Reddit RDDT
Communication Services · Internet Content & Information · Synthos Deep Dive · 2026-08-04
$154.71
Stage-In — buy a partial position into a -43% drawdown that has just been contradicted by the numbers (Q2'26 revenue +61% YoY, a 32% beat to the consensus EPS line), but stage the rest, because the de-rating is being driven by a structural question about AI-era search referral that the reported quarter does not answer
—
—
The Overview
Reddit is a website made of thousands of user-run forums. It makes money almost entirely by selling ads next to those conversations — $2.06 billion of its $2.20 billion in FY25 revenue was advertising. Because users write all the content for free, the cost of delivering it is almost nothing: 91 cents of every revenue dollar is gross profit, and last year the company spent just $6.7 million on physical equipment while generating $684 million of free cash flow. Two years ago it lost $484 million; last quarter it earned $253 million. That is one of the faster swings from loss to profit you will see in a company this size.
So why is the stock down 43% from its high? Not because of the numbers — the most recent quarter beat what analysts expected by about a third. It is down because of a worry about where the visitors come from. A large share of Reddit's audience arrives through search engines. If AI assistants start answering questions directly instead of sending people to Reddit threads, the audience shrinks and so does the ad business. There is a second, related worry: as AI-written text spreads across the internet, the thing that makes Reddit valuable — that a real person wrote it — gets harder to guarantee. One of the sources we track reported in April that roughly 10% of Reddit content is already AI-written.
The bull answer is that the same trend cuts the other way. AI models need human-written training data, and Reddit has more of it than almost anyone. Reddit has already licensed that data (a $60 million non-exclusive deal with Google is the known example), and as the rest of the web closes off, private, human-authored corpora become more valuable, not less. Right now that revenue line is small — $140 million of $2.20 billion — so it is a promise rather than a proof.
- Downside Risk 6/10. No solvency question at all, but a very high-beta stock (1.938) whose key operating variable we cannot see.
- Growth Quality 9/10. +61% revenue growth at 91% gross margin with margins expanding — about as clean as growth gets.
- Exponential Potential 7/10. Real optionality in data licensing and international monetization; not yet visible in the reported segments.
Putting a number on it: our fair-value estimate is $195 against a current price of $154.71 — real upside if our numbers are right.
Our summary metrics
Downside Risk (lower = safer)
6/10 · High
Not a balance-sheet risk — $2.48B cash and short-term investments versus $23.2M of total debt, a 10.5x current ratio, $684M of FY25 free cash flow on $6.7M of capex, and no dividend or buyback obligation. The 6 is entirely about (a) beta 1.938 and a -43% drawdown that shows how violently this trades, (b) genuine platform dependency — the file gives no daily-user disclosure at all, so the single most important operating variable is UNOBSERVABLE in our data, (c) stock-based compensation at 12.2% of revenue, and (d) an ownership structure the profile describes as a controlled subsidiary of Advance Publications, which limits minority-holder leverage. Rated 6 rather than 4 because the drawdown risk is demonstrably large even with a fortress balance sheet.
Growth Quality
9/10 · Very High
Among the highest-quality growth profiles in the entire universe. Q2'26 revenue $804.9M, +61.1% YoY; Q1'26 +69.1%; gross margin 91.2%; operating margin swung from -43.1% (FY24) to +20.1% (FY25) to 28.8% in Q2'26. Advertising was $2.06B of $2.20B FY25 revenue, +74% YoY, and non-US revenue grew +76% off a small base — international is a second, undermonetized leg. This is operating leverage arriving all at once on a fixed cost base. Capped at 9 rather than 10 only because consensus itself models deceleration to ~32% (FY27) and ~25% (FY28), and because we cannot verify user growth from this data.
Exponential Potential
7/10 · High
Two real exponential vectors. First, the monetization gap: advertising ARPU on a platform with this much first-party intent data is early, and 28.8% operating margins on 91% gross margins mean incremental revenue drops through at extraordinary rates. Second, data licensing — the knowledge base names this directly (latent_space, 2024-03-29: the Reddit-Google $60M non-exclusive deal, "data is the second most expensive input after GPUs"; macrovoices, 2024-05-02: expensive APIs restricting scraping "pushing value to private data"). Reddit sits on the single largest corpus of unstructured human opinion on the open internet. Held to 7 rather than 9 because "Other Revenue" — the line that would carry licensing — was only $140.0M of $2.20B in FY25, so the exponential is asserted by the panel and not yet visible in the segment disclosure.
Fair value$195 $105–$265
What does “fair value” mean?
Fair value is Synthos’s estimate of what one share is worth today, based on our model of the company’s future cash generation and the risks to it. It is not a price target or a prediction of where the stock trades next quarter — it’s the price where we think risk and reward are balanced. Above it, you’re paying for outcomes better than our base case; below it, the market is offering a margin of safety.
The model’s inputs and weightings are proprietary — they’re the product. What we publish is the output, We don’t publish a reverse-DCF cross-check for pre-profit companies — negative or missing earnings break that math — so take this number on our modeling alone.
What do the 5 tiers mean? (Core · Tactical · Watch · Hold · Avoid)
Buy — CoreOwn it as a foundation — start or add now, size it for years, let dips be gifts.
Buy — TacticalGood price + confirmed trend + a defined exit — buy the setup, not a marriage.
WatchWe want the business, just not at this price/setup — act only when the listed trigger hits.
HoldFine to keep if you own it — no reason to buy more; new money does better elsewhere.
AvoidDon't own it — the problem is the business or the expectations, so a cheaper price won't fix it.
The Road Ahead
What we expect to matter in each window, and the evidence that would prove us wrong.
Short term 0-6 months
No differentiated view- Driver
- A violent, unresolved technical setup. $154.71 is -42.9% from the $270.71 52-week high, ~12% below the 50-DMA ($175.4) and ~14% below the 200-DMA ($179.9), with RSI at 27.3 — deeply oversold. But the stock is +9.98% TODAY, five sessions after a Q2 print (2026-07-30) that beat the consensus EPS line by 32% ($1.25 vs $0.949) and the revenue line by 10.1% ($804.9M vs $731.0M). A monster beat that leaves the stock 43% below its high tells you the de-rating is thesis-driven, not results-driven; a 10% up-day tells you positioning was extremely short into it. Both facts can be true and both matter for entry.
- What we’re watching
- Whether the bounce holds above the recent low and reclaims the 50-DMA (~$175); whether the 12 months of insider selling continues (COO Jennifer Wong sold ~26,000 shares on 2026-07-15 at $190-198, i.e. materially above today's price, all option-exercise-and-sell with zero open-market buying anywhere in the file); and any disclosure at all touching daily active users or search-referral mix, which is the actual bear case.
- Confidence
- Low
Medium term 6-24 months
Tailwind- Driver
- Consensus models revenue $3.340B (FY26E, +51.6%) to $4.403B (FY27E, +31.8%) to $5.497B (FY28E, +24.9%) on 21-26 revenue analysts — deep, credible coverage. EPS runs $5.25 (FY26E) to $6.95 (FY27E) to $8.82 (FY28E), which compresses the multiple from 29.5x to 22.3x to 17.5x at an unchanged price. Reddit does not need multiple expansion to work over this window; it needs the earnings to arrive, and the last four quarters have arrived ahead of schedule every time.
- What we’re watching
- Advertising revenue growth staying above ~35%; international mix continuing to outgrow US (non-US was $417.0M of $2.20B in FY25, +76% YoY); operating margin holding above ~25%; and whether "Other Revenue" (the data-licensing proxy, $140.0M in FY25) inflects — that line turning into a real growth driver is the cleanest confirmation the panel's licensing thesis is monetizing.
- Confidence
- Medium
Long term 2+ years
No differentiated view- Driver
- The long-run case is that a durable, human-authored community corpus becomes MORE valuable, not less, as generated text floods the open web — scarcity value in a world of abundance. Our own knowledge base is the best statement of both sides of this: we_study_billionaires (2025-12-04, conviction 66) frames Reddit as "an advertising business monetizing a massive network effect as the internet's front page," while odd_lots (2026-04-02, conviction 70, bearish) reports that "about 40% of internet pages are now AI-generated... and ~10% of Reddit is AI-written," and (conviction 65) that bot-farm startups now sell "organic" Reddit brand mentions precisely because Reddit ranks in search and feeds training data. The same property that makes Reddit valuable makes it a target for contamination.
- What we’re watching
- Any measurable degradation in content authenticity or moderation efficacy; the structure of AI-answer products and whether they cite-and-link (traffic-positive) or absorb-and-summarize (traffic-negative); and whether data-licensing contracts renew at higher rates, which would convert the corpus into an annuity rather than a traffic funnel.
- Confidence
- Low
Exponential Potential
Exponential Potential
7/10 · High
Two real exponential vectors. First, the monetization gap: advertising ARPU on a platform with this much first-party intent data is early, and 28.8% operating margins on 91% gross margins mean incremental revenue drops through at extraordinary rates. Second, data licensing — the knowledge base names this directly (latent_space, 2024-03-29: the Reddit-Google $60M non-exclusive deal, "data is the second most expensive input after GPUs"; macrovoices, 2024-05-02: expensive APIs restricting scraping "pushing value to private data"). Reddit sits on the single largest corpus of unstructured human opinion on the open internet. Held to 7 rather than 9 because "Other Revenue" — the line that would carry licensing — was only $140.0M of $2.20B in FY25, so the exponential is asserted by the panel and not yet visible in the segment disclosure.
What could take this further than the base case — and where the market may be underpricing it. Full forward-growth and acceleration math in Deeper analysis, §4 below.
Deeper analysis
Technicals, fundamentals, valuation, and the full expert-claim evidence panel — the detail behind the numbers above.
Reference table
| Street consensus | $207.05 target (high $265 / low $110) · grades 1 strong-buy, 18 buy, 10 hold, 0 sell |
| Valuation | profitable · P/E ~36x TTM GAAP diluted / 29.5x FY26E / 22.3x FY27E / 17.5x FY28E · EV/Sales 9.8x TTM |
| Conviction | Moderate — 8 usable KB claims across 6 voices, but genuinely two-sided |
| Technicals | -42.9% from the $270.71 52-wk high, below 50-DMA ($175.4) and 200-DMA ($179.9), RSI 27.3, MACD -4.94, beta 1.938 |
| Position sizing | Consumer-internet / attention-monetization sleeve. 1.5-3% core, built in 3 tranches |
What the experts actually said 1 traceable claims on RDDT · showing the highest-conviction voices
“Reddit is an advertising business monetizing a massive network effect as the internet's front page; it was at one point a triple for us.”
Every claim reconciles to a real claim_id in the Synthos knowledge base — this is the evidence the verdict is built on, not vibes. Management (the company itself) is shown but half-weighted; one cautionary voice is included on purpose.
Price & moving averages 12 months · 50 & 200-day averages · 52-week range
Solid line = price · dashed line = 50-day average · dotted line = 200-day average · the two thin horizontal lines mark the 52-week high and low. Price above both averages is an uptrend.
Data summary: last close $159.85, 9% below the 50-day average ($176), 11% below the 200-day average ($180) — a downtrend. 41% below the 52-week high of $271, 31% above the 52-week low of $122.
Bollinger Bands 20-day average ± 2 standard deviations
The shaded band widens when the stock gets more volatile. Riding the upper edge = strong momentum (sometimes stretched); the lower edge = weak / potentially oversold.
Data summary: price $159.85 is currently inside the band (band $148–$212).
RSI (14) momentum gauge · 0–100
Above 70 (overbought zone, shaded) = overbought, below 30 (oversold zone, shaded) = oversold. Currently 42.
MACD 12 / 26 / 9 · trend & momentum
The MACD line crossing above the signal line (bars flip to the up color) = momentum turning up; crossing below (bars flip to the down color) = turning down. Bar height = the size of that gap.
Data summary: MACD is currently below its signal line by 3.77, negative momentum.
Relative performance vs S&P 500 & its sector (XLC (sector)), set to 100 a year ago
Solid = RDDT · dashed = S&P 500 · dotted = XLC (sector). A rising line means it is beating that benchmark — the sector line shows whether it is a leader or laggard within its own group.
Forward revenue & earnings actual → estimate · "FY" = fiscal year, "E" = estimate
Darker bars = actual results, brighter = analyst estimates. Taller bars to the right = expected growth.
Key stats an RIA wants
Price$154.71
Market cap$30B
P/E trailing34×
P/E FY26E / FY27E29× / 22×
EV / Sales9.8×*
EV / EBITDA30.6×*
Gross margin91.4%
Net margin31.3%
Dividend yield0.00%
Beta1.938
52-wk range$122 – $271
RSI(14)27
50 / 200-DMA$175 / $180
12-mo return+-4% (SPY +20%)
Street target$207 ($110–$265)
Analyst grades18 Buy · 10 Hold · 0 Sell
FMP ratingB
Next earnings2026-10-29 (Q3'26 earnings; consensus EPS $1.27, revenue ~$867.5M — implying +48% YoY. Reddit has beaten the consensus EPS line in four of the last five reported quarters, in two of them by more than 30%)
1. What they actually sell — the segments, with real numbers
Reddit's disclosure is thin — two product lines and two geographies — which is itself worth knowing. FY25 total revenue was $2.203B, +69.5% over FY24's $1.300B.
| Line (FY25) | Revenue | FY24 | YoY | Share of total |
|---|---|---|---|---|
| Advertising | $2,062.5M | $1,185.5M | +74.0% | 93.6% |
| Other Revenue | $140.0M | $114.7M | +22.0% | 6.4% |
| Geography (FY25) | Revenue | FY24 | YoY | Share |
|---|---|---|---|---|
| United States | $1,785.6M | $1,063.6M | +67.9% | 81.1% |
| Non-US | $417.0M | $236.6M | +76.2% | 18.9% |
Three things fall straight out of this table.
1. This is an advertising company, full stop. 93.6% of revenue. Any thesis about Reddit is a thesis about ad load, ad pricing, and audience.
2. "Other Revenue" is where data licensing lives, and it is small and slow. $140.0M, growing +22% — a fraction of the advertising line's +74%. The knowledge base's licensing thesis (latent_space, macrovoices, ml_street_talk) is real as a strategic idea, but it is not currently a material earnings driver, and honesty requires saying so plainly. If licensing is the reason you own this, you are underwriting 6% of revenue.
3. International is the undermonetized leg. Non-US grew faster than US (+76.2% vs +67.9%) off a base less than a quarter the size. Reddit's user base has always been more global than its revenue; that gap closing is a multi-year revenue source that requires no new product.
What is NOT in this file, and it matters: there is no daily-active-user disclosure, no engagement metric, no traffic-source breakdown, and no advertiser-count or ARPU line. For an advertising business, those are the four numbers that determine everything. We are underwriting the P&L without the operating metrics behind it. That is stated here rather than buried, and it is the reason Downside Risk is 6 and not 4.
2. The profit inflection — what actually happened over eight quarters
This is the strongest part of the story and the part the drawdown ignores.
| Quarter | Revenue | YoY | Gross margin | Operating income | Op margin | Net income | Diluted EPS |
|---|---|---|---|---|---|---|---|
| Q2'24 | $281.2M | — | 89.5% | -$31.0M | -11.0% | -$10.1M | -$0.06 |
| Q3'24 | $348.4M | — | 90.0% | $6.9M | 2.0% | $29.9M | $0.17 |
| Q4'24 | $427.7M | — | 92.6% | $52.9M | 12.4% | $71.0M | $0.39 |
| Q1'25 | $392.4M | — | 90.5% | $3.9M | 1.0% | $26.2M | $0.14 |
| Q2'25 | $499.6M | +77.7% | 90.8% | $67.7M | 13.6% | $89.3M | $0.48 |
| Q3'25 | $584.9M | +67.9% | 91.0% | $138.5M | 23.7% | $162.7M | $0.80 |
| Q4'25 | $725.6M | +69.7% | 91.9% | $231.8M | 31.9% | $251.6M | $1.24 |
| Q1'26 | $663.4M | +69.1% | 91.5% | $182.9M | 27.6% | $204.0M | $1.01 |
| Q2'26 | $804.9M | +61.1% | 91.3% | $231.7M | 28.8% | $252.8M | $1.25 |
Read the middle two columns together. Gross margin has been pinned between 89.5% and 92.6% for two years — this is a structurally fixed-cost business. So every incremental revenue dollar arrives with ~91 cents of gross profit, and because operating expenses grow far slower than revenue (R&D went from $142.8M to $231.3M, +62%, while revenue went from $281.2M to $804.9M, +186%), operating margin has gone from -11.0% to +28.8% in eight quarters. That is not cost-cutting; that is scale meeting a fixed cost base.
Full-year picture. FY25 revenue $2.203B (+69.5%), operating income $442.0M (20.1% margin) versus FY24's -$560.6M (-43.1%), net income $529.7M versus -$484.3M. The FY24 loss is almost entirely a post-IPO stock compensation artifact — FY24 stock-based compensation was $801.6M, versus $343.2M in FY25 — which is why FY24 operating cash flow was still positive at $222.1M despite a $484M reported loss.
Cash generation is real. FY25: operating cash flow $690.9M, capital expenditure $6.7M, free cash flow $684.2M — a 31.1% free-cash-flow margin. Capex is 0.3% of revenue. There is no factory, no fleet, no content budget. This is as capital-light as public companies get.
3. The balance sheet — nothing to discuss, which is the point
At FY25 year-end: cash $953.6M, short-term investments $1,523M, total $2,477M; total debt $23.2M; total current liabilities $271.3M; shareholders' equity $2,929M.
- Net cash ≈ $2.45B, about 8.2% of the market capitalization.
- Current ratio 10.5x. There is no working-capital risk.
- Note the FY23 comparison: total liabilities were $2.009B and equity was negative $412.9M pre-IPO. The IPO and two years of profits rebuilt the entire capital structure. Stock issuance contributed $600.0M in FY24.
Our computed enterprise value is $27.33B (market cap $29.78B less $2.45B net cash), against the vendor's $28.32B — the vendor nets only the $953.6M cash line and ignores the $1,523M of short-term investments. All EV multiples in this dive use our figure.
4. Valuation — priced in or room?
At $154.71 (market cap $29.78B, EV ~$27.33B), TTM revenue is $2.779B (Q3'25 through Q2'26) and TTM diluted EPS is $4.30.
| TTM | FY26E | FY27E | FY28E | FY29E | FY30E | |
|---|---|---|---|---|---|---|
| Consensus revenue | $2.779B | $3.340B | $4.403B | $5.497B | $6.094B | $7.357B |
| YoY growth | — | +51.6% | +31.8% | +24.9% | +10.9% | +20.7% |
| Consensus EPS | $4.30 | $5.25 | $6.95 | $8.82 | $10.33 | $14.34 |
| P/E at $154.71 | 36.0x | 29.5x | 22.3x | 17.5x | 15.0x | 10.8x |
| Revenue analysts | — | 21 | 26 | 22 | 15 | 11 |
6a. What today's price assumes (the inversion)
At $154.71 — 22.3x FY27E consensus EPS and 9.8x TTM EV/Sales — the price embeds roughly the following falsifiable claims:
- Revenue compounds at ~35% CAGR from the TTM $2.78B to the FY28E $5.50B, i.e. Reddit adds ~$2.7B of annual revenue in two and a half years without a step-down in ad pricing. (Source: consensus, 22-26 analysts.) Falsifier: any quarter printing below ~+30% YoY revenue growth before Q4'27.
- Operating margin holds at or above ~28% and grinds higher, delivering the FY28E $8.82 EPS on $5.50B of revenue — implying roughly a 32% net margin, above the current 31.3% TTM net margin. (Consensus-implied, our arithmetic.) Falsifier: operating margin printing below 24% in any quarter absent a one-time charge.
- The market keeps paying only ~22x forward — i.e. NO re-rating is required for the base case. At an unchanged price, the multiple mechanically rolls from 29.5x (FY26E) to 17.5x (FY28E). The price is currently assuming that the rate of deceleration justifies a market multiple for a 60% grower. Falsifier: a sustained move above ~28x forward would mean the crowd has resolved the search-referral question in Reddit's favour.
- THE MOST FRAGILE ASSUMPTION: that audience reach is stable. Every bullet above rests on the number this data file does not contain. If search-driven arrival narrows materially, the revenue path breaks first, the margin path breaks second (advertising deleverages fast), and both the E and the multiple compress together. Falsifier: any company disclosure of flat or declining logged-in/daily users, or a step-down in US revenue growth while non-US growth holds — the signature of a US-search-referral problem rather than a demand problem.
6b. The return bridge (why the multiple moves)
Expected return ≈ EPS growth + multiple drift + shareholder yield.
The base case at $195 is built as 22.3x — today's exact FY27E multiple — applied one year forward to the FY28E consensus EPS of $8.82 ($8.82 × 22.1 = $195). This explicitly assumes NO multiple expansion. The entire base-case return is earnings growth: consensus EPS rises 27% from FY27E to FY28E, and the price follows it at a constant forward multiple. Shareholder yield is zero — no dividend, no buyback (FY25 repurchases: $0), and stock compensation at 12.2% of revenue is a modest dilution headwind partially offset by the fact that diluted share count has been nearly flat (202.9M in Q4'25 to 202.0M in Q2'26).
Note the guardrail: the roll from 29.5x FY26E to 17.5x FY28E at a constant price is mechanical, not a de-rating. The actual de-rating already happened — it is the 43% drop from $270.71, which took the FY27E multiple from roughly 39x to 22.3x. The bull case requires expansion; the base case does not. That is the fragile-leg distinction that matters here.
6c. Variant perception (where we differ, what would surprise)
- We think the market is pricing a traffic shock as more likely than the reported numbers support — but we cannot prove it, and we say so. Consensus targets average $207 (+34%) with 19 of 29 analysts at buy or better, so the analyst community broadly shares our view; the price does not. When the sell side and the price action disagree this sharply, the price action is usually pricing a risk the models cannot quantify. Here that risk is identifiable: referral dependency. Our variant perception is narrow — that Q2'26's +61% growth and 28.8% margin deserve more weight than they are receiving, not that the structural risk is fake.
- Positive surprise that would force a repricing: "Other Revenue" (the licensing proxy) inflecting from +22% toward +50%+, or any new named licensing agreement. This would convert Reddit from a traffic-dependent ad business into a corpus-annuity business, and it is the single cleanest resolution of the bear case. Watchable at: the FY26 revenue disaggregation, and the Q3'26 print on 2026-10-29.
- Negative surprise: a Q3 or Q4 revenue print below ~+30% YoY, particularly with US decelerating faster than non-US. Consensus needs $867.5M in Q3'26 (+48%); anything with a 3 in front of the growth rate would be read as the referral thesis arriving. Watchable at: 2026-10-29.
- The other negative surprise, sourced directly from our knowledge base: measurable content-authenticity degradation. odd_lots reported (2026-04-02, conviction 65) that bot-farm startups now sell "organic" Reddit brand mentions specifically because Reddit ranks in search and feeds model training. If platform integrity visibly erodes, both the ad business and the licensing business degrade at the same time — the one scenario where Reddit's two revenue engines are correlated rather than offsetting.
Synthos fair values
- Bear ~$105 (-32%). The referral thesis arrives: revenue growth decelerates toward ~20%, FY27 EPS lands near the analyst low of $4.38, and the market pays 24x for a decelerating advertising asset. $4.38 × 24 = $105. This brackets the street's low target of $110 and sits ~14% below the 52-week low of $121.84.
- Base ~$195 (+26%). Consensus executes; FY28E EPS $8.82 at today's unchanged 22.1x forward multiple. No re-rating assumed.
- Bull ~$265 (+71%). Growth holds above 30% into FY28, licensing inflects, and the multiple re-rates to 30x on FY28E $8.82 = $265. This is the street's high target and still ~2% below the 52-week high — i.e. the bull case is a recovery, not a new regime.
5. Knowledge base — a genuinely two-sided lane
8 usable claims across 6 tracked voices. This is a real lane, but it is not a conviction pile-on, and it required careful filtering.
The bullish side — network effects and the value of the corpus:
- we_study_billionaires (2025-12-04, conviction 66, bullish, entity-tagged RDDT): "Reddit is an advertising business monetizing a massive network effect as the internet's front page; it was at one point a triple for us." This is the only claim in the file tagged to the ticker rather than the platform, and it is the cleanest statement of the bull case.
- latent_space (2024-03-29, conviction 70, bullish): "Data wars are heating up; licensing (Reddit-Google $60M, non-exclusive) becomes table stakes — data is the second most expensive input after GPUs."
- latent_space (2024-03-29, conviction 50, bullish): "Reddit's data-licensing angle is more interesting to investors than its consumer business; stock up ~40% on IPO day."
- macrovoices (2024-05-02, conviction 70, neutral): "Public training data is a limiting factor; the internet is going dark — lawsuits (Getty vs Midjourney) and expensive APIs (Twitter, Reddit) restrict scraping, pushing value to private data."
- ml_street_talk (2025-11-03, conviction 60, bullish): "Human expertise will be licensed like NYT/FT/Reddit data — ongoing passive royalties (Spotify-analogy) for contributions improving models or personal digital twins."
- real_vision (2022-12-12, conviction 70, bullish): "Mainstream platforms (Reddit 3M onboards, Starbucks, Instagram) drive the big adoption moments by abstracting away tokens/wallets for normal users." — dated and crypto-specific; included for completeness, weighted near zero.
The bearish side — content contamination, and it is the sharper argument:
- odd_lots (2026-04-02, conviction 70, bearish): "About 40% of internet pages are now AI-generated; over 50% of new Medium articles and ~10% of Reddit are AI-written."
- odd_lots (2026-04-02, conviction 65, neutral, entity-tagged Reddit): "Bot-farm startups sell 'organic' Reddit brand mentions because Reddit ranks in search and feeds LLM training — gaming both human and model outputs."
Excluded after review (name-collision discipline): two claims that surfaced on the word "Reddit" but are about retail investor flows, not the company — jordi_visser (2021-03-25) on small caps outperforming as "Reddit flows" lift smaller names, and gavin_baker (2021-11-30) on "social-media-fuelled collaboration among tens of thousands of retail investors on Reddit/Twitter/YouTube." Both are market-structure observations from 2021 with no bearing on RDDT the security. Counting them would have inflated the lane by 25% with claims that say nothing about the asset.
Read. The lane is mixed-positive, breadth 6, and unusually well-balanced. The bulls own the asset thesis (network effect, corpus scarcity, licensing); the bears own the mechanism (contamination and search dependency). Critically, the two most recent claims in the file are the two bearish ones (2026-04-02) — the panel's most current thinking on Reddit is the risk side, and that is exactly consistent with what the stock has done since. This is a case where the knowledge base explains the drawdown rather than contradicting it. It raises our respect for the bear case and is the primary reason this is Stage-In rather than Buy.
6. Technicals — oversold, bouncing, unrepaired
- Price $154.71, +9.98% today. A 10% single-session move five days after earnings implies heavy short positioning being covered.
- -42.9% from the 52-week high of $270.71 (the quote series shows a $282.95 intraday high); +27.0% from the 52-week low of $121.84. The full drawdown from peak is -42.9%.
- Below both moving averages: 50-DMA $175.39 (-11.8%), 200-DMA $179.87 (-14.0%). The 200-DMA is above the 50-DMA, which is the classic broken-trend configuration — the intermediate trend rolled over and the primary trend has now rolled over behind it.
- RSI 27.3, MACD -4.94. Deeply oversold with momentum still negative. Note that this reading includes today's +10% session, which tells you how far the stock had fallen before it.
- Beta 1.938 — the highest in this batch. Size the position accordingly.
- Relative performance: -3.7% over 12 months against SPY +19.9% and QQQ +23.9%. Reddit has underperformed the index by roughly 24-28 points over a year in which it grew revenue ~65%.
The technical structure argues for staging. The 200-DMA at ~$180 is 16% above spot and is the level that would confirm the drawdown as a re-load rather than a top. There is no completed base here — one 10% day off an RSI-27 low is a bounce, not a bottom.
7. Moat & competitive position
Reddit's moat is a community network effect with unusually high switching costs at the sub-community level. The asset is not the software; it is roughly two decades of accumulated, topic-organised, human-authored discussion plus the volunteer moderation layer that maintains it. A competitor can copy the interface trivially and cannot copy the corpus or the moderators at all. This is why we_study_billionaires' framing — "the internet's front page" — is more than a slogan: Reddit occupies a structural position in how people search for opinion rather than fact, which is precisely the query class that general-purpose answer engines handle worst.
The moat's shape also defines its vulnerability. Because the value sits in the corpus, the two threats are (a) the corpus being contaminated by generated content — the odd_lots claim — and (b) the corpus being consumed without the traffic, which is the search-referral question. Neither is a competitive-displacement risk. No named competitor threatens Reddit's position; the threats are to the channel and to the integrity of the asset itself. That is an unusual moat profile and it is why we rate Growth Quality 9 while rating Downside Risk 6.
Structural note: the profile describes Reddit as "a controlled subsidiary of Advance Publications, Inc." Minority shareholders should assume limited influence over strategic direction. The file does not disclose the class structure or voting split — a gap.
8. Management, capital allocation & the insider signal
- Capital allocation is currently: do nothing. FY25 shows $0 buybacks, $0 dividends, $0 acquisitions, capex of $6.7M, and $25.1M of stock issuance. The company is simply accumulating cash — $2.48B and rising. For a business generating a 31% free-cash-flow margin with no reinvestment needs, this is a live question. A buyback at $154 would be a strong signal; its absence is neutral-to-mildly-negative.
- Insider activity is one-directional and the file contains no purchases. The complete insider record here: COO Jennifer Wong, 2026-07-15, exercised 39,200 options at $5.35 and sold approximately 26,000 shares across six tranches at $190.54-$198.03. That is textbook option-exercise-and-sell, and the prices are 23-28% above today's $154.71. There is zero open-market insider buying anywhere in the file, including during the drawdown. Read it as neutral-to-mildly-negative: not a red flag on its own, but a missing bullish tell. Insider buying into this drawdown would be a meaningful upgrade to conviction.
- The CEO is Steven Ladd Huffman (Steve Huffman), a co-founder, still in the seat. Founder-led, 2,555 employees, IPO'd 2024-03-21 — this is a young public company with two years of operating history as a listed entity, which is itself a reason to demand a wider fair-value band.
9. Data integrity — what we rejected and why
Vendor data corruption is common in this feed and we check before we use. For RDDT:
- REJECTED: the entire consensus EBITDA and EBIT series. The
estblock reports negative EBITDA in every single forecast year — -$459.1M (FY26E), -$605.1M (FY27E), -$755.6M (FY28E), -$837.5M (FY29E), -$1,011M (FY30E) — while simultaneously reporting positive net income of $1.068B, $1.296B, $1.736B, $2.089B and $2.899B in the same rows. A company cannot have -$1.0B of EBITDA and +$2.9B of net income. Against reported FY25 EBITDA of +$544.6M, the FY25 estimate row shows -$294.8M. This series fails internal consistency in every year and is excluded from all analysis. Only the revenue, net income and EPS lines fromestare used. - REJECTED: the historical net-income estimate rows. FY23E net income of -$1.992B against actual -$90.8M, and FY24E of -$804.2M against actual -$484.3M. Historical estimate rows in this vendor block are unreliable; forward EPS rows cross-check adequately against realised results (FY25E EPS $2.315 versus $2.62 reported diluted) and are used with that caveat.
- CORRECTED: enterprise value. Vendor EV of $28.32B nets only the $953.6M cash line. Reddit also holds $1,523M in short-term investments. Our EV is $27.33B, and EV/Sales is 9.8x rather than the vendor's 10.2x.
- FLAGGED, NOT REJECTED:
interestCoverageRatioTTM= 0 anddebtServiceCoverageRatioTTM= 0. These are zero because interest expense is genuinely $0 across every reported period — Reddit has no meaningful debt. The zeros are correct but read as data errors on a screen; noted so no one treats them as such. - GAP: no operating metrics. No daily/weekly active users, no engagement, no ARPU, no traffic source, no advertiser count. For an advertising business this is the most important omission in the file, and it is the direct reason the score_risk is 6.
- GAP: seg_prod and seg_geo cover only FY24 and FY25 — two years, no quarterly disaggregation. We cannot see whether the licensing line inflected within FY26.
- GAP:
ptcarries a single consensus row with no per-analyst detail, andgradesgives counts without firm names or dates.
10. Verdict, kill-criteria & flip conditions
Stage-In. Reddit is a profitable, fast-accelerating, capital-light, net-cash advertising monopoly trading at 22.3x FY27 consensus earnings after a 43% drawdown, five sessions after beating the consensus EPS line by 32%. The economics are not in dispute: +61% revenue, 91.3% gross margin, 28.8% operating margin, $684M of free cash flow on $6.7M of capex, $2.45B of net cash. Base fair value ~$195 (+26%) requires no multiple expansion whatsoever — only that consensus earnings arrive and the price tracks them at today's forward multiple.
The reason it is Stage-In and not Buy: the drawdown is not irrational. It is pricing a structural question about AI-era discovery that our own knowledge base states in specific terms — and the two most recent panel claims on this asset are the bearish ones. Critically, the operating variable that would resolve the argument (audience) is absent from our data. Buying a full position on a P&L when the driver of that P&L is unobservable is not conviction, it is faith.
Staged entry (build a 1.5-3% position in 3 tranches):
1. Tranche 1 — now (~1/3 of target), at ~$155. RSI 27, -43% from the high, 22.3x forward, net cash, a 32% earnings beat five days old. Underwrite the demonstrated economics.
2. Tranche 2 — on either a reclaim of the 50-DMA (~$175) with the Q3 print in hand, OR a retest of the $122-135 zone (the 52-week low area) that holds. Buy confirmation or buy capitulation; do not buy the middle.
3. Tranche 3 — only after a data point that speaks to the referral question — a disclosed user metric, a new licensing agreement, or "Other Revenue" inflecting. This tranche is deliberately gated on new information rather than on price.
Pre-registered KILL / avoid-adding criteria:
- Any quarter below ~+30% YoY revenue growth before Q4'27 (consensus needs +48% in Q3'26). This is the cleanest single falsifier of the entire thesis.
- US revenue decelerating materially faster than non-US — the signature of a search-referral problem rather than a demand problem.
- Operating margin below 24% in a quarter without a disclosed one-time charge, indicating the fixed-cost leverage has reversed.
- Any disclosure of flat or declining users/engagement.
- Continued heavy insider selling with zero buying through a further leg down.
Pre-registered FLIP TO HIGH-CONVICTION (upsize toward 3%):
- "Other Revenue" inflecting above ~+50% growth, or a new named data-licensing agreement. This resolves the bear case structurally rather than tactically.
- A share repurchase authorisation deployed at these levels — management putting the $2.48B to work below the 200-DMA.
- Open-market insider buying into the drawdown.
- Two consecutive quarters above +45% revenue growth with the 200-DMA reclaimed.
Where RDDT fits in the Synthos Framework Portfolio. The consumer-internet / attention-monetization sleeve, as a high-beta growth position at 1.5-3%. Note the correlation profile: with beta 1.938 and a -43% drawdown already delivered, RDDT is a position that will dominate sleeve volatility. It is not a substitute for a diversified advertising exposure — it is a concentrated bet on one platform's audience durability. Logged as a tracked Synthos call (Stage-In) as of 2026-08-04 at $154.71.
Single biggest risk: audience/referral dependency that this data cannot measure. Every number in this dive — the +61% growth, the 28.8% margin, the $684M of free cash flow, the $195 base case — is downstream of people arriving at Reddit. The file contains no user metric, no engagement metric and no traffic-source disclosure. If AI answer engines absorb the discovery layer rather than routing through it, the revenue line breaks first and the margin line breaks immediately after, because advertising deleverages violently. The most fragile assumption in today's price is not the growth rate — it is that the channel delivering it is stable.
Provenance & disclosures
- Traceability: 8 usable KB claims across 6 tracked voices (breadth 6, net conviction mixed-positive). Bullish lane: we_study_billionaires (2025-12-04, conv 66, the only ticker-tagged claim), latent_space (2024-03-29, conv 70 and 50), macrovoices (2024-05-02, conv 70), ml_street_talk (2025-11-03, conv 60), real_vision (2022-12-12, conv 70, weighted near zero as dated and off-topic). Bearish lane: odd_lots (2026-04-02, conv 70 and 65) — the two most recent claims in the file. Two claims excluded as name collisions (jordi_visser 2021-03-25 and gavin_baker 2021-11-30, both about retail-investor flows rather than the company).
- Data as-of: fundamentals 2026-06-30 (Q2'26, filed 2026-07-31) · estimates 2026-08-04 · prices 2026-08-04 ($154.71, +9.98%; 50-DMA $175.39, 200-DMA $179.87, RSI 27.3, beta 1.938) · KB claims 2026-08-04. Sole data source:
scripts/deepdive/vti_data/RDDT_data.json. No figure in this dive comes from any other source. - Rejected data: the entire consensus EBITDA/EBIT series (negative in all forecast years alongside positive net income — internally impossible); historical net-income estimate rows (FY23E/FY24E off by 5-20x versus actuals). Vendor enterprise value corrected upward for $1,523M of unrecognised short-term investments.
- Disclosed gaps: no user, engagement, ARPU, traffic-source or advertiser-count metrics anywhere in the file — the single most important omission for an advertising business; segment data covers FY24-FY25 only with no quarterly split; no per-analyst price-target detail; class/voting structure of the Advance Publications control position not disclosed.
- Fair-value caveat: the $105 / $195 / $265 anchors are scenario multiples applied to consensus EPS, labelled at each step. The base case assumes no multiple re-rating and derives entirely from consensus earnings growth. It is scenario arithmetic, not a discounted-cash-flow model.
- Non-GAAP note: all P/E multiples in this dive use GAAP diluted EPS for trailing figures and the vendor consensus EPS series for forward figures. The FY25 consensus EPS of $2.315 versus $2.62 reported diluted suggests the consensus series is close to GAAP for this name, unlike several others in this batch.
- Not investment advice. Independent research, educational and informational only, never personalized. Nothing here is a recommendation to buy or sell any security.
- Version: 2026-08-04-full.