Digitalocean Holdings, Inc.NYSE: DOCN

Earnings Presentation Q2 2026

Aug 4, 2026 at 11:10 AM UTC · MarketScreener

Watch Digitalocean explain itDigitalOcean Holdings, Inc. (DOCN) Q2 2026 Earnings Call · Replay · Aug 4

Q2 2026 Eaínings Písssntation

August 4, 2026

6ccslsíating Gíowtk, Disciplinsa Exscution

6ccslsíating Gíowtk Mats

Revenue grew 29%, more than double a year ago, with record incremental ARR and strong profitability

Infsísncs Tíaction

Inference services grew nearly 800% year over year as AI Natives scale production workloads on our platform

6I Nativs Cloua Llywkssl

Our five-layer platform is becoming a flywheel: 80+ releases since April, and every layer adopted pulls customers into the next

Disciplinsa Exscution

Delivered 40% adjusted EBITDA margins while deploying new capacity on or ahead of schedule and reducing leverage

Mscoía Q2 Msvsnus Mssults & Stíong Píofitability

Ǫ2 Fsvsnus

$281M

29% YoY

Incís"sntal 6FF

$93M 191% YoY

40%

Ǫ2 6ajustsa EÐITD6 Maígin

24%

Ǫ2 6ajustsa Opsíating Inco"s Maígin

17%

TTM 6ajustsa FCF Maígin

Note: Refer to the Appendix for definitions of metrics and reconciliations of Non-GAAP measures to GAAP.

$1M+ Customsí 6MM Msacksa $259M, Gíowing 214% YoY

$100K+ Customer ARR($M)

$500K+ Customer ARR($M)

$1M+ Customer ARR($M)

% of total ARR

23%

35%

13%

26%

23%

Note: Refer to the Appendix for definitions of customer categories, customer count and ARR metrics and our Form 10-Q filed with the SEC for a description of our prior methodology. Prior periods have been recast using our current methodology.

Certain growth rates may not recalculate precisely to numbers shown due to rounding.

Q2 2026

Q2 2025

+214%

Q2 2026

Q2 2025

+160%

Q2 2026

Q2 2025

+98%

% Non Ðaís-mstal

43%

61%

70%

♦1%

♦5%

6I Customsí 6MM Msacksa $234M, Gíowing 212% YoY

AI Customer ARR($M)

Q2 YoY

ARR Growth

Total

Coís Cloua

Infsísncs Ssívicss

Ðaís Mstal

Note: Refer to the Appendix for definitions of AI Customers and AI Customer ARR.

Core Cloud includes: VM, VPC, Storage, High Performance NFS. CPU Kubernetes & Droplets. Inferences Services includes: all non bare metal inference capabilities.

Managsa 6gsnts

Open harness · Agent sandboxes · Tools · State · Planning

Web search · MCP calling · Function calling · Knowledge retrieval

Data & Lsaíning

DBaaS · Advanced Edition (PG/MySQL) ·

Knowledge Bases ·

Weaviate

Serverless

Dedicated BYOM

Infsísncs Moutsí

Píompt cacking

Continuous batcking

Moasl svaluations

Moasl syntkssis

Inrsísncs Engins

Moasl Catalog

Guaíaíails

Woíkloaa-awaís scksauling

Ssívsí-sias tools

Obssívability & insigkts

6utoscaling & loaa balancing

Quantization (4-bit / ♦-bit)

Spsculativs ascoaing

KV-cacks optimization

Ksínsl optimization

Cloua Píi"itivss

Kubernetes ·

Droplets (CPU & GPU) ·

VPC · Storage (object, block & file) ·

High performance NFS

Inríastíuctuís

20 data centers, liquid-cooled · NVIDIA: H100, H200, HGX B300 · AMD Instinct™: MI300X, MI350X, MI355X

Ksy upaatss sincs 6píil

DigitalOcsan 6I-Nativs Cloua

Inrsísncs Engins kas bsco"s a rull píoauction íunti"s solving toaay's "ost písssing nssas

Mapia Innovation 6cíoss 6ll 6I-Nativs Cloua Laysís

S0¦ píoauct íslsasss sincs 6píil '26

Managsa 6gsnts

OpenCode integration
Grok Build marketplace integration
- Codex in the Cloud / DigitalOcean for Codex
- Server-Side Tools for agentic applications
  Data & Lsaíning
Knowledge Bases
Managed MySQL Advanced
- Managed Weaviate
- DBaaS available across regions
Managed PostgreSQL Advanced
Vector databases portfolio

Inrsísncs Engins

AI-Native Cloud & Inference Engine launch
Dedicated Inference GA
Inference Router (public preview)
Unified Batch Inference
Prompt Caching (public preview)
Model Catalog & BYOM GA
Server-Side Tools (public preview)
Model Synthesis (public preview)
Model Evaluations
Multimodal Serverless Inference
Guardrails
Kernel Optimization
Model additions (29+): DeepSeek V3.2, DeepSeek V4 Pro, DeepSeek V4 Flash, Kimi K2.6, Kimi K3, GLM-5.1, GLM-5.2, GLM-B.2, Gemma 4, Qwen3 Coder Flash, Qwen3 Embedding 0.6B, Qwen 3 TTS (1.7B), Qwen 3.5 397B A17B, Llama 4 Maverick 17B 128E, NVIDIA Nemotron 3 Nano Omni, NVIDIA Nemotron 3 Ultra, Nemotron Nano 12B v2 VL, Claude Opus 4.7, Claude Sonnet 5, Claude Fable 5, Claude Opus (5), GPT Image 2, GPT-5.5, GPT
5.6 Sol, Terra, Luna, Stable Diffusion 3.5 Large, MiMo V2.5, MiMo V2.5 Pro.
Cloua Píi"itivss
- Security Posture Management · Private Droplets GA · GPU-ready Kubernetes hardening with automatic node remediation
- New York backbone upgrade to 1.6 Tbps · Droplets (CPU & GPU) · VPC · Storage (object, block, and file)
- High-performance NFS
  Inríastíuctuís
  - 20 data centers · NVIDIA: H100, H200, HGX B300 · AMD Instinct: MI300X, MI350X, MI355X
  - Liquid-cooled AI infrastructure · High-speed networking foundation

Selected externally announced launches and availability milestones since Deploy (late April 2026) - not the full platform capability list.

Tks 6I-Nativs Cloua Llywkssl

Evsíy laysí a custo"sí aaopts pulls tks" into tks nsxt

Ðsttsí

píics-psíroí"ancs

Inrsísncs Engins

Moís consu"ption

ADOPTION UNIT ECONOMICS

Five Layers

One Integrated Platform Owned Infrastructure

Cloua Píi"itivss

Managsa 6gsnts

Compounding on infrastructure we own

Lowsí cost to ssívs

Data & Lsaíning

Higksí utilization

Opsה-sonícs 6I codiהg agsהt ussa by ovsí 7.5M asvslopsís "ontkly to buila, tsst, ana skip sortwaís autono"ously

Ðuilt on DigitalOcsan roí:

Tíillions or opsn-wsigkt "oasl toesns aslivsísa by an sntsípíiss-gíaas, píivats inrsísncs sngins

Sscnís 6I saהdbox platfoíc roí íunning 6I-gsnsíatsa coas ana autono"ous agsnt woíerlows

Ðuilt on DigitalOcsan roí:

GPU Sanaboxss roí

CUD6-accslsíatsa woíeloaas, witk custo" snviíon"snts asploysa as isolatsa co"puts in "illissconas

6gsהtic iהfíastínctnís platfoíc wksís ku"ans ana agsnts skip sortwaís togstksí

Ðuilt on DigitalOcsan roí:

Intsgíating DigitalOcsan's inrsísncs sngins into tksií 6I Gatsway to píovias tksií pís"isí custo"sís asaicatsa 6I íssouícss

Lsadiהg LLM íontiהg caífistplacs connscting "illions or asvslopsís to "oasls acíoss tks 6I scosysts"

Ckannsl roí DigitalOcsan:

Millions or asvslopsís ana agsnts acíoss tks scosysts" ísacking ouí Inrsísncs Engins on as"ana

6I Ðuilasís 6ís Spinning Tks Gíowtk Llywkssl

DigitalOcsan Compstitivs 6avantags

Ssíving 6I-Nativss witk ons intsgíatsa platroí", puíposs built roí agsntic applications

Strength Partial Gap

Managsa 6gsnts

Data ana Lsaíning

Inrsísncing

Cloua Píi"itivss

Inríastíuctuís

Nativs agsnt íunti"s

Sandboxes, evals, observability Built for agents

Intsgíatsa aata laysí

Knowledge Bases, DBs, RAG-ready Data becomes learning

Full-stace inrsísncs ssívicss

Router, Synthesis, Evals, Batch, Caching - 75+ models, one API

~800% YoY growth

Co"plsts coís cloua

CPU/GPU, managed K8s, storage, network

70% of $100K+ AI customers attach

20 global aata csntsís, ownsa

~155MW committed capacity Owning the stack lowers cost to serve

Hypsíscalsís

Large enterprises & frontier labs

DIY acíoss ssívicss

Optimized for large enterprises

20¦ DÐ options

Powerful, but assembly required

Ðsaíoce / Vsítsx / Founaíy

Strong, within their own ecosystem

Ðíoaasst poítrolio

Built for enterprise scale and complexity

Hypsíscals rootpíint

Built for the largest workloads

Nsoclouas

Optimized for training clusters

No agsnt laysí

Separate vendor needed

No aata laysí

Separate vendor needed

Sslr-"anagsa on baís "stal

Some software, mostly via acquisitions

Stíong GPU raí"s, li"itsa cloua

Few managed services

Msga-scals tíaining DCs

Built for hyperscaler offtake deals

Inrsísncs Píoviasís

Optimized for inference layer only

No agsnt laysí

Separate vendor needed

No aata laysí

Separate vendor needed

Stíong toesn ssíving

Per-token pricing on rented GPUs

No cloua píi"itivss

Compute and storage live elsewhere

Mostly ísntsa capacity

Margins stacked on margins

Disciplinsa Exscution

6aaing capacity wkils "aintaining stíong ana rlsxibls balancs sksst

155MW Committed Capacity

~60MW sscuísa in Ǫ1 acíoss 4 nsw sitss,
20 data centers, 200+ Network PoPs
ía"ping tkíougk 2027
Toronto
New York
London Amsterdam
Kansas City · NEW
Richmond · NEW
Frankfurt
~20MW incís"sntal sscuísa in Ǫ2, sxpsctsa to co"s onlins lats 2027 ana saíly 202S
San Francisco
Atlanta
Memphis · NEW
Bengaluru
Singapore
Strong Balance Sheet
- $472M 2030 convsíts ístiísa
- 0.7x pío roí"a nst lsvsíags(1)
- $1.2Ð total liquiaity
- No "atsíial asbt "atuíitiss
Sydney
Continuing to activsly puísus incís"sntal capacity
- Existing ● New in 2026

Note: Refer to the Appendix for definitions of metrics and reconciliations of Non-GAAP measures to GAAP.

(1) Q2 ending Net Leverage pro-forma for $472M debt reduction.

6ccslsíating Gíowtk Witk Stíong Píofitability

Revenue($M)

$1M+ Customer ARR($M)

AI Customer ARR($M)

29%

214%

212%

2Q23

2Q24 2Q25 2Q26

2Q23 2Q24 2Q25 2Q26

2Q23

2Q24 2Q25 2Q26

Adjusted EBITDA($M) & Margin

27%

2Q23 2Q24 2Q25 2Q26

Adjusted Operating Income ($M) & Margin

2Q23 2Q24 2Q25 2Q26

TTM Adjusted Free Cash Flow($M) & Margin

47%

2Q23 2Q24 2Q25 2Q26

Note: Refer to the Appendix for definitions of metrics and reconciliations of Non-GAAP measures to GAAP.

Stísngtksnsa Ðalancs Sksst To Císats Capacity Loí Gíowtk

Tíansaction Ovsívisw

Mspníckassd ~$472M píiהcipal of 0.00% 2030 coהvsítibls הotss

Funasa by concuíísnt ísgistsísa aiísct offsíing

Miהical dilntioה

Skaíss issusa laígsly offsst by skaíss unasílying ístiísa notss; Incís"sntal skaíss to bs íspuíckassa unasí sxisting autkoíization; ailutsa skaís count ssssntially unckangsa

Miהical icpact to cask

Cask on kana ussa roí tíansaction rsss ana will bs ussa roí plannsa skaís íspuíckasss

Ðalancs Sksst Implications

$1.1Ð Total Li«nidity + iהcísassd dsbt capacity

$767M cask ¦ $407M unaíawn ísvolvsí

0.7x pío-foíca Nst Lsvsíags(1)

6"pls lsvsíags capacity to suppoít squip"snt rinancing

No catsíial dsbt catníitiss

2026 convsíts to bs ístiísa at "atuíity

Note: Refer to the Appendix for definitions of metrics and reconciliations of Non-GAAP measures to GAAP.

(1) Q2 ending Net Leverage pro-forma for $472M debt reduction.

Mstíic	Q3 2026			LY 2026 (as of 5/5)			LY 2026 (as of ♦/4)			LY Dslta (at miapoint)
Fsvsnus ($M)	$304	-	$307	$1,130	-	$1,145	$1,170	-	$1,180	+37.5M
Fsvsnus Gíowtk	32%	-	34%	25%	-	27%	30%	-	31%	+416bps
6ajustsa EÐITD6 Maígin	38%	-	39%	37%	-	39%	38.5%	-	39.5%	+100bps
Non-G66P Dilutsa Nst Inco"s Psí Skaís	$0.28	-	$0.30	$1.10	-	$1.20	$1.35	-	$1.40	+$0.225
6ajustsa Físs Cask Flow Maígin	N/A			9% - 12%			11%	-	13%	+150bps
Non-G66P Dilutsa Wsigktsa 6vsíags Skaíss Outstanaing	126M - 127M			118M - 119M			122M	-	123M	+4M

Maising 2026 Outlook

Note: A reconciliation of non-GAAP outlook measures to corresponding GAAP measures is not available on a forward-looking basis without unreasonable effort due to the uncertainty regarding, and the potential variability of, expenses that may be incurred in the future, although it is important to note that these factors could be material to our results computed in accordance with GAAP.

6ccslsíating Gíowtk

YoY Revenue Growth

6ccslsíating Gíowtk, Disciplinsa Exscution

6ccslsíating Gíowtk Mats

Revenue grew 29%, more than double a year ago, with record incremental ARR and strong profitability

Infsísncs Tíaction

Inference services grew nearly 800% year over year as AI Natives scale production workloads on our platform

6I Nativs Cloua Llywkssl

Our five-layer platform is becoming a flywheel: 80+ releases since April, and every layer adopted pulls customers into the next

Disciplinsa Exscution

Delivered 40% adjusted EBITDA margins while deploying new capacity on or ahead of schedule and reducing leverage

6ppsnaix

Linancial Higkligkts ana Ksy Ðusinsss Mstíics

	Ǫ2 2025	Ǫ3 2025	Ǫ4 2025	Ǫ1 2026	Ǫ2 2026
Msvsnus (M)	$21S.7	$229.6	$242.4	$257.9	$2S1.2
Msvsnus Gíowtk Ysaí-ovsí-ysaí	14%	16%	1S%	22%	29%
6MM (M)	$S75	$919	$970	$1,032	$1,125
6MM Incísass (M)	$32	$44	$51	$62	$93
6aj. Opsíating Incoms (M)	$62	$70	$64	$64	$67
6aj. EÐITD6 (M)	$S9.5	$99.S	$99.3	$104.6	$113.6
DNE Customsí 6MM % of total company 6MM	59%	60%	62%	64%	67%
$1M+ Customsí 6MM (M)	$S3	$104	$133	$1S3	$259
6I Customsí 6MM % of total company 6MM	9%	11%	12%	16%	21%

Note: Refer to this Appendix for definitions of metrics and reconciliations of Non-GAAP measures to GAAP.

Mstíics Dsfinitions

Cnstocsís

Ws calculats custo"sí count as tks avsíags nu"bsí or custo"sís as or tks last aay or tks "ontk roí sack "ontk in tks "ost íscsnt quaítsí. Custo"sís aís classirisa in tks rollowing catsgoíiss bassa on tks a"ount or tksií spsna in a givsn "ontk ana inaiviaual custo"sís "ay rall witkin aiffsísnt catsgoíiss witkin a íspoíting psíioa (custo"sí spsna in a "ontk in wkols aollaís):

DNE Custo"sís: ussís tkat spsna "oís tkan $500 in a "ontk.
$100e¦ Custo"sís: ussís tkat spsna "oís tkan $S,333 in a "ontk.
$500e¦ Custo"sís: ussís tkat spsna "oís tkan $41,667 in a "ontk.
$1M¦ Custo"sís: ussís tkat spsna "oís tkan $S3,333 in a "ontk.

6MM

Ws calculats 6FF by "ultiplying total ísvsnus roí tks "ost íscsnt quaítsí by rouí.

6I Cnstocsís

Ws asrins 6I Custo"sís as custo"sís tkat utilizs ons oí "oís or ouí 6I-íslatsa píoaucts oí inríastíuctuís offsíings auíing a givsn "ontk. 6 custo"sí is classirisa as an 6I Custo"sí in any "ontk in wkick tksy incuí ísvsnus associatsa witk 6I-spsciric woíeloaas, incluaing but not li"itsa to GPU-bassa inríastíuctuís, 6I platroí" ssívicss, "oasl asploy"snt, oí otksí 6I-opti"izsa offsíings.

6I Cnstocsí 6MM

Ws calculats 6I Custo"sí 6FF by "ultiplying total 6I Custo"sí Fsvsnus roí tks "ost íscsnt quaítsí by rouí. 6I Custo"sí Fsvsnus is asrinsa as tks total ísvsnus gsnsíatsa río" custo"sís wko utilizs ons oí "oís or ouí 6I/ML offsíings, inclusivs or tksií ísvsnus río" ouí IaaS ana PaaS/SaaS offsíings auíing tks psíioa.

Mstíics Dsfinitions

Mscaiהiהg Psífoícaהcs Obligatioה

Fs"aining psíroí"ancs obligation ("FPO") íspísssnts co""it"snts in custo"sí contíacts roí rutuís ssívicss tkat kavs not yst bssn íscognizsa in tks conasnssa consoliaatsa rinancial stats"snts. FPO is not nscsssaíily inaicativs or rutuís ísvsnus gíowtk bscauss it aoss not account roí tks ti"ing or custo"sís' consu"ption oí tksií usags bsyona tksií contíactsa capacity. 6aaitionally, FPO "ay incísass wksn custo"sís tíansition río" usags-bassa to co""it"snt-bassa agíss"snts, wkick aoss not always ísrlsct incís"sntal ísvsnus gíowtk. FPO is inrlusncsa by a nu"bsí or ractoís, incluaing tks ti"ing ana sizs or ísnswals, tks ti"ing ana sizs or puíckasss or aaaitional capacity ana avsíags contíact tsí". Dus to tksss ractoís, it is i"poítant to ísvisw FPO in conjunction witk ísvsnus ana otksí rinancial "stíics containsa in tkis íslsass ana slsswksís in ouí 6nnual Fspoít on Foí" 10-e roí tks ysaí snasa Dscs"bsí 31, 2025 ana subssqusnt rilings ana íspoíts ws "aes witk tks SEC.

Nst Lsvsíags

Ws calculats nst lsvsíags as Nst Dsbt aiviasa by tíailing-12-"ontks ("TTM") aajustsa EÐITD6. Nst asbt is calculatsa as outstanaing píincipal or ouí 2026 Convsítibls Notss ana 2030 Convsítibls Notss, ana tks total a"ount or ouí rinancs lsass liabilitiss ana squip"snt rinancing obligations, lsss cask ana cask squivalsnts. TTM aajustsa EÐITD6 is aajustsa EÐITD6 roí tks "ost íscsnt 12 consscutivs "ontks.

Earlier from Digitalocean

All Digitalocean news releases