Executive Summary

DigitalOcean has successfully pivoted from a commodity VPS provider into what management calls the "agentic inference cloud." The thesis is structurally sound: AI agents need always-on compute, OpenClaw adoption is creating organic demand for cheap VPS instances, and real enterprise customers (Character.ai, Fal, NewsBreak) are signing 8-figure committed contracts. Revenue is accelerating — 16% YoY in Q3 2025 with management guiding 18-20% for FY2026, pulled forward a full year from the original 2027 target.

The problem: at $68.16, the stock is at the 95th percentile of its 52-week range, up 53% in the last year, and trading above the $66.51 capped call cap price where dilution protection from the 2030 convertible notes expires. Analyst upgrades have been chasing the move (Cantor to OW at $68, BofA at $72). The mean analyst target is actually below the current price at $59.58. The agentic cloud thesis is real. The question is whether there's alpha left.

Verdict: No edge at current price. Probability-weighted EV of $64.40 implies -5.5% downside from $68.16. The thesis is right but the stock is ahead of the thesis. Pass for new positions. For existing holders: Feb 24 earnings is the next catalyst — RPO inflection and FY2026 guidance quality determine whether the premium sustains.

Past Performance: The Turnaround Story

DigitalOcean's trajectory over the past three years is a genuine inflection:

The company hit a trough in FY2024 with growth decelerating to 13%, but Q3 2025 marked the highest organic incremental ARR in company history at $44M, and the highest revenue growth rate since Q3 2023. The business swung from operating losses in FY2022 to 43% adjusted EBITDA margins in Q3 2025. Operating cash flow grew from $195M (FY2022) to $283M (FY2024), with CapEx increasing from $106M to $178M as the company invested in GPU capacity and data centers.

Basic share count declined from 100M+ at IPO to 91.2M through $1.6B in cumulative buybacks, though the pace has slowed dramatically — from $500M in 2023 to just $100M authorized now. Important caveat: the diluted share count is 105.5M (Q3 2025 10-Q), reflecting convertible note dilution. With the stock at $68 — well above the $39.17 conversion price — the diluted count is what matters for valuation. Capital allocation priority has shifted to growth investment: 30MW of incremental data center capacity for 2026, new GPU orders, and equipment financing.

The earnings beat history is strong: four consecutive quarters of EPS beats, though the magnitude has been shrinking (43.7% in Q4 2024 down to 9.5% in Q3 2025), suggesting consensus is catching up.

The OpenClaw Catalyst: Real Demand, Real Revenue, Real Risk

OpenClaw is the most important near-term demand catalyst for DOCN, and it deserves granular analysis.

What's actually happening

OpenClaw — an open-source AI agent framework — went viral in late January 2026, accumulating 176,000+ GitHub stars at the fastest rate in GitHub history. The framework gives AI models (Claude, GPT, etc.) tools and autonomy to interact with computers and the internet. To run 24/7, these agents need always-on VPS instances — exactly what DigitalOcean sells.

On February 10, 2026, DigitalOcean launched a 1-Click OpenClaw deployment starting at $12/month with security hardening: container-based isolation, authentication tokens, and firewall rules pre-configured. On February 13, they announced they'd "streamlined" the deployment further. Board member Warren Adelman (ex-GoDaddy COO) was personally amplifying these posts on X.

The organic developer adoption is visible and specific:

• Engineers deploying trading bots, team assistants, and daily automation on $12-32/month Droplets
• An ex-Meta founder running an AI assistant "embedded in all team group chats, shipping real PRs" on a $32/mo Droplet
• An ex-Google/Rubrik engineer (a16z scout) deploying OpenClaw + Composio agents with a 15-minute setup
• Multiple developers citing DigitalOcean specifically as the preferred host over AWS and Hetzner

Cross-ticker validation

Cloudflare CEO Matthew Prince mentioned OpenClaw directly on NET's Q4 2025 earnings call (Feb 10, 2026): "All hear news around AI agents like OpenClaw... driving more need for classically Cloudflare's services." When the CEO of a $45B infrastructure company names your demand catalyst on their earnings call, the trend is real.

The security problem (and DOCN's response)

This is also a risk worth flagging: OpenClaw has major security issues. Within two weeks of going viral, researchers found 135,000 internet-facing OpenClaw instances, a CVE with 8.8 CVSS score enabling one-click remote code execution, and 12% of skills in the ClawHub registry were malicious (341 out of 2,857). Fortune, Cisco, Trend Micro, and VentureBeat have all published pieces on the security nightmare.

DigitalOcean's hardened 1-Click deployment directly addresses this gap. If OpenClaw adoption continues but security concerns intensify, DOCN's managed, security-hardened offering becomes more valuable — developers will pay $12-24/month for "it just works securely" rather than risk bare-metal exposure. This is a legitimate competitive moat within the OpenClaw ecosystem.

Revenue impact: material but not yet quantifiable

Management has never mentioned OpenClaw by name on an earnings call (the product launched after Q3 2025 results). The Q4 2025 call on Feb 24 will be the first opportunity to hear management frame the OpenClaw demand. Here's the math on potential impact:

• OpenClaw has 100,000+ active installations (per web reports)
• If even 10% run on DigitalOcean at $12-32/month average = $120K-$320K/month incremental MRR
• That's $1.4M-$3.8M annualized — immaterial to $1B ARR

The direct OpenClaw revenue won't move the needle. What matters is whether OpenClaw is a gateway drug: developers deploying agents discover they need databases, storage, Kubernetes, monitoring — the full "agentic cloud stack" that DOCN has built. Management has repeatedly described this land-and-expand motion: customers "start journey fairly small, now starting to scale." The 8-figure contracts from Character.ai and Fal demonstrate the endpoint of this funnel.

The Enterprise AI Thesis: Bigger Than OpenClaw

OpenClaw gets the headlines, but the real revenue driver is DOCN's enterprise AI infrastructure:

Character.ai (announced Jan 13, 2026): DOCN's inference cloud with AMD Instinct MI300X/MI325X GPUs delivers 2X production inference throughput and 50% lower token costs for Character.ai's 1B+ daily queries. This is the kind of customer Cantor Fitzgerald specifically cited in their upgrade. Character.ai is processing over a billion daily queries — this is production-scale inference, not experimentation.

Fal (strategic partnership): Generative media platform serving Canva, Shopify, and Perplexity. Hosts and runs hundreds of models on DOCN infrastructure. The partnership expanded to bring fal's image/audio generation to DOCN's 640,000+ customer base via serverless inference.

The 8-figure contracts: Management disclosed "multiple 8-figure committed contracts" in Q3, expected to "materially increase RPO in Q4." This is transformative for a company whose RPO was only $47.4M — just 5% of ARR. If RPO doubles or triples in Q4, it shifts the revenue visibility narrative entirely.

AI revenue trajectory: Doubling YoY for 5 consecutive quarters through Q3 2025. CFO Matt Steinfort guided AI revenue into "mid-teens, maybe high teens as percent of revenue" — implying $30-40M/quarter or $120-160M annualized. This is NOT included in NDR calculations yet, meaning reported NDR of 99% understates actual customer expansion.

Gradient AI Platform metrics: 19,000+ agents deployed, 7,000+ in production. GPU Droplets available with NVIDIA H200, AMD MI300X, RTX 400, and L40s. Inference-specific optimizations: page retention, flash attention, FP8 quantization, speculative decoding.

Three-Quarter Outlook: Q4 2025 Through Q2 2026

Q4 2025 (reports Feb 24, 2026)

Guided: Revenue $237-238M (≈16% YoY), EBITDA margin 38.5-39.5%, non-GAAP EPS $0.35-0.40.

Consensus: EPS $0.38. The lower EBITDA margin vs Q3 (43%) reflects data center ramp costs — NRC (non-recurring charges) for new facilities plus lease-up expenses before revenue fills capacity.

What to watch: RPO inflection (management said "material" increase from committed contracts), AI revenue disclosure (will they finally give a number?), FY2026 formal guidance (18-20% previewed), any mention of OpenClaw demand.

Risk: Q4 EPS guidance implies a sequential step-down from $0.54 to $0.35-0.40 — largely explained by new debt interest ($0.05-0.10 reduction) and DC ramp costs. If the miss is worse than guided, the stock at 95th percentile of its range has significant downside.

Q1 2026 (reports ~May 2026)

Expected: Revenue ≈$245-255M (18-20% growth implies $917M-$935M FY2026, so ≈$230M+/quarter average ramping through the year). First full quarter with 30MW incremental DC capacity coming online. OpenClaw demand should be visible in Droplet metrics. EBITDA margin temporarily depressed in "high 30s to 40%" per CFO as new capacity ramps.

Catalyst: First quarter where OpenClaw-driven demand would appear in financials. If management quantifies "agentic workload" growth separately, the narrative could strengthen significantly.

Q2 2026 (reports ~Aug 2026)

Expected: Revenue growth should accelerate toward the upper end of 18-20% as new DC capacity fills. GPU capacity additions ramp inference revenue. AI revenue could reach 20%+ of total if doubling trend continues (from ≈15-18% in Q3 2025). EBITDA margins should recover toward 40%+ as lease-up expenses normalize.

Key question: Does demand exceed the 30MW of new capacity? If demand > supply persists, management could signal further investment acceleration, which would be bullish for long-term growth but could pressure near-term margins.

Valuation, Alpha Calculation, and Risk Assessment

Current valuation

At 7.5x trailing revenue, DOCN trades at a premium to its historical average but below pure-play AI infrastructure names. The forward P/E of 34.8x assumes consensus EPS that may be conservative if AI revenue inflects.

Scenario analysis and alpha calculation

Why these probabilities: Bull at 25% because it requires guidance above 18% AND RPO inflection AND AI revenue disclosure — multiple conditions. Base at 50% because management has been guiding conservatively and delivering; 18% growth-as-guided is the most likely outcome. Bear at 25% because the CPTO vacancy, 17.4% short interest, hyperscaler competitive response, and convertible dilution are real risks that the market may be underweighting at the 95th percentile of range.

Alpha calculation:

Current: $68.16
EV target: $65.13
Total expected return (12mo): -4.4%
Tech sector expected return (XLK 1Y): assume 12%
Idio return: -4.4% - 12% = -16.4%
Idio variance: 65% (below 75% target)
Edge%: 65% (company-level thesis only, no sector edge)

α = -16.4% × 0.65 = -10.7% annualized

This is a negative alpha stock at current prices. The thesis is correct but the price already reflects it. No position justified.

Bull case ($80-85, 12 months)

• FY2026 revenue growth hits 20%+ (above guidance)
• AI revenue disclosed at 20%+ of total, doubling trend continues
• RPO inflects to $150M+ showing visibility transformation
• OpenClaw creates sustainable developer funnel
• Multiple re-rates to 9-10x forward revenue
• Upside: +17-25%

Base case ($62-70, 12 months)

• FY2026 delivers 18% growth as guided
• AI revenue grows but competitive pressure from hyperscalers intensifies
• Margins compress temporarily during DC ramp
• Stock consolidates around current levels
• This is roughly where the stock is now

Bear case ($42-50, 12 months)

• OpenClaw demand proves fleeting (security concerns kill adoption)
• Hyperscalers (AWS, Azure, GCP) release competitive $10-15/mo agent hosting
• CPTO departure impairs AI product execution (no replacement announced)
• Debt burden constrains flexibility (net debt/EBITDA mid-3x)
• Convertible dilution accelerates (see below)
• Short interest (17.4%) could amplify any negative catalyst
• Downside: -26-38%

Key risks

1. Competitive sandwich: Hyperscalers above (more GPUs, better pricing at scale), cheap VPS providers below (Hetzner, Vultr offering similar compute cheaper). DOCN's moat is the integrated "agentic cloud stack" — if agents remain simple enough to run on bare Droplets, the software differentiation doesn't matter.

2. CPTO departure and insider selling: Bratin Saha resigned Nov 20, 2025. He was the AI strategy architect referenced repeatedly on earnings calls. No replacement announced. Notably, Saha sold 13,010 shares on September 3, 2025 — two months before his departure. No other insiders have made open-market purchases. All recent insider transactions are director stock awards (compensation grants), not conviction buys. For a company whose growth thesis depends on AI product execution, the AI chief selling and leaving is a yellow flag.

3. Convertible dilution mechanics: The 2030 convertible notes ($625M, 0% coupon) have a conversion price of $39.17 per share. At $68, the stock is well above conversion price — dilution has been active since the stock crossed $39.17. The capped call transactions offset dilution between $39.17 and $66.51. Above $66.51, the capped call protection expires and the company bears full dilution. With the stock at $68.16, DOCN is in unprotected dilution territory. The fully diluted share count of 105.5M (vs 91.2M basic) reflects this — a 15.7% difference that materially affects per-share valuation. The remaining ≈$130M of 2026 converts also need refinancing.

4. Factor exposure: Idiosyncratic variance is only 65% — below the 75% target. Beta of 1.77 with significant XLK (tech sector) loading means a tech selloff hits DOCN regardless of fundamentals. The stock amplifies market moves nearly 2:1. This is a momentum-loaded, high-beta name.

5. OpenClaw security backlash: If a major security incident occurs involving OpenClaw on DigitalOcean infrastructure specifically, the reputational damage could be significant despite the hardened deployment.

6. Debt structure: $1.32B total debt, net debt/EBITDA mid-3x by end 2026. For a company with $897M revenue, this is meaningful leverage.

The Bottom Line

DigitalOcean has a genuine structural thesis: the agentic AI wave needs cheap, reliable, developer-friendly compute, and DOCN is the best-positioned VPS provider to capture it. The evidence is accumulating — Character.ai, Fal, OpenClaw organic adoption, 8-figure contracts, AI revenue doubling for 5 consecutive quarters.

But the stock is ahead of the thesis. At $68, it's above every analyst's median target, at 95% of its 52-week range, in unprotected convertible dilution territory above $66.51, and trading 34.8x forward earnings for a company growing 18-20%. The probability-weighted EV of $65.13 implies negative alpha at current prices. The short interest at 17.4% could amplify moves in either direction. The CPTO who architected the AI strategy sold shares and left. No insider is buying.

For new positions: pass. Negative alpha at current price. The risk/reward at 95th percentile of range with earnings 9 days away doesn't justify entry. If earnings disappoint, a pullback to $50-55 (the 200-day MA is $38.86, 50-day is $53.14) would create a much better entry with 30-40% upside to fair value.

For existing holders: Hold through earnings. Watch for RPO inflection above $100M and AI revenue disclosure as the key signals that could push the bull case probability higher. If FY2026 guidance comes in above 20%, re-evaluate — the probability-weighted EV shifts materially. If guidance merely confirms 18%, consider trimming into strength.

Revisit trigger: Stock below $55, OR FY2026 guidance above 20%, OR RPO above $150M. Any of these changes the alpha calculation.