1; that cap was historically $0.0030/share, but the SEC's 2024 Reg NMS amendments cut it to $0.0010/share with a compliance date of November 3, 2025 (per SEC final rule and law-firm summaries) — so the lower cap is now in force.
• Taker-maker / inverted venues flip the sign to attract marketable flow.
• Payment for order flow (PFOF): a wholesaler pays the retail broker to receive its orders, then internalizes them — typically offering modest price improvement over the NBBO. Reg NMS access-fee caps do not apply off-exchange.

For an institution, the practical workflow is a smart order router (SOR) that slices a parent order across lit books, dark pools, and ECNs to minimize market impact. For a retail customer at a zero-commission broker, the order is most often sold to a wholesaler and never touches a lit exchange.

Adoption, debate & evidence

Off-exchange execution is no longer a niche. Through 2024, off-exchange (TRF-reported) volume ran on the order of ~42–48% of total consolidated volume depending on the vendor and period (FINRA's own figures put TRF-reported trades around 42% of consolidated tape; some vendor measures reach the high-40s), and by January 2025 Bloomberg data reported dark/off-exchange venues briefly crossing ~51% — a frequently cited milestone meaning a majority of trades no longer print on a lit exchange. Within off-exchange flow, the large majority is wholesaler internalization (commonly cited as ~80%+ of TRF volume) rather than formal dark-pool ATSs (~17% of TRF volume per late-2024 industry data — these are estimates from market-data vendors, not exact SEC figures).

Wholesaling is highly concentrated: regulatory and press summaries report Citadel Securities and Virtu together handling on the order of 60–70% of marketable retail flow over 2017–2021. PFOF is genuinely contested. Critics (and the SEC's 2022–2023 reform proposals) argue it creates a conflict of interest and that internalized retail orders may not get the best achievable price; defenders point to studies (including SEC DERA working-paper analysis) finding PFOF venues often deliver lower effective spreads and meaningful price improvement vs. lit-exchange execution. The empirical picture is mixed and order-type-dependent — there is no settled consensus that PFOF systematically harms retail traders, nor that it is benign. PFOF has been effectively banned in the UK (treated as an incompatible inducement under FCA rules since ~2012) and the EU has outlawed it outright under MiFIR Article 39a, with the bloc-wide ban taking full effect on 30 June 2026 (Germany's temporary carve-out expiring) — itself evidence of the controversy.

Strengths & limitations

Strengths. Competition across venues has driven explicit trading costs and quoted spreads to historic lows, and the NBBO/Rule 611 framework gives even a small retail order a credible best-price guarantee. Dark pools genuinely reduce the market impact of large institutional orders.

Limitations. Fragmentation means liquidity is splintered — the displayed size at the NBBO can be a fraction of true available depth, and quotes can be fleeting. Heavy off-exchange routing means the public lit quote increasingly reflects less than half of real trading, degrading price discovery (the central concern as off-exchange share approaches/exceeds 50%). The single most common misconception is that "the NBBO is the price" or that all orders compete in one book — in reality a retail marketable order is usually internalized, and the visible tape is an aggregation of many disconnected venues.

Sources

• SEC — National Securities Exchanges directory (~24 registered, 2024); Rule 611/Reg NMS final rule (sec.gov)
• SEC 2024 Reg NMS amendments (access-fee cap cut $0.0030 → $0.0010/share, compliance Nov 3 2025) — confirmed via Sidley, Davis Wright Tremaine, WilmerHale client alerts and SEC final rule
• ESMA — MiFIR Article 39a PFOF prohibition (EU-wide ban effective 30 June 2026); UK FCA inducement treatment
• SEC DERA working paper, How Does Payment for Order Flow Influence Markets? (2025) (sec.gov)
• Congressional Research Service — Payment for Order Flow and Broker-Dealer Regulation (IF12594, IF11800) (congress.gov)
• CFA Institute — Dark Pools, Internalization, and Equity Market Quality (rpc.cfainstitute.org)
• Wikipedia — Electronic communication network, Payment for order flow, Dark pool (cross-checked against primary sources for history/definitions)
• Cboe U.S. Equities exchange overview (cboe.com); Nasdaq U.S. Equity Market Data Whitepaper (2024)
• Off-exchange share figures: market-data-vendor industry summaries and Bloomberg-reported Jan-2025 dark-pool data — flagged as estimates, not exact SEC counts