The PLUG POWER INC / US DOLLAR (Equity.US.PLUG/USD) price feed delivers a live PLUG price, updated in real time on Pyth Network. Prices are aggregated from first-party publishers - exchanges, market makers and trading firms pricing from their own order flow - with a confidence interval on every update. Available on 100+ blockchains.
US equities do not trade continuously. Each weekday cycle is split into four sessions with different venues, participants and liquidity: overnight (8:00 PM-4:00 AM ET, Sunday-Thursday, on alternative trading systems such as Blue Ocean ATS), pre-market (4:00-9:30 AM ET), regular hours (9:30 AM-4:00 PM ET, with opening and closing auctions on the primary listing exchange and LULD circuit breakers active) and post-market (4:00-8:00 PM ET, typically with wider spreads, and the window in which many companies release earnings). From Friday 8:00 PM to Sunday 8:00 PM ET no session runs, a 48-hour gap.
Within the regular session, trading is fragmented across many registered exchanges, dark pools and alternative trading systems, with the SIP consolidating exchange quotes into the NBBO. Each stock has a primary listing venue, usually NYSE or NASDAQ, which runs its opening and closing auctions.
Pyth US equity prices are contributed directly by first-party institutions that make markets, run venues or route order flow in the stock. Every update carries a price and a confidence interval, and consumers pull the latest price on demand instead of waiting for a scheduled push. Publisher prices are benchmarked against the mid-point at the stock's primary listing venue. On Pyth Pro, supported tickers are priced across all four sessions, and every update carries a marketSession field so risk logic can be session-aware.
Unlike crypto, which trades 24/7, US equity trading activity is fragmented across multiple venues and four distinct trading sessions within each 24-hour cycle. Each session may have different participants, liquidity profiles and venues, which directly impacts how prices are formed and how much they can move.
| Tier | Participants | Role in price formation |
|---|---|---|
| Primary listing exchanges (NYSE, NASDAQ) | Designated or registered market makers, institutional and retail flow during regular hours | Run each stock's opening and closing auctions; the reference venue for Pyth's publisher benchmark |
| Other registered exchanges (including Cboe, IEX and MEMX) | Broker-dealers, market makers | Contribute quotes to the NBBO; share of volume varies by venue and by stock |
| Dark pools and alternative trading systems (ATS) | Institutional block flow; overnight venues | Off-exchange matching; the overnight session (8:00 PM-4:00 AM ET, Sunday-Thursday) trades on alternative trading systems |
| SIP (Securities Information Processor) | Quotes from all registered exchanges | Consolidates exchange quotes into the NBBO, the best bid and offer across registered exchanges |
Venues do not contribute equally to price discovery. A price taken from one venue reflects that venue's quotes and liquidity, not the market as a whole.
| Session/Window | Time (ET) | Character |
|---|---|---|
A price gap occurs when a stock opens at a significantly different price from its previous session's close. Because each of the four sessions has a different liquidity profile and participant base, price information does not flow smoothly across boundaries. The resulting gaps are a structural feature of equity markets, not anomalies.
| Factor | Impact |
|---|---|
| Earnings | Often released after the 4:00 PM ET close, so the initial move forms in the post-market session on thinner liquidity |
| Company, sector and macro news | Contract awards, guidance changes, regulatory decisions and macro data can move a stock sharply in the overnight or pre-market sessions |
| Analyst upgrades/downgrades | Rating changes and large disclosed positions can trigger multi-day volatility. Pre-market reactions can set the tone before RTH |
| Weekend / holiday gaps | The 48-hour weekend void (Friday 8:00 PM-Sunday 8:00 PM ET) is the longest scheduled closure of a normal week. The overnight session from the Sunday open provides the first price signal |
| LULD halts | Limit Up-Limit Down circuit breakers pause trading during RTH when price moves too fast. On resumption, price can gap up or down |
| Corporate actions | Stock splits, reverse splits and delistings create planned price discontinuities. A 4:1 split drops the price 75% overnight: not a market move, but a structural repricing that must be handled explicitly |
Worked example. A company reports quarterly results after the 4:00 PM ET close. The first prices reacting to the print form in the post-market session, on thinner liquidity and wider spreads, and the next regular-session open is a gap against the prior 4:00 PM close rather than a continuation of it. If the move is large enough to trip an LULD band after the open, trading pauses: no trades, no feed updates, and a discontinuous price on resumption. Neither event is an anomaly; both are the normal mechanics of a four-session market.
Liquidity in US equities is not a single number. It is typically deepest in the regular session, when market makers quote across every exchange and the SIP consolidates exchange quotes, and it thins outside it. The windows where liquidity disappears or degrades are:
| Event | What happens |
|---|---|
| LULD halt | Trading pauses during regular hours when price breaches its band. No trades and no feed updates while halted; price free to gap on resumption |
| After-hours earnings | The initial move forms in the post-market session, on thinner liquidity and wider spreads, before the next regular open |
| Stock split | A planned discontinuity. Publishers apply the split ratio before the ex-date session opens, under the workflow in the Asset-Specific Handling section |
| Weekend void, every week | 48 hours (Friday 8:00 PM-Sunday 8:00 PM ET) with no session. Positions held through the weekend carry gap risk across the longest closure of the week |
Pyth US equity feeds aggregate first-party contributions from institutions that make prices in this market: firms with direct market access, their own pricing engines and the ability to distribute their own data, including exchanges and trading venues, banks, non-bank market makers and trading firms. Pyth does not enumerate individual publishers per feed.
The publisher set for US equities draws on four types of contributor:
Publishers are selected per session on two criteria: price freshness through the trading day, and accuracy against the primary-market benchmark described in the What Makes Pyth Data Different section.
Publisher accuracy is benchmarked against the mid-point price at each stock's primary listing venue, and publishers are held to a deviation limit against that benchmark. This anchors the aggregate to the venue that runs each stock's opening and closing auctions.
| Traditional US equities data | Pyth Network |
|---|---|
| Redistributed venue feeds, per-terminal licensed | First-party contributions from the institutions that own the data |
| A single price with no uncertainty measure | Price plus confidence interval with every update |
The Pyth price for a US equity feed is an indicative reference price derived from an aggregate of publisher submissions. It is venue-agnostic: it does not represent the price available on any particular exchange or trading system, and it is not an offer to buy or sell the security.
Always verify price freshness before use with
getPriceNoOlderThan()and a threshold appropriate to this feed's schedule. A threshold that is correct during the trading session will reject every price while the underlying market is closed; threshold logic must be schedule-aware, not a single global constant.
For US equities, thresholds are session-specific as well as schedule-specific:
getPriceNoOlderThan() with strict thresholds appropriate for the session. Regular-session prices update most frequently; extended-session prices may update at longer intervals.Confidence intervals are especially important for equities because spread and liquidity vary widely across sessions.
| Use case | Recommended approach | Formula |
|---|---|---|
| Valuing collateral (lending) | Use lower bound | μ - σ |
| Outstanding loan positions | Use upper bound | μ + σ |
| Extreme volatility (σ/μ exceeds threshold) |
Live Pyth symbology for US equities is Equity.US.{TICKER}/USD, with the feed ID, schedule and exponent in each feed's metadata. On Pyth Pro each ticker is one feed: its reference data carries a schedule for each supported session (regular, pre_market, post_market, over_night), and the marketSession field on each update names the active session. Pyth Core carried separate session feeds with the suffixes .PRE, .POST and .ON until their deprecation on 2026-06-15; Hermes lists regular-session Equity.US.* feeds only (checked 2026-11-11).
This pack applies at asset-class level to every US-listed spot equity feed. The regular-session schedule is Monday-Friday 9:30 AM-4:00 PM ET, with US exchange holidays and early closes carried in the Hermes schedule string. 24/7 Equity.Index.* feeds and ETF-based feeds with their own methodology are described separately.
Sources:
Data vintage: 2025-2026.
Links: price feeds https://app.pyth.com/explore; documentation https://docs.pyth.network.
This document is for informational and educational purposes only and does not constitute legal, financial, investment or other professional advice. Market-structure descriptions are provided for general context and are not a complete or authoritative statement of applicable law or market structure. Accuracy and performance figures reflect historical observations over specific periods and are not a guarantee of future performance or data quality. References to exchanges, venues, protocols or products are illustrative and do not constitute an endorsement. Digital assets and decentralised finance protocols carry significant risk, including total loss. Conduct independent due diligence and consult qualified professionals before acting on this information.
The Pyth US equities feed is a set of real-time price feeds for US-listed stocks, each aggregated from first-party publishers such as market makers, exchanges, trading firms and broker-dealers that trade the stock directly. Every update carries a price and a confidence interval. Because US equities trade in four sessions with different venues and participants, Pyth Pro reports the session behind every update in a marketSession field, so a consumer can apply rules for the session it cares about.
Each feed is benchmarked against the mid-point on the stock's primary listing venue. Pyth Core carries regular-session prices; Pyth Pro delivers every supported session over WebSocket and REST.
Pyth Pro covers all four US equity sessions for supported tickers: overnight (Sunday-Thursday 8:00 PM-4:00 AM ET), pre-market (4:00-9:30 AM ET), regular hours (9:30 AM-4:00 PM ET) and post-market (4:00-8:00 PM ET). Pyth Core carries regular-session prices. Each session has its own publisher requirements because the firms with order flow differ by session: retail-exposed broker-dealers often matter most after the close.
Overnight trading takes place on alternative trading systems, and the overnight session opens the trading week on Sunday evening. The one window no session covers is the weekend void, Friday 8:00 PM to Sunday 8:00 PM ET.
The primary listing venue, usually NYSE or NASDAQ, runs each stock's opening and closing auctions. US equity trading is spread across many registered exchanges, dark pools and alternative trading systems, and volume is not evenly distributed, so a price taken from one venue reflects that venue's quotes, not the market as a whole.
Pyth measures each publisher against the mid-point at the primary listing venue and holds publishers to a deviation limit against it. This anchors the aggregate to one consistent reference point for each stock.
| Overnight |
| 8:00 PM-4:00 AM |
| Sun-Thu. Trades on alternative trading systems. The overnight session opens the trading week (Sunday 8:00 PM ET) |
| Pre-Market | 4:00 AM-9:30 AM | Mon-Fri. Early reactions to overnight news. Liquidity typically builds towards the open. Many brokers accept limit orders only |
| Regular (RTH) | 9:30 AM-4:00 PM | Mon-Fri. Opening and closing auctions on the primary listing exchange. Typically the deepest liquidity and tightest spreads of the day. LULD circuit breakers are active |
| Post-Market | 4:00 PM-8:00 PM | Mon-Fri. Many companies release earnings in this window. Spreads are typically wider than in regular hours |
| Weekend void | Fri 8:00 PM-Sun 8:00 PM | No session runs for 48 hours. The Sunday overnight open provides the first price signal |
When the underlying market is closed (the weekend void, exchange holidays, and the overnight-before-ex-date halt described in the Asset-Specific Handling section), the regular-session feed produces no new prices. The last published price stays in place with an ageing timestamp, which is why staleness thresholds in the Staleness Checks section must be schedule-aware.
Different publishers participate in different sessions, so Pyth sets publisher requirements per session. On Pyth Pro, each supported US equity is a single feed whose reference data carries a schedule for every session it supports (regular, pre-market, post-market, overnight), and every update reports the active session in the marketSession field. Pyth Core carries regular-session prices; its separate session-suffixed feeds were deprecated on 2026-06-15. The naming convention is in the Symbols & Feed Metadata section.
Stocks differ in how much of this matters. Heavily traded, news-driven names with large retail and institutional followings typically see more activity outside regular hours and larger session-boundary moves; less active names may trade thinly or not at all overnight.
In traditional finance, stop-loss orders cannot protect against gaps: they execute at the next available price, not the intended stop level. For DeFi protocols using equity price feeds, the same risk applies to automated logic:
Engineers building products on equity feeds need to understand that a stock's "price" is not continuous. It varies by session, venue and liquidity regime. Any system using equity prices for mark-to-market, funding rates or liquidations should source from where genuine price discovery happens: a price that deviates significantly from the primary venue can trigger false liquidations or create arbitrage opportunities against the consumer.
Pyth Pro (formerly Lazer) is the subscription-based, low-latency delivery layer for institutional and latency-sensitive consumers, delivered over standard WebSocket and REST APIs (FIX on request). Figures below are per Pyth's published documentation as of September 2026.
| Metric | Pyth Pro |
|---|---|
| Update frequency | Configurable channels: real-time, 50 ms, 200 ms, 1 s |
| End-to-end latency | Sub-100 ms |
| Uptime SLA | >99.99% |
| Coverage | 3,000+ feeds across FX, equities, commodities, crypto, indices and rates (September 2026) |
| Chain availability | 100+ blockchains plus off-chain APIs (REST, WebSocket) |
| Confidence intervals | Included with every update |
For equities, Pyth Pro provides a unified interface that abstracts the multi-session complexity:
The marketSession field. Pyth Pro equity payloads carry a marketSession field that reports which trading session produced the update (regular, preMarket, postMarket, overNight or closed). This allows protocols to implement session-aware risk logic:
{ "priceFeedId": 42, "price": "13489000000", "confidence": "15000000", "marketSession": "regular", "exponent": -8, "publisherCount": 9 }
24/7 index feeds. Selected stocks also have a 24/7 Equity.Index.{TICKER}/USD feed that continues pricing outside the four sessions. Each of those feeds carries its own methodology description.
| Pause new activity for that feed |
| - |
| Session transition | Widen the confidence band (example: 2σ) | μ ± 2σ |
| Overnight / low-liquidity | Apply wider haircuts to collateral value | μ - 2σ |
Unlike pricing gaps caused by microstructure phenomena or macro shocks, corporate actions produce planned price discontinuities. A stock split, reverse split or delisting is announced in advance, and the price change on the ex-date is structural, not driven by supply and demand. If a price feed does not handle these events explicitly, the result is a serious mispricing: a 4:1 split looks like a 75% crash to any system treating the raw price as a market move.
The ex-date is the first trading day on which the share trades excluding the right in question.
Pyth continuously monitors corporate actions. When a split or reverse split is announced, a structured change-management process is initiated across all publishers contributing to the affected feed.
| Phase | Timing | What happens |
|---|---|---|
| Up to 4 weeks before the ex-date | N/A | Reference data produced by Pyth reflects the corporate action information, enabling users to initiate their own change-management processes to adjust positions, collateral and risk parameters ahead of the ex-date |
| Overnight halt | 8:00 PM-4:00 AM ET (night before ex-date) | The overnight venue halts trading on the affected ticker. No feed updates are generated. A market schedule holiday is added via governance to avoid producing any new price while the market is closed |
| Publisher adjustment | Before 4:00 AM ET on ex-date | Publishers apply the split ratio to their pricing engines. Any publisher not ready must stop publishing to avoid corrupting the aggregate. Non-compliance exposes publishers to slashing risk |
| Rejection window | 4:00 AM ET, for up to 12 hours | The aggregator computes the expected post-split price (last_price × denominator / numerator) and rejects any publisher whose price deviates from the expected value. Only correctly adjusted prices enter the aggregate |
| Normal aggregation | After the rejection window expires | Standard aggregation resumes |
Pyth Pro implementation: corporate actions are stored as feed metadata via governance instructions (updateFeedMetadata with the corporate_actions key). The schema includes the action kind (split or reverse_split), the adjustment factor as numerator/denominator, the activation type (us_equity_ex_date with the ex-date) and configurable rejection thresholds. The process operates at the aggregator level, where per-feed state and last_price persistence let rejection hold across restarts.
Users can select favourable historical prices within staleness constraints. Guard against this with strict
getPriceNoOlderThan()thresholds calibrated to the feed and the session, and gate on confidence width in addition to age.
For US equities this is most exploitable at the four session boundaries (4:00 AM, 9:30 AM, 4:00 PM and 8:00 PM ET), across the weekend void, during and immediately after LULD halts, and in the overnight window before a corporate-action ex-date, when the last published price can be many hours old and materially different from the next one.
Access. Pyth Pro: WebSocket, REST — request a demo key at data@dourolabs.xyz.
A gap is when a stock reopens at a materially different price from its last close. It happens because price information does not flow smoothly across session boundaries: earnings land after the close, company news lands overnight, and the 48-hour weekend accumulates information that is priced all at once on Sunday evening. LULD halts during regular hours add mid-session gaps on resumption.
Stop-losses do not protect against gaps, and neither does a naive liquidation engine. Tighten collateral or leverage 24-48 hours before earnings, require higher margin through weekends, apply session-specific risk parameters, widen confidence bands across session transitions, and freeze activity when updates stop mid-session.
A split is a planned discontinuity: a 4:1 split cuts the quoted price 75% on the ex-date with no market move behind it. Pyth runs a structured change-management process for each announced split or reverse split. Up to four weeks ahead, reference data reflects the action so integrators can adjust positions and risk parameters.
On the night before the ex-date, the overnight venue typically halts the ticker and a schedule holiday is added via governance, so no prices are produced. Before 4:00 AM ET publishers apply the split ratio; any publisher not ready must stop publishing. For up to 12 hours after 4:00 AM the aggregator computes the expected post-split price and rejects any contribution that deviates from it.
Pull the latest price on demand and verify it before use. Call getPriceNoOlderThan() with a threshold that matches the feed's schedule and session: regular-session prices update most frequently, extended-session prices may update less often, and while the market is closed no threshold will pass, so the logic must be schedule-aware rather than a single constant. Gate on confidence width as well as age, using the lower bound to value collateral and the upper bound for outstanding loans, and widen thresholds around session transitions.
On Pyth Pro, read the marketSession field to apply per-session risk parameters and pause new activity for a few seconds either side of a boundary. Prices are fixed-point: multiply by 10^exponent, so 13489000000 with exponent -8 is $134.89. Request a demo key at data@dourolabs.xyz.