← Sticker Wall MODEL · DETAIL

Serverless Function-as-a-Service (FaaS) Pay-per-Invocation Platform

1) Pay-per-invocation billing, calculated based on execution duration multiplied by memory tiers; subscription fees for

MODEL

Key Fields

FIELD STAMPS

IndustryCloud Computing

RegionMulti-region

ScaleGiant

ChannelOnline

📌 Background

As cloud computing enters a mature phase, enterprises are prioritizing cost reduction, efficiency, and on-demand elasticity, moving away from the inefficiencies of traditional reserved instances. With the surge in AI inference and agent-based workloads expected by 2026, function computing is naturally suited for event-driven, short-duration tasks. As cold-start optimization matures, it is increasingly covering mid-to-long-tail scenarios. Major cloud providers in both the US and China have positioned Serverless as the core of next-generation infrastructure, with economies of scale continuously driving down the cost per invocation.

👤 Target Customers

Application developers, technology-focused SMEs, and AI application/agent development teams seeking pay-per-use models.

💰 Revenue Streams

1) Pay-per-invocation billing, calculated based on execution duration multiplied by memory tiers; subscription fees for provisioned concurrency, function egress traffic fees, and trigger surcharges; sales of value-added modules such as enterprise-grade monitoring, logging, and observability tools. 2) Training and Workshops: Charging technical teams for intensive training sessions or course packages. 3) Proprietary Integration: Charging deployment and integration fees for clients requiring private cloud or custom system connectivity.

🧮 Cost Structure

Depreciation costs for underlying compute and storage, R&D costs for bandwidth and cooling scheduling, and personnel costs for security, compliance, and operations.

🛡️ Moat

Accumulated expertise in cold-start optimization, deep integration with cloud storage and message queues, and economies of scale that dilute the cost per invocation.

🔑 Keys to Success

Cold-start latency optimization and provisioned concurrency strategies
Breadth of trigger ecosystem and deep integration with cloud services
Granular billing and cost transparency

⚠️ Risks

SLA breaches caused by cold starts
Vendor lock-in disrupted by regulations or standardization
Price wars diluting marginal profits

🏢 Cases

Baidu Cloud CFC
Tencent Cloud SCF
Huawei Cloud FunctionGraph

📊 SWOT Analysis

Strengths

Pay-per-use significantly reduces idle costs for clients
Automatic elastic scaling eliminates maintenance and lowers the barrier to entry
Event-driven architecture is naturally suited for high-frequency short tasks and AI agent calls

Weaknesses

Cold-start latency is unfavorable for real-time tasks
Vendor lock-in leads to high migration costs
Complex state management still requires traditional backend support

Opportunities

High-frequency AI agent calls drive a new wave of growth
Convergence of edge computing and function computing expands use cases
Preferred lightweight entry point for SMEs migrating to the cloud

Threats

Containerization and Kubernetes platforms pose a competitive threat
Large clients building private Serverless environments reduce reliance on public clouds
Price wars compress marginal profit margins

Sources