top of page
Business & Enterprise
Cloud Technology
Cloud Infrastructure
Scaleway
Won European builders by pairing developer-friendly cloud services with transparent pricing, sovereign infrastructure and accessible GPU capacity, despite AWS, Azure and Google Cloud controlling broader enterprise ecosystems.
1
MODEL
BUSINESS MODEL
IaaS
- Developers and companies provision virtual instances, bare metal, GPUs, storage, databases, Kubernetes and managed AI services through self-service APIs and console.
- Scaleway operationalizes infrastructure through European data centers, standardized compute, block storage, networking and automated billing.
- It uses transparent pricing—including included egress and IPv6 on listed virtual instances—to reduce the billing uncertainty that often slows cloud adoption.
- GPU instances and managed inference turn scarce accelerator capacity into higher-value AI workloads.
- Sovereignty, GDPR alignment and European data residency create a differentiated wedge for startups, scale-ups and regulated organizations.
1. Start with a foundational resource your target builder consumes repeatedly—compute, storage, connectivity or specialized capacity.
2. Remove one major adoption fear, such as unpredictable egress fees, complex billing or data-residency risk.
3. Make provisioning self-serve and API-first so technical buyers can evaluate without sales.
4. Add a scarce or high-value resource—such as GPUs—only when the core platform can provision, monitor and support it reliably.
5. Use transparent unit economics to build trust, then expand into managed services that reduce customer operational burden.
6. Avoid competing on catalog breadth before utilization and support economics are proven; the common mis-build is launching many services while losing money on idle capacity and support.
DISTRIBUTION MODEL
Scaleway traces its roots to 1999 through Online SAS, began renting dedicated servers in 2006, and evolved into a European cloud and AI infrastructure provider owned by Iliad Group. It differentiated through European data centers, GDPR-aligned data handling, no egress fees on listed virtual instances and straightforward pay-as-you-go pricing. Its GPU portfolio includes NVIDIA L4, L40S, H100 PCIe and B300 instances, with L4 pricing from €0.79 per hour and managed inference starting at €0.93 per hour. The strategic bet was that European developers and AI builders wanted cloud capability without hyperscaler complexity, opaque egress costs or non-European data control.
HOW THEY BUILT IT
HOW TO ARCHITECT IT
Self-Serve / Website, SEO, Developer/API-Led, Community-Led
HOW THEY OPERATIONALIZED
HOW TO REPLICATE WHAT WORKED
dm
- Scaleway used self-serve signup, console onboarding and API documentation to convert developers searching for European cloud infrastructure.
- SEO targeted cloud servers, GPU instances, managed Kubernetes, European hosting and AI-inference queries.
- Developer communities, startup ecosystems and technical content reinforced its price-transparency and sovereignty positioning.
- Partnerships with AI tooling, open-source communities and technology providers extended its reach into AI and platform-engineering audiences.
What worked: Removing a specific billing or sovereignty fear can differentiate a smaller infrastructure provider even when it cannot match hyperscaler breadth.
Trap if copied blindly: Transparent pricing works only when the provider’s underlying capacity, utilization and support costs are controlled; otherwise the promise becomes a margin problem rather than a growth advantage.
| PATTERNS OF THIS MODEL
PATTERNS IN REGIONAL CLOUD CHALLENGERS:
1. A removed fee can become the positioning. Eliminating egress charges or billing complexity gives technical buyers a concrete reason to evaluate a smaller provider.
2. Sovereignty is a trust wedge, not a feature. Data residency and regulatory alignment matter most where compliance, privacy or public-sector procurement shapes the purchase.
3. Specialized capacity creates premium expansion. GPUs, managed inference and AI tooling raise revenue per customer beyond commodity virtual machines.
THE WEAKNESS: Regional challengers face a structural tradeoff between infrastructure investment and service breadth; without enough utilization, capital intensity overwhelms the price advantage.
What companies with this model reveal
| OPPORTUNITY INTELLIGENCE
GOLDMINE 1 — SOVEREIGN AI PLATFORM. The opportunity is packaging European GPU capacity with data residency, model governance, private networking and compliance-ready deployment; the buyer wants an AI operating environment, not merely an H100 rental.
GOLDMINE 2 — PREDICTABLE AI INFERENCE ECONOMICS. Startups fear variable GPU bills; fixed-capacity inference plans, autoscaling guardrails and cost dashboards can convert price-sensitive builders into long-term customers.
GOLDMINE 3 — VERTICAL AI FACTORIES. Media, healthcare, industrial and public-sector customers need domain-specific pipelines—rendering, imaging, simulation or document intelligence—rather than raw infrastructure.
THE PIT — GPU CAPACITY BET. Buying accelerators ahead of demand creates depreciation, power and utilization risk; GPU supply cycles can turn quickly.
THE SECOND PIT — HYPERSCALER FEATURE GAP. A regional provider cannot win broad enterprise workloads if it lacks the managed data, security and integration ecosystem buyers expect.
MOVE WITH CAUTION — Do not use “sovereign” as a slogan without operational proof: residency, certifications, support and exit paths must be explicit and auditable.
Untapped Business Model / Gaps / Goldmines / Pits
Patterns & Insights
2
MARKET
mkt mt es
Scaleway entered cloud infrastructure against AWS, Azure, Google Cloud, OVHcloud and Hetzner. Its wedge was not catalog breadth but a European, developer-friendly alternative with transparent pricing, no egress fees on listed virtual instances and accessible GPU capacity. By offering H100, L40S, L4 and B300 options alongside managed inference, it reframed the choice from “which hyperscaler?” to “which infrastructure gives European builders control, cost clarity and AI access?”
MARKET TYPE
Red Ocean
ENTRY STRATEGY
Direct Entry
WHY THEY WON
Cloud infrastructure is a mature, crowded market dominated by hyperscalers with massive capital and ecosystem advantages. Scaleway’s opening was the European builder or AI team that wanted cloud capability but objected to opaque egress pricing, non-European data control or complex enterprise procurement. The structural blind spot was the gap between commodity VPS providers and hyperscalers: developers wanted modern cloud services, but with simpler economics and local trust.
EXECUTION
Scaleway entered through its own hosting and server business, beginning with web hosting and domain registration in 1999 and adding dedicated servers in 2006. Its initial bet was that European developers and businesses would pay for reliable infrastructure from a local provider. This was viable because it could standardize servers, control data-center operations and build direct self-serve distribution before hyperscalers dominated the European cloud market.
FOOTHOLD STRATEGY
Beachhead Strategy
Scaleway’s beachhead was European developers, startups and technical SMBs that needed affordable servers, cloud instances and storage. These buyers were reachable through search, developer communities and self-serve onboarding; they valued transparent pricing, European data residency and API-driven provisioning. The path outward was managed Kubernetes, databases, AI accelerators and enterprise-grade sovereign cloud workloads.
GROWTH CAMPAIGN
CAMPAIGNS THAT WORKED
SO WHAT: Scaleway’s strongest commercial mechanism is converting billing transparency and European trust into a first workload, then expanding into higher-value AI capacity.
KEY LEARNING: In crowded infrastructure markets, a smaller provider wins by making one hidden cost or trust concern disappear.
HOW TO USE: Identify the buyer’s biggest cloud anxiety—egress fees, lock-in, residency or GPU cost—then make its elimination central to the offer.
KEY LEARNING
In a red-ocean cloud market, technical buyers compare performance, price predictability, compliance and ecosystem depth. A regional challenger must make its tradeoff explicit: fewer services in exchange for simpler economics, local control and specialized capacity.
gc
Market Context
| MARKET INTELLIGENCE
PATTERNS IN TRANSPARENT-PRICING PLUS DEVELOPER-ECOSYSTEM GROWTH:
1. Removed fees create a memorable wedge. “No egress fees” is easier to remember and evaluate than a broad feature list.
2. Developer self-service reduces acquisition cost. APIs, documentation and instant provisioning let engineers adopt before procurement engages.
3. AI capacity creates a premium expansion path. GPU instances and managed inference turn infrastructure into higher-value workloads.
4. THE WEAKNESS: Developer-led growth can attract short-lived experiments; without workload retention, the provider pays for capacity that does not generate durable revenue.
| MARKET ENTRY PLAYBOOK
Use direct entry when the buyer can self-serve and your cost or trust advantage is real. It fails when the buyer needs broad enterprise integration, global compliance coverage or managed services that your operating model cannot support profitably.
How to enter
| FOOTHOLD STRATEGY PLAYBOOK
Choose a beachhead where the buyer is technical, price-sensitive and able to operate infrastructure independently. Expand into managed and AI services only after core provisioning, utilization and support economics are stable.
How to get the first strong position
MARKET PATTERNS & PLAYBOOK
3
MONEY
money rev pri
Scaleway bills virtual instances hourly and GPU capacity by the minute under pay-as-you-go terms. A GP1-XS instance costs €0.09282 per hour, roughly €67.75 per month, while an L4-1-24G GPU instance costs €0.79 per hour, roughly €574.87 per month. Managed inference starts at €0.93 per hour for Llama 3.1 8B on an L4 instance. The meter is consumed infrastructure and AI capacity, with egress and IPv6 included in listed virtual-instance prices.
REVENUE MODEL
Usage-Based, Subscription
PRICING MODEL
Usage-Based Pricing, Tiered Pricing, Value-Based Pricing
WHY THEY WON
Customers pay for compute hours, GPU minutes, storage, databases, Kubernetes and managed inference. Revenue expands as workloads remain live, data grows and AI usage increases; transparent pricing reduces adoption friction but requires tight control over capacity utilization and support cost.
Scaleway uses pay-as-you-go pricing by hour for virtual instances and by minute for GPUs, with listed egress and IPv6 included on many virtual instances. A current anchor is €0.79 per hour for an L4-1-24G GPU instance and €0.93 per hour for managed Llama 3.1 8B inference. The meter is consumed infrastructure and AI capacity, making cost predictable at the resource level.
TARGET AUDIENCE
European startups and scale-ups (virtual instances/Kubernetes; want transparent cloud economics); AI developers and research teams (GPU/managed inference; want accessible accelerator capacity); regulated enterprises and public sector (sovereign cloud; want GDPR-aligned data residency and control).
CUSTOMER BUYING BEHAVIOUR
The trigger is a new application, cloud-cost review, AI experiment, data-residency requirement or need for GPU capacity. Developers evaluate through documentation, pricing calculators, trials and community recommendations, then self-provision. Adoption begins with a virtual instance or GPU workload and expands through Kubernetes, storage, databases and managed inference; enterprise buyers add procurement, compliance and architecture review.
| PRICING INTELLIGENCE
What makes this model effective & make customers pay
Builders are not buying cloud resources; they are buying the ability to ship without fearing a surprise bill or losing control of their data.
RULE 1 — PRICE THE ADOPTION FEAR AWAY. Removing egress uncertainty can matter more than a marginal compute discount.
RULE 2 — PRICE SCARCE CAPACITY CAREFULLY. GPU pricing must reflect utilization, power, depreciation and demand volatility—not merely GPU list cost.
RULE 3 — EXPAND FROM COMPUTE TO OPERATIONS. Managed inference and Kubernetes reduce customer burden and create higher-value revenue.
RULE 4 — WATCH IDLE CAPACITY. A GPU that is not utilized is a depreciating, power-consuming liability.
The buyer is really buying predictable infrastructure and AI capacity, proven by the engineering, cloud or AI-compute budget.
PRICE & REVENUE
| Revenue Risk - The biggest threat to revenue stability
GPU utilization risk: Accelerators are expensive, power-intensive and prone to demand cycles; idle capacity destroys margin.
Hyperscaler competition: AWS, Azure and Google Cloud can bundle AI services, enterprise support and global compliance.
Price transparency: Publishing rates invites direct comparison and limits pricing flexibility.
Churn from experiments: Developers may run short AI tests without creating durable workloads.
Sovereignty niche risk: A European positioning can constrain global expansion if compliance, support and capacity do not scale.
Scaleway is private and Iliad-owned; its 2024 Impact Report is cited by third parties as showing approximately €15 million revenue, but this figure should be treated as company-reported and not independently audited ARR.
Where the model can break
4
MOTION
motion ge cs
Scaleway’s self-serve console, API, transparent pricing and pay-as-you-go GPU billing support product-led adoption for developers and startups. Its sovereign-cloud, enterprise and large GPU deployments require consultative sales, architecture support and compliance discussions. The primary motion is product-led; sales becomes important when capacity commitments, residency requirements or enterprise governance are involved.
GROWTH EXPANSION MODEL
Product Line Expansion, Geographic Expansion
HOW THEY EXPAND
Scaleway expanded from web hosting and dedicated servers into virtual instances, bare metal, storage, databases, Kubernetes, managed inference and GPU infrastructure. It expanded geographically across French, Dutch and Polish data centers, strengthening European coverage. It expanded the platform through AI accelerators, managed inference and developer tooling, moving from commodity infrastructure toward AI-native workloads. The dependency is capacity utilization: every new GPU class and region must generate enough demand to justify power, depreciation and operations.
COMPETITIVE STRATEGY
Differentiation
HOW THEY COMPETE
Against AWS, Azure and Google Cloud, Scaleway focuses on European builders, transparent economics and AI accessibility rather than catalog breadth. Against Hetzner, it differentiates through broader managed cloud and AI services. Against OVHcloud, it emphasizes developer experience and GPU-led AI positioning. Its constraint is that hyperscalers can outspend it on global capacity, enterprise sales and ecosystem depth.
GROWTH ENGINE
GTM
Content Flywheel
Developers discover Scaleway through search, documentation and community recommendations. Transparent pricing and self-serve provisioning activate them; a live workload expands into storage, Kubernetes and databases. AI demand pulls customers into GPU instances and managed inference, increasing revenue per account. The loop weakens if workloads are short-lived, GPU utilization remains low, or customers need hyperscaler-level managed services.
Scaleway runs a product-led, developer-first GTM with a sales-assisted enterprise layer. Self-serve pricing, APIs and documentation convert European builders into paying infrastructure customers; transparent egress policy and sovereignty positioning reduce adoption objections. GPU and managed-inference offers expand accounts into AI workloads, while enterprise sales handles residency, compliance and larger capacity commitments. The transferable components are a removed-fee wedge, developer self-service and premium capacity expansion; the model requires disciplined utilization and support economics.
SUSTAINING MOATS
High Customer Lock-In
moat
The moat is Scaleway’s European infrastructure footprint and operational ability to provision compute, storage, networking and AI capacity with local data control. Its owned or controlled data-center relationships, European positioning and developer tooling make it a credible alternative for buyers whose requirements hyperscalers address less directly.
| MOAT INTELLIGENCE
THE STANDARD: A regional cloud moat lasts when locality, compliance and operational reliability create value that a global provider cannot easily replicate without changing its own economics.
RULE 1 — DATA RESIDENCY CREATES TRUST. European locations and GDPR alignment matter to regulated and sovereignty-sensitive buyers.
RULE 2 — OPERATIONAL RELIABILITY PROTECTS THE WEDGE. Transparent pricing only works if provisioning, uptime and support remain dependable.
RULE 3 — SPECIALIZED CAPACITY CREATES EXPANSION. GPUs and managed inference deepen the customer relationship beyond commodity compute.
THE SIGNAL: Scaleway’s GPU portfolio, European data centers and no-egress positioning support its differentiation; the moat does not protect against GPU supply cycles, underutilization or hyperscalers offering stronger sovereign and AI services.
Why this company remains defensible
ARR & TAKEAWAY
ARR Journey - what to do at each stage
PRE-$1M REVENUE — PROVE LOCAL INFRASTRUCTURE. Sell web hosting and domains to European developers and businesses; the key move was establishing reliable local infrastructure. WATCH: capital discipline and support capacity.
$1–5M REVENUE — STANDARDIZE DEDICATED SERVERS. Add server rental in 2006 and automate provisioning for developers and growing businesses.
$5–10M REVENUE — BUILD CLOUD-NATIVE SERVICES. Expand from bare metal into virtual instances, storage and developer APIs.
$10–50M REVENUE — ADD MANAGED PLATFORM SERVICES. Introduce Kubernetes, databases, managed services and broader cloud workflows.
$50–100M REVENUE — ENTER AI INFRASTRUCTURE. Add H100, L40S, L4 and B300 GPU options plus managed inference, creating a premium AI workload layer.
$100M+ REVENUE — NOT REACHED / UNVERIFIED. Third-party reporting cites roughly €15 million 2024 revenue from Scaleway’s Impact Report; this is not independently audited public ARR.
Rule: In regional cloud infrastructure, prove utilization and operational reliability before expanding into capital-intensive AI capacity.
COPY PLAYBOOK : What Worked → What Failed → What to Replicate → What to Avoid
THE STANDARD: Build a regional infrastructure challenger by removing a major adoption fear, winning technical buyers through self-service, then expanding into scarce or higher-value capacity.
SEQUENCE: 1. Identify the buyer’s biggest infrastructure objection—cost opacity, residency, latency or complexity; prerequisite: you can remove it operationally, not just in messaging. 2. Make the core resource self-serve and API-first; gate: a technical buyer can deploy and evaluate without sales. 3. Add specialized capacity or managed services only after utilization and support economics are stable; gate: the new offering improves revenue per customer without creating idle-capacity risk.
WORKED: Scaleway used transparent pricing, European residency and GPU/managed-inference expansion to differentiate from larger cloud providers.
CAUTION: Do not buy GPUs ahead of demand visibility; do not claim sovereignty without auditable residency, compliance and operational controls. Scaleway’s reported 2024 revenue is not independently verified ARR, so replication should be measured by workload retention, GPU utilization and contribution margin.
bottom of page