# How OpenAI Perps Work

> Why OpenAI can trade as a perp even without a public listing: the market only needs a credible oracle, and Nasdaq Private Market is the right reference layer.

- Canonical: https://www.liquid.trade/learn/how-openai-perps-work
- Published: 2026-03-21
- Category: Pre-IPO Perps
- Tags: Pre-IPO, Perps, Markets

---

OpenAI is one of the largest private AI companies — but it still does not trade like AAPL or NVDA on a public exchange. There is no ordinary listed OpenAI common stock market with a continuous exchange tape available to everyone.

That does not prevent a perpetual futures market from existing. A perp does not require physical delivery of OpenAI shares. **It requires a credible reference price, a mark-price methodology, a liquidation engine, and a funding mechanism** that keeps the traded market anchored to the reference over time.

<StatBar
  stats={[
    { value: "$852B", label: "Post-money valuation", caption: "Round closed Mar 31 2026" },
    { value: "$122B", label: "Committed capital", caption: "Largest private raise on record" },
    { value: "3×", label: "Max leverage", caption: "Pre-IPO risk tier" },
    { value: "Private", label: "Listing status", caption: "IPO chatter ≠ live tape" }
  ]}
/>

## Current OpenAI Context

As of the May 26, 2026 update, OpenAI remains a private-market exposure rather than a normally listed public equity.

<KeyMetric
  value="$852B"
  label="March 31, 2026 funding round · post-money"
  tone="neutral"
  detail="OpenAI announced on March 31, 2026 that it closed a funding round with $122 billion in committed capital at an $852 billion post-money valuation. For a name this large, the oracle methodology should continuously incorporate the best available private-market and company-specific information."
/>

Reports about a possible IPO also need to be handled carefully. A confidential filing, reported preparation, or expected listing is **not** the same as a public stock already trading. Until exchange-listed shares actually trade, OpenAI exposure remains a private-market reference problem.

## A Perp Does Not Need Share Delivery

A perpetual futures market is a derivative. When someone trades an OpenAI perp, they are not buying OpenAI shares, joining the cap table, or receiving shareholder rights.

<CompareTable
  caption="What the perp needs vs. what it doesn't"
  columns={["", "Doesn't need", "Does need"]}
  rows={[
    { label: "Share custody", values: ["✕ public-share settlement", "Index or oracle price"], emphasis: [1] },
    { label: "Cap-table access", values: ["✕ transfer-agent processing", "Mark price for PnL + liquidations"], emphasis: [1] },
    { label: "Tender offers", values: ["✕ tender-offer participation", "Risk parameters per asset"], emphasis: [1] },
    { label: "Primary access", values: ["✕ direct access to primary shares", "Orderly trading liquidity"], emphasis: [1] },
    { label: "Governance", values: ["✕ voting rights", "Funding logic to manage deviations"], emphasis: [1] }
  ]}
/>

They are taking synthetic long or short exposure to a reference price.

## Why the Oracle Is the Product

For private-company perps, the oracle is not a side detail — it is the **core of the market**.

A credible oracle methodology should prefer institutional private-market data over social-media screenshots or isolated broker chatter. [Nasdaq Private Market](https://www.nasdaqprivatemarket.com/) and its [Tape D data products](https://www.nasdaqprivatemarket.com/product/tape-d-financial-data-products/) are built for private-company price discovery: secondary-market signals, reported trade levels, bid/offer history, 409A valuations, mutual fund marks, primary rounds, and related market intelligence.

The point is not that any single private-market datapoint is perfect. The point is that a defensible OpenAI oracle should **triangulate** across multiple signals and publish a methodology users can understand.

## What the Oracle Should Consider

<StepFlow
  caption="Inputs a credible OpenAI reference methodology weighs"
  steps={[
    { title: "Recent primary-round valuation context", body: "The most recent priced equity sale anchors the order of magnitude — but on its own, it can be stale within months." },
    { title: "Secondary-market bid and offer indications", body: "Where dealers stand right now. The most timely signal, but indicative rather than executable in size." },
    { title: "Reported secondary trade levels where available", body: "Real shares changing hands at known prices. Thin, episodic, but high signal." },
    { title: "Mutual fund marks and other institutional marks", body: "Institutional NAVs report dated valuations. Disciplined and recurring; smoothed by accounting policy." },
    { title: "409A valuation context where relevant", body: "Independent appraisal that regulates option grants. Lags fast moves." },
    { title: "Share-class differences and liquidation preferences", body: "Common vs. preferred can imply different values. The oracle has to normalise — not pretend every print is identical." },
    { title: "Major company events", body: "Financings, restructurings, IPO milestones, partnership disclosures." },
    { title: "Liquidity and staleness of each input", body: "Weight signals by how recent and how thick they are — not just by their face value." }
  ]}
/>

Private-company prices are not as clean as public exchange prices. Different share classes, transfer restrictions, information rights, and transaction sizes can produce different implied values. **The oracle has to normalise those differences rather than pretending every private-market print is identical.**

## How OpenAI Perps Trade Between Oracle Updates

Private markets do not print every second like listed equities. That is normal.

<Callout variant="key">
The oracle provides the anchor. The Liquid market provides continuous price discovery around that anchor. If traders push the perp materially above the reference, funding can make longs pay shorts. If they push it below, shorts pay longs. **Same broad mechanism as every other perp market: traded price moves continuously, while funding and mark-price logic keep the contract tied to a reference.**
</Callout>

## What Traders Get

<ScenarioCompare
  caption="Why the order book has two sides"
  scenarios={[
    {
      label: "Bull thesis",
      headline: "Long OPENAI",
      tone: "long",
      body: "ChatGPT keeps growing · enterprise API revenue compounds · multimodal moat widens · next round prints higher · IPO narrative pulls private marks up.",
      metrics: [
        { k: "Catalyst", v: "Revenue acceleration" },
        { k: "Reference cue", v: "Up-round leaks" }
      ]
    },
    {
      label: "Bear thesis",
      headline: "Short OPENAI",
      tone: "short",
      body: "Margin compression from compute costs · governance instability · open-weight competition · agentic platform fragmentation · IPO at a discount to last mark.",
      metrics: [
        { k: "Catalyst", v: "Margin / governance" },
        { k: "Reference cue", v: "Mutual fund mark cuts" }
      ]
    }
  ]}
/>

A perp is a **trading instrument**, not private-share ownership.

## The Main Risks

<Callout variant="warn" title="Six risks worth pricing in">
**Oracle lag** — private-market data can update slower than the traded perp market. **Valuation dispersion** — different share classes and transaction contexts can imply different values. **Liquidity risk** — a market can be directionally right but hard to exit during stress. **Funding risk** — funding can become expensive if one side is crowded. **Event risk** — funding rounds, lawsuits, governance changes, strategic partnerships, or IPO developments can move the reference quickly. **IPO transition risk** — if OpenAI eventually lists publicly, the reference methodology should migrate from private-market inputs toward public-market pricing; understand how that transition is handled.
</Callout>

## Further Reading

- [OpenAI: March 31, 2026 funding announcement](https://openai.com/index/accelerating-the-next-phase-ai/)
- [Nasdaq Private Market: Tape D data products](https://www.nasdaqprivatemarket.com/product/tape-d-financial-data-products/)
- [Liquid docs: funding rates](https://docs.liquid.trade/trading/funding-rates)
- [Liquid docs: liquidations](https://docs.liquid.trade/trading/liquidations)
- [How Anthropic Perps Work](/learn/how-anthropic-perps-work)
- [How SpaceX Perps Work](/learn/how-spacex-perps-work)

OpenAI perps make sense when you understand the distinction: you are not buying OpenAI stock — you are trading synthetic exposure to a private-market reference.