Phase Space AI

Trade Construction

Reddit [RDDT]

Reddit, Inc. [RDDT] — Trade Construction & Liquidity Criteria

As of 2026-07-29 · spot $178.345 (close 2026-07-28) · chain pulled live from Alpaca, 2026-07-28

This document constructs vehicles. It does not recommend a position, a direction or a size. The memo issues no verdict; the book decides whether to own this, on which side, and how large. What follows is the evidence a book needs to know that whatever it decides is fillable.

---

1. Liquidity Criteria (BINDING) — the chain was pulled first

The framework's rule exists because of a real failure: on HCA the maximum open interest across an entire March-2027 chain was 18 contracts, the default defined-risk spread was uninvestable at any size, and nothing tested for it. A vehicle that cannot be filled is not a vehicle. So the chain comes first.

1.1 Equity liquidity

Stated limitation, not hidden. The Alpaca feed=iex series prints only IEX-executed volume. IEX is a single venue and routinely represents a low-single-digit percentage of consolidated US equity volume, so $37.4m/day is a floor, not an estimate of consolidated ADV. No multiplier is applied here because none can be evidenced from the data pulled. The floor alone is sufficient: at $37.4m/day a $50m position is ~1.3 days of IEX-only volume and far less of the consolidated tape.

Equity Liquidity: PASS, on the floor alone.

1.2 Options chain — actual open interest and quoted size, January 2027 expiry

Every row below is a live quote from data.alpaca.markets/v1beta1/options/snapshots, 2026-07-28. Expiry 2027-01-15 (171 days), the only expiry on the board with institutional open interest.

Whole-chain totals, January 2027: call open interest 40,938 contracts across 20 strikes in the 0.6×–1.6× spot band; maximum single-strike call OI 28,823; maximum single-strike put OI 26,361. Other listed expiries: 2026-11-20 (1,393 call OI), 2026-12-18 (4,628), 2027-03-19 (3,043), 2027-06-17 (10,777).

Liquidity Criteria (BINDING): PASS. The January 2027 chain is genuinely institutional. At 28,823 contracts of open interest on a single strike, a 250-lot is under 1% of open interest.

1.3 Volatility pricing — is the option cheap or dear?

Options are priced at a 9% premium to trailing realised volatility — a normal, unremarkable variance risk premium, and notably not elevated into an imminent earnings print. There is no volatility edge here in either direction. This argues for defined-risk structures over outright long premium, and equally against any thesis that rests on selling "expensive" vol: at 1.09x it is not expensive.

---

2. Vehicles — constructed, priced at both mid and at the spread

Direction is not asserted. Both sides are constructed because the memo's job is to make either executable.

2.1 Upside expression — Jan-2027 220/270 call spread (the deepest liquidity on the board)

What it expresses. The 12-month target band's lower half ($216–$267, +21% to +50%) fully covers the $270 upper strike within the option's life only at the top of the band. This structure is therefore a tail expression, not a base-case one: it pays if the name re-rates back toward the middle of its own multiple history, and it is worth zero if the name simply grinds sideways.

2.2 Upside expression — Jan-2027 200/250 call spread (lower breakeven, still fillable)

This is the structure that maps to the base case. Its breakeven of +20.0% sits just below the 25th percentile of Reddit's own multiple history applied to NTM revenue ($215.80, +21.0%). In other words: this spread pays if Reddit does nothing more than trade at a multiple it has been above three-quarters of the time since IPO. Open interest of 2,241/1,438 supports meaningful but not unlimited size — a 100-lot is 4.5% and 7.0% of the respective strikes' open interest, which is executable; a 500-lot is not, at these quotes.

2.3 Downside expression — Jan-2027 150/115 put spread

What it expresses. The Downside Criteria bear case is $100 (−44%), driven by a named cause — further loss of US search-referred logged-out traffic — with p=0.30. The 115 strike sits above that bear case, so the structure captures the bulk of it. The $115 put is the single most heavily owned put on the board (26,361 contracts), which is itself information: the market has already built a large position at exactly the level implied by a referral shock.

2.4 Equity, for a book that simply wants the exposure

Nothing about this name requires an option. At $37.4m/day of IEX-only notional and $36bn of market cap, a position of any size a single-name book would take is executable in the common stock, and the common stock avoids paying a 9% variance risk premium for a thesis whose horizon (five years, per the implied-path test) is far longer than any listed expiry.

---

3. Sizing input — not a sizing decision

The framework's active protection is inverse-volatility sizing, and it works because a fat-left-tail name is almost always a high-volatility name and is sized down automatically. The inputs:

RDDT sizes small on volatility alone. The Downside Criteria is MEASURED and constrains nothing; the volatility does the constraining, which is the framework's stated interim control.

---

4. Invalidation — what would refute the analysis, stated in advance

These are thesis invalidations, not stop-losses. They are the events that would make the Research and Valuation documents wrong rather than merely early.

1. US total DAUq growth turns negative in any reported quarter. It is +7% today, down from +59% eight quarters ago. Negative US DAUq removes the impressions leg entirely, and the implied path needs ARPU to carry a 29%/yr revenue CAGR alone.
2. Ad price growth decelerates below ~15% YoY. Price is half the mechanism (research §1). At +32% today, a fall below 15% breaks the ARPU series that the entire 12-month band rests on.
3. A quarter in which revenue growth falls below the guided +44% trajectory without an ARPU offset. Q1'26 +69% → Q2'26 guided +44% is already a 25pp deceleration; a further step down without ARPU support is the sequence the Downside Criteria describes.
4. A large GAAP EPS beat that turns out to be a valuation-allowance release. Not an invalidation of the business, but an invalidation of any target built on that EPS. $779.0m of valuation allowance is sitting there. Read the tax note before reading the headline.

---

5. Execution notes

• Every quote above is a live snapshot from 2026-07-28 and will be stale on any trading day after it. Re-pull before working an order. Do not trade off these figures.
• Slippage is real and quantified. Paying the spread costs 14–20% of the debit on these structures. Work mid; do not lift offers on a 171-day option in a name with a 9% variance premium.
• Earnings risk is imminent and unpriced. Reddit reported Q2 on 2025-07-31 and 2024-08-06. Q2 2026 is due within days and is not in this analysis. IV at 74.5% against 68.4% realised is not pricing an event premium, which means either the market has the date wrong or the print has already happened. Verify the reporting date before entering any option position.
• Options open interest is not the same as quoted depth. OI evidences that the strike trades; it does not guarantee size on the screen at a given moment. The spread percentages in §1.2 are the honest read on immediate executability.