The theme of CLOUD 2026 is:

Agentic AI as a Service for Cloud Computing

Agentic AI systems can understand objectives, construct plans, access models and data, invoke tools, collaborate with other agents, and execute complex tasks. The cloud provides the infrastructure and platform capabilities required to make these systems scalable, continuously available, and broadly accessible.

For the cloud computing community, the central question is:

How can cloud infrastructures and platforms host, scale, coordinate, secure, and govern agentic AI as dependable services?

Agentic workloads differ from conventional cloud applications. They may dynamically select models and tools, maintain persistent state and memory, generate unpredictable execution paths, and interact with external systems. They may also require combinations of CPUs, GPUs, AI accelerators, storage, networking, vector databases, sandboxes, and external services.

Supporting these workloads transforms cloud computing:

• From request–response applications to long-running agentic tasks

• From predefined workflows to dynamic execution plans

• From homogeneous workloads to heterogeneous AI resource demands

• From CPU-centric scheduling to accelerator-aware orchestration

• From reactive autoscaling to intent- and workload-aware provisioning

• From application monitoring to end-to-end agent-task observability

• From conventional multitenancy to secure and isolated agent execution

• From infrastructure availability to dependable task completion

CLOUD 2026 invites researchers and practitioners to develop the cloud foundations required for scalable, reliable, secure, efficient, and responsible Agentic AI as a Service.

CLOUD focuses on cloud infrastructure, platforms, runtimes, resource management, operations, and cloud-specific system engineering.

Research whose primary contribution concerns Web interfaces, general service engineering, AI models, big-data analytics, edge-first systems, Internet of Things applications, cognitive computing, or blockchain should be submitted to the corresponding SCF conference unless it makes a clear systems contribution to cloud computing.

Topics include, but are not limited to:

1. Agentic Cloud and AI Infrastructure

• Cloud platforms and runtimes for Agentic AI

• Agent execution, isolation, state, and lifecycle management

• Cloud infrastructure for AI training and inference

• Resource support for models, memory, tools, and retrieval

• Distributed execution of agentic and multi-agent workloads

• Serverless and event-driven execution for AI agents

2. Cloud Architectures and Platforms

• Infrastructure, Platform, Software, Function, and AI as a Service

• Virtual machines, containers, unikernels, and serverless computing

• Public, private, hybrid, multi-cloud, and federated clouds

• Cloud control planes and orchestration platforms

• Cloud storage, networking, and distributed systems

• Cloud-native application platforms and DevOps

3. Cloud Resource Management and Scheduling

• Resource provisioning, placement, scheduling, and autoscaling

• CPU, GPU, and accelerator-aware resource management

• Workload modeling, prediction, and capacity planning

• Performance isolation and multitenant resource sharing

• Cost-, performance-, and SLO-aware optimization

• Resource federation and cross-cloud scheduling

4. Cloud Reliability and Operations

• Cloud availability, resilience, and fault tolerance

• Monitoring, observability, tracing, and diagnostics

• AIOps and autonomous cloud operations

• Failure prediction, self-healing, and disaster recovery

• Workload migration, replication, and live system evolution

• Cloud performance evaluation and benchmarking

5. Cloud Security, Privacy, and Governance

• Cloud security and zero-trust architectures

• Identity, authentication, authorization, and access control

• Confidential computing and trusted execution environments

• Secure multitenancy and workload isolation

• Cloud data privacy, sovereignty, compliance, and auditability

• Security and sandboxing for tool-using AI agents

6. Sustainable and Industry Cloud Computing

• Energy- and carbon-aware cloud computing

• Sustainable data centers and resource optimization

• Cloud economics, pricing, accounting, and FinOps

• Industry-specific and sovereign cloud platforms

• Enterprise cloud migration and modernization

Large-scale cloud deployments and operational experience

All submissions must report original and unpublished work and must not be under review elsewhere at the time of submission.

Each manuscript will be reviewed by at least three program committee members. Papers will be evaluated according to:

• Relevance to CLOUD

• Originality and significance

• Technical or practical quality

• Strength of evaluation

• Clarity and organization

• Reproducibility

• Research or operational impact

Papers should position their contributions accurately within the relevant international literature. Authors are encouraged to cite CLOUD and other conferences or journals when genuinely relevant. Citation of any particular venue is not required.

Research Track and Application and Industry Track manuscripts are limited to 15 pages in the Springer LNCS format. Short Paper Track manuscripts are limited to 8 pages. Authors may purchase up to two additional pages, subject to the applicable publication fee. Special Paper Track submissions should follow the limits stated in the corresponding call.

Authors must use the official Springer LNCS templates:

• Microsoft Word template

• LaTeX template

Electronic submission in PDF or Word format is required. Manuscripts that do not follow the required format or exceed the permitted page limit may be returned without review.

A Research Track submission that primarily presents a deployed application or operational solution may be recommended for consideration by the Application and Industry Track.

At least one author of each accepted paper must register for CLOUD 2026 and present the paper at the conference.

The CLOUD 2026 technical program includes four submission tracks.

Research Track

The Research Track welcomes original and technically rigorous contributions that advance the theories, architectures, algorithms, systems, or engineering foundations of cloud computing.

Research papers should:

• Clearly define the cloud-computing problem

• Explain the novelty relative to relevant literature

• Present technically sound methods or system designs

• Provide rigorous and reproducible evaluation

• Use appropriate workloads, baselines, and performance measures

• Discuss limitations and threats to validity

• Demonstrate a generalizable cloud-systems contribution

Application and Industry Track

The Application and Industry Track focuses on deployed cloud systems, operational innovations, migration experience, industry platforms, and evidence-based cloud practices.

Papers should:

• Describe a significant real-world cloud problem

• Explain the architecture and implementation

• Present deployment evidence or measurable outcomes

• Discuss scalability, reliability, security, cost, and operational constraints

• Identify lessons that can be transferred to other environments

• Clarify the practical or system-level innovation

Papers in this track do not need to introduce a new algorithm. However, they should contribute through cloud-system architecture, large-scale deployment, operational evidence, performance improvement, or generalizable engineering knowledge.

Short Paper Track

The Short Paper Track welcomes focused contributions, including:

• Promising early-stage cloud research

• Novel architectures or system concepts

• Emerging challenges and research visions

• Cloud datasets, benchmarks, tools, and artifacts

• Focused empirical findings or negative results

• Concise deployment and operational reports

Short papers must clearly state their contribution, evidence, relationship to prior work, and limitations.

Special Paper Track

The Special Paper Track addresses designated emerging topics and community priorities announced by CLOUD 2026. Submissions should follow the requirements of the corresponding special-track call.

Accepted and presented CLOUD 2026 papers will be published in the conference proceedings, which are scheduled to appear in Springer’s Lecture Notes in Computer Science (LNCS) series.

Springer makes LNCS proceedings available to major abstracting and indexing services. Final inclusion of a volume or individual paper is determined by the policies and selection criteria of each indexing service.

Under the current SCF publication arrangement, the CLOUD 2026 proceedings are expected to become freely accessible through SpringerLink four years after publication.

Extended versions of selected papers may be invited to relevant partner journals. Journal submissions must include substantial extensions and will undergo the journal’s independent editorial and peer-review process.

CLOUD 2026 will recognize outstanding contributions through:

• Best Paper Award

• Best Student Paper Award

For the Best Student Paper Award, the first author must be a full-time student.

CLOUD has served as an international forum for exchanging advances in cloud-computing research, systems, and practice. Its scope has evolved alongside cloud computing itself—from virtualized infrastructure and utility computing to cloud-native platforms, containers, serverless systems, heterogeneous accelerators, autonomous operations, and AI infrastructure.

Recent CLOUD conferences were held in Seattle in 2018, San Diego in 2019, online with satellite activities in Shenzhen in 2020 and 2021, Hawaii in 2022, Shenzhen in 2023, Bangkok in 2024, and Hong Kong in 2025.

CLOUD 2026 continues this development by examining the infrastructure and platform foundations needed to deliver Agentic AI as a scalable and dependable cloud service.

CLOUD is a member of the Smart Computing Frontier, a federation of ten co-located international conferences.

Within SCF, CLOUD focuses on the computing foundation of smart systems: cloud infrastructure, platforms, runtimes, resource management, operations, security, and sustainability.

Its scope is complementary to the other SCF conferences:

• ICWS focuses on Web interfaces, APIs, protocols, and Web-based services.

• SCC focuses on service modeling, engineering, lifecycle management, and value creation.

• BigData focuses on large-scale data management and analytics.

• AIMS focuses on AI models, multimodal intelligence, and AI-enabled services.

• EDGE focuses on computing systems located near data sources and users.

• ICIOT focuses on connected devices and Internet of Things systems.

• ICCC focuses on cognitive computing.

• ICBC focuses on blockchain technologies and trusted ecosystems.

• METAVERSE focuses on immersive and persistent digital environments.

This structure enables cross-conference collaboration while preserving CLOUD’s distinctive identity as the conference devoted to cloud computing systems and infrastructure.