DigitalOcean (DOCN) Deep Dive: Why We Own It

Why We Own DigitalOcean (DOCN)

DOCN is the rare cloud-infrastructure name that combines three things that almost never travel together: triple-digit price momentum (+404% over the trailing 252 sessions), real GAAP profitability (16.4% Operating Margin TTM, while the median direct cloud peer is at -7.2%), and a positive earnings-revision trail (14 upward FY 2026 revisions, 0 downward over the past 30 days). It's the kind of profile our three-pillar process is built to spot: a hot field inside a mature sector, where one player is doing something structurally different.

The "developer cloud" angle has been DOCN's pitch for a decade — simple pricing, a clean dashboard, and infrastructure aimed at startups rather than Fortune 500 IT departments. What changed in the last 18 months is the layer the company built on top: an agentic inference cloud that lets a two-person team deploy and scale AI workloads without the complexity of hyperscaler architectures. The data says the market believes it.

At a Glance

Numbers below are computed from primary fundamentals data and 10-Q / 10-K filings as of the most recent reporting period. Both peer-set medians are calculated by us from the same underlying data using the methodology shown.

*Net margin includes a non-operating gain booked in Q3 2025 (likely a deferred-tax valuation-allowance release of roughly $113M). Operating margin is the cleaner read on the underlying business — that's the line we anchor to.

Direct peer set (n=9): the public cloud-infrastructure and managed-data names closest to DOCN's operating model — NET (Cloudflare), AKAM (Akamai, owns Linode), FSLY (Fastly), MDB (MongoDB), CFLT (Confluent), ESTC (Elastic), DDOG (Datadog), GTLB (GitLab), SNOW (Snowflake). All US-listed, all with mature reporting. The hyperscalers themselves (AMZN, MSFT, GOOGL) are excluded because cloud is a small fraction of their reported financials and would dilute the comparison. Broad sector set (n=100): the top 100 US-listed names by market cap in the GICS Information Technology sector, filtered for $300M+ market cap and at least 8 quarters of reporting history.

A Profitable Cloud in a Loss-Making Niche

The peer chart below is the single picture that explains why DOCN is in the portfolio. Direct cloud peers grow faster on average (22.8% vs. 17.6%), but they do it while burning cash — the median operating margin in the direct set is negative 7.2%. DOCN runs at +16.4% while still posting double-digit revenue growth. The broad IT sector median is similar on margin (15.6%) but only grows at 13.4%. DOCN sits in a rare quadrant: above-sector growth, sector-level profitability, in a niche where neither is the norm.

Two honest caveats. First, DOCN's TTM revenue growth (17.6%) is below the direct peer median, but it's reaccelerating: the most recent reported quarter (Q1 2026) grew +22.4% year-over-year, the fastest pace in two years, driven by the AI-infrastructure ramp. Second, the +404% one-year price move is partly a base effect — the stock came off a $25 floor twelve months ago when sentiment was deeply negative. Some of the easy rerating is now behind us; the next leg has to be earned with execution.

The forward-revision trail says the sell side increasingly believes that execution. Over the past 30 days we count 14 upward revisions to FY 2026 EPS estimates and zero downward — a strikingly clean signal. The forward quarter (Q3 2026) shows 10 up and 3 down over the same window. That's the opposite of what aggregator screeners were flagging a few months ago and reflects the ARR disclosures from the company's last two earnings calls.

The Business — Developer Cloud at Scale

DigitalOcean was founded in 2012 in Broomfield, Colorado, and went public on the NYSE in March 2021. The company runs IaaS and PaaS for one specific customer profile: technology companies that are too big for a single VPS but too small (or too cost-sensitive) to want a dedicated AWS solutions architect. Common workloads include online gaming, fintech APIs, cybersecurity infrastructure, and — increasingly — AI inference.

The competitive picture is asymmetric. On one side: AWS, Azure, and GCP — vastly larger, with effectively unlimited capital and product breadth. On the other side: the tiny self-hosted alternatives (Vultr, Hetzner) and a handful of acquired-and-folded competitors (Linode now sits inside Akamai). DOCN's defensible middle is the workflow itself: pricing you can predict on a napkin, a dashboard a junior developer can navigate, and integrated services that compress the operational complexity of running cloud infrastructure.

In 2024–2025 the company layered an agentic inference cloud on top of that base — managed GPU instances, model hosting, and integrated tooling for retrieval-augmented and agent workloads. That's where most of the recent ARR growth has come from, and it's why the latest-quarter revenue reaccelerated even as the broader cloud market matured.

“The hyperscalers are massive but architecturally complex. DOCN's pitch is the opposite: predictable pricing, a clean dashboard, AI infrastructure that a two-person startup can actually use. That asymmetry — designed for the customers the giants don't profitably serve — is the moat.”

Institutional-Grade Profitability

TTM revenue is $948M. TTM EBITDA is $298M — an EBITDA margin of roughly 31%, well into territory that institutional investors recognise as durable. Operating margin is 16.4% on a GAAP basis (no adjustments, no exclusions). Stock-based compensation runs higher than we'd like — common in cloud — but the company is still printing positive free cash flow while investing aggressively in data-centre capacity.

The balance sheet is solid but not pristine. Cash and short-term investments stand at $741M; total debt is $1.09B (a mix of convertibles and a credit facility). Net debt of roughly $158M is comfortable against TTM EBITDA — leverage just over 0.5x. Tangible book is supportive but, importantly, this is not a cheap balance-sheet story; this is a profitability and growth story.

The Earnings Trajectory

DOCN has now posted 10 consecutive earnings beats, with the most recent quarter (Q1 2026) coming in at $0.44 versus a $0.26 consensus — a 69% surprise. The chart shows the trajectory: a 2024 ramp into a Q2 2025 peak ($0.59), a Q3-Q4 dip as AI-infrastructure investment hit the income statement, and a stable $0.44 print in Q1 2026 despite the heavier capex.

The honest reading: near-term EPS is being held back by infrastructure spend. Q1 2026 EPS of $0.44 is down 21% versus the year-prior $0.56 — and that's the bear case. But the analyst community is reading through it. Forward FY 2026 consensus EPS has been revised up 14 times and down zero in the past 30 days, with average upward revisions of $1.20 per share over the past quarter. The narrative shift is from "AI capex weighing on earnings" to "AI capex producing the next leg of growth."

Valuation — Paying for Quality and Growth

This is where we are the most honest with ourselves. DOCN's forward P/E of 158.7 is well above both peer-set medians (direct: 62.1, broad sector: 22.4). Price-to-book of 19.3 is more than double the direct peer median (8.2) and almost 2.5x the broad sector (7.7). On every static valuation metric, DOCN looks expensive.

Two things justify holding the position at this multiple. First, the EBITDA multiple is far more sensible — at $298M TTM EBITDA against a $16.3B market cap, the EV/EBITDA is in the high-50s, which is rich but in line with profitable hyper-growth cloud names that have inflected on AI. Second, the forward consensus EPS estimate has been moving up faster than the price for several months; if the FY 2027 EPS estimate of $1.72 lands accurately, the implied forward P/E compresses to roughly 92, and FY 2028 lower still. The valuation is a bet that the upgrade cycle continues, not that the multiple stays where it is.

Wall Street's average price target of $177 implies modest single-digit upside from current levels — confirming that consensus, including the sell side, considers most of the rerating to have already happened. We don't disagree. We hold DOCN for the next earnings leg, not the next multiple expansion.

How DOCN Passed Our Three-Pillar Process

DOCN didn't make the portfolio because we liked the AI narrative — every cloud name has an AI narrative right now. It made the portfolio because it cleared all three pillars of our systematic process, in order, on the numbers.

Pillar 1 — Stock Universe Filter

The Universe stage looks for companies with durable profitability, accelerating fundamentals, and clean balance sheets. DOCN cleared on operating margin (+16.4% vs. -7.2% direct peer median), Return on Equity TTM (high, helped by tight equity base), and the 30-day revision spread (+14 / 0 for FY 2026). It cleared decisively on momentum metrics. The only filter it sits closer to the line on is valuation — but the Universe stage uses growth-adjusted valuation, where DOCN's 158x forward P/E divided by the consensus FY 2026 EPS growth rate is supportable.

Pillar 2 — Optimizer

The Optimizer selects which 15–30 names from the qualified universe make the live portfolio. It does not size positions — all portfolio positions are equal-weighted. DOCN was selected for its diversification properties: cloud-infrastructure exposure that doesn't correlate one-for-one with the semis or with the large-cap software names we already hold, plus a beta of 1.42 that's high enough to participate in upside but not so high that it dominates the portfolio's variance. More on how this stage works in Portfolio Construction: Why 15–30 Stocks Is the Sweet Spot.

Pillar 3 — Risk Overlay

The overlay decides how much of the portfolio is actually invested at any time — irrespective of how good the individual names are. As of late April we shifted from Braking to Cruising mode, meaning we run roughly partial equity exposure with the rest in cash. DOCN's beta of 1.42 means our effective DOCN exposure scales with the overlay: when the overlay calls for less risk on, our DOCN position contributes proportionally less to portfolio variance. Detail in Monthly Review: April 2026.

Catalysts on the Radar

• Next earnings: Q2 2026 in early August. Consensus EPS of $0.25 looks low given the cadence; another beat would extend the 10-quarter streak.
• AI-infrastructure ARR disclosure on the next earnings call — the company has been incrementally adding detail on the inference-cloud customer mix.
• Continued AI-capex normalisation through 2H 2026; operating margin can re-expand once peak capex rolls off.
• Read-across from hyperscaler AI capex commentary (AWS, Azure, GCP) — sustained demand confirms the bull thesis.
• FY 2027 EPS consensus currently $1.72. If revisions continue at the current pace, this is the number we watch for the next leg of the valuation re-rating.

Risks worth naming: a hyperscaler price war on developer-tier infrastructure, an AI-spending pullback, or stock-based-compensation dilution faster than EPS growth. All three are real; all three are why this is one name out of 15–30, not a concentrated position. AI exposure across the wider portfolio is discussed in How Our Portfolio Captures the AI Boom — While Staying Balanced.

“A profitable cloud business growing at the rate of an unprofitable one is unusual. The valuation reflects that. The position size reflects discipline. The overlay scales it back when markets demand caution. That is what owning DOCN looks like inside a systematic process.”

Related reading. Previous deep dives: Seagate (STX), Sandisk (SNDK), Nextpower (NXT). Process context: Portfolio Construction: Why 15–30 Stocks Is the Sweet Spot.

*DigitalOcean Holdings (DOCN) is a current Veloris Capital portfolio holding as of the publication date. Past performance is not an indication of future results. Your capital is at risk.*

Weekly investment newsletter, every Sunday — positioning, performance vs SPY/QQQ, and what we're watching next. No spam.