Built by Claude Fable

DOCN | Investor Sentiment: 5/10

Inverted scoring: the question is not whether DOCN is liked or disliked, but whether management is repeatedly bullish on a specific driver the street does not believe or does not model. Data note: Daloopa is not connected this session (see company_context.md), so no daloopa.com/src citations are possible. Figures below come from the FY2024Q3 to FY2026Q2 call transcripts in transcripts/, FMP consensus files in this run folder (fmp_analyst_estimates.json, fmp_grades_consensus.json, fmp_price_target_consensus.json, fmp_quote.json), SEC Form 4 filings, and web sources listed at the end.

4a. Management-Street Divergence (50% weight)

How management's outlook has moved (transcripts)

Sources: DOCN_FY2025Q4.txt (prepared remarks, Q&A line 18), DOCN_FY2026Q1.txt (Crane Q&A), DOCN_FY2026Q2.txt (Steinfort guidance; Mohan, J. Ader, Singh Q&A).

What the street models vs. what management says (FMP consensus)

Where the divergence actually is

1. Headline growth: no divergence. Management walked FY2027 from 30% (Feb) to 50%+ (May) and the sell side moved with it; consensus FY2027 revenue of $1,820M (+54%) already sits above management's "50% plus." Q&A on the Q2 call was the street asking management to confirm upside (BofA: "should we not assume directionally that there is further upside to 2027"), not doubting it. That is the opposite of NVDA-style disbelief.
2. The specific contrarian thesis: "software makes megawatts more valuable." Management repeats every call that DOCN is not a bare-metal neocloud: 85% of AI customer ARR is non-bare metal, inference services grew ~800% YoY and are 70%+ of AI ARR, ~70% of $100K+ AI customers attach core cloud, and therefore incremental ARR per MW should rise (FY2026Q2 prepared remarks and Singh Q&A). The street keeps anchoring to megawatt math: Canaccord's "low to mid-teens revenue per megawatt," Morgan Stanley's "you previously said revenue per megawatt will come down," and UBS cutting its target twice (to $155 in July, to $140 in August) on capacity, not on demand. The street values DOCN as a capacity-bound GPU landlord; management argues it is a software platform whose revenue per MW and stickiness compound. This is a genuine, repeated, specific divergence.
3. Durability vs. neoclouds and lease-adjusted cash flow. William Blair asked in back-to-back quarters for FCF including leases and whether neoclouds' shift to full-stack inference erodes the moat. Management answered qualitatively ("can't provide specificity on lease payments"). Skepticism here is reasonable, not a mispricing.
4. Market vs. analysts. The tension is less between management and the sell side than between both of them and the tape: shares fell after the Q2 beat-and-raise (Investing.com) and fell 14.8% on Oct 14 despite Oppenheimer raising to $220 (GuruFocus, Finimize). Investors are discounting credibility and duration of the ramp, but that is not management pushing a thesis the street ignores.

Verdict on divergence: partial. The ARR-per-MW / software-attach thesis is a real management-only view, but the bigger-ticket bullish claim (50%+ growth) is fully in consensus already. Sub-score ~5.5/10.

4b. Analyst consensus

Sell side is constructive and tone on calls is effusive ("amazing results," "congrats on the momentum, you've earned it"). Targets far above the current price mean the sell side has not lost faith; the gap reflects the market's doubt, not analyst skepticism. Rubric-wise this is the "Buy-heavy, target well above current" profile, partly offset by a third of coverage at Hold. Sub-score ~5/10.

4c. Insider activity

Sources: SEC Form 4 filings (EDGAR CIK 0001582961), InsiderTrades, Motley Fool. Steady pre-planned selling, small relative to holdings, no open-market buying through the drawdown. No confirming signal; mildly negative. Sub-score ~4/10.

4d. Retail / social sentiment

• Momentum attention is high: the stock was up ~336% over the 12 months to Aug 17 (Motley Fool) and ~127% YTD as of mid-October (Citizens via Investing.com); small-cap momentum outlets (StocksToTrade / Timothy Sykes) ran repeated "DOCN surges/jumps" pieces around the Oct 9 to 10 AI cloud announcements.
• Short interest is elevated: ~7% to 12% of float through summer 2026 (Equibles, Benzinga, FINRA-based tracker), so it is a contested, high-beta AI name, not a quiet one.
• Volatility both ways (52-week range $37.84 to $187.50) points to a crowded, trading-driven holder base. Not low attention. Sub-score ~4/10.

Score build

Web sources: Investing.com, Citizens after summit; StockAnalysis forecast; MarketBeat Oct 15; GuruFocus 14.8% drop; Finimize; Investing.com Q2 transcript, shares fall premarket; MarketBeat UBS cut; SEC Form 4; InsiderTrades CFO sale; Equibles short interest; StocksToTrade.

Assessment. DOCN does not have an NVDA-style contrarian setup on its headline story: management's big bullish call, 50%+ revenue growth in 2027 (up from 30% in February), is already in consensus at +54%, and the sell side is 68% Buy with an average target of $175 against a $98.72 price, so analysts are believers, not doubters. The one real management-only thesis is that DOCN's software (inference engine, core cloud attach, 85% non-bare-metal AI ARR) makes each megawatt earn more over time, while the street still values it on megawatt math and capacity (Canaccord's low-to-mid-teens ARR/MW, UBS cutting twice on capacity). That gap is specific and repeated, and the beat-and-raise record since February gives it some credibility. It is offset by steady 10b5-1 selling from the CEO and CFO with no buying through the drawdown, elevated short interest, and a momentum-driven retail base after a ~336% one-year run. The selloff reflects the market doubting how long the ramp lasts, not management being ignored, which lands at 5/10: partial divergence, mostly priced into estimates, with no insider or low-attention confirmation.