InfoQ 2025 Technology Adoption Survey Results

Modernization, AI, and Mastering Complexity: The New Mandate for Software Leaders in 2025

Apr 21, 2025

*Total survey respondents = 191

Key Technical Challenges Faced by Software Practitioners in 2025

The latest InfoQ reader survey reveals several recurring technical challenges that are top of mind for senior software engineers, architects, and technical leaders in 2025. Across 191 survey respondents, a few clear themes emerge—legacy modernization, technical debt, AI integration, scalability, and cost management dominate the landscape.

1. Legacy Systems & Technical Debt - The most frequently cited challenge is the burden of legacy systems, often intertwined with technical debt. Respondents highlight difficulties in maintaining aging codebases while simultaneously adopting modern frameworks, tools, and architectures. Many are caught in a cycle of delivering new features while lacking the time or resources to refactor or retire outdated systems. This tension is exacerbated when legacy architectures create bottlenecks or architectural inconsistencies that hinder scaling, deployment speed, and integration with modern services.

2. AI Integration & Adopting AI — especially Generative AI — is both a promising opportunity and a growing pain point. Practitioners report challenges in integrating AI/ML into existing workflows, managing AI/ML model accuracy and training, and understanding the actual ROI of AI initiatives. Others express concerns about keeping up with the fast pace of AI advancements, upskilling their teams, and avoiding hype-driven decision-making. The application of GenAI across the software development lifecycle (SDLC) is still experimental for many, and organizational alignment around it remains immature.

3. Scalability, Reliability & Complexity - Engineering teams are increasingly grappling with the scalability and reliability of distributed systems, often in the context of cloud-native and microservices architectures. These concerns extend to observability, deployment reliability, and the management of cross-team dependencies. As applications become more modular and interconnected, ensuring fault tolerance, consistent system performance, and coherent team collaboration is an escalating challenge.

4.Cloud Complexity and Cloud costs, infrastructure sprawl, and cost optimization are major concerns. Teams are balancing the need for speed and flexibility with the pressure to control operational expenses. Respondents frequently mention the difficulty of managing multi-cloud environments, monitoring cloud spend, and implementing effective FinOps practices.

5. Skill Gaps and Organizational Alignment - Beyond technical hurdles, many cite lack of skilled developers, internal silos, and misalignment between business and engineering as barriers to success. Upskilling, retaining talent, and maintaining high-quality developer experience were also prominent challenges—especially in relation to newer tech like AI and evolving cloud platforms.

Key Challenges in AI/ML Adoption

Organizations face a range of barriers when adopting and operationalizing AI/ML technologies, with skills shortages standing out as the top challenge. Nearly half (48%) of respondents cited a lack of skilled AI/ML professionals as their primary hurdle. This highlights an ongoing talent gap that continues to limit organizations' ability to scale AI efforts.

Another major obstacle is the difficulty in measuring and demonstrating ROI from AI/ML initiatives, selected by 38% of respondents. This suggests that while many companies are experimenting with AI/ML, proving its business value remains elusive, potentially slowing down broader investment and adoption.

Additional significant challenges include:

Integration complexity: 35% report struggles with embedding AI/ML into existing systems and workflows, reflecting technical and organizational silos.
Pace of change: 32% say that keeping up with evolving technologies is a persistent issue, underlining the rapid innovation cycles within AI/ML ecosystems.
Data hurdles: 30% identify difficulties with obtaining and preparing high-quality training data, and another 30% point to the high costs of infrastructure and tooling.

Managing model performance in production (29%) and addressing ethical concerns and bias (18%) were noted but ranked lower compared to operational and technical barriers, suggesting that while ethics is recognized, immediate practical concerns dominate current adoption efforts.

Interestingly, 10% of respondents indicated that none of these challenges significantly affect their organizations, hinting at either higher maturity levels or more limited AI/ML initiatives.

Overall, the survey reveals that successful AI/ML adoption is as much a people and process challenge as it is a technical one. Organizations aiming to lead in AI will need to invest in skills development, ROI measurement frameworks, and modernizing infrastructure to overcome these barriers.

AI Infrastructure Priorities for 2025

Survey data reveals strong momentum toward cloud-based AI adoption across organizations:

Cloud-based AI platforms (59%) such as AWS SageMaker, Google Vertex AI, and Azure ML are the top priority, establishing the cloud as the foundation for AI development and deployment.

AI development environments like Jupyter Notebooks and AI-enhanced IDEs (46%) are also widely used, supporting experimentation and model iteration.

Adoption is more measured for hybrid and on-premises approaches:

20% plan to implement hybrid AI infrastructure, combining cloud and on-prem solutions—indicating a need for flexibility in data governance and performance.

On-premises AI infrastructure (17%) and AI model hosting platforms (20%) are less common, suggesting more targeted or advanced use cases.

Edge AI adoption remains limited (7%), with only niche applications driving current interest.

Notably, 13% of respondents report no plans to adopt AI infrastructure in 2025—reflecting either organizational constraints or a wait-and-see approach.

Key Insight: Cloud-native AI platforms and development tools are becoming the default, while hybrid, on-premises, and edge deployments remain specialized strategies driven by specific operational or regulatory needs.

Cloud-Native Technology Adoption

Survey respondents indicated strong adoption of foundational cloud-native technologies across their organizations:

Containerization (87%), microservices (81%), and infrastructure as code (70%) are firmly established as standard practices.
Kubernetes is widely adopted (68%), though 21% of respondents do not plan to use it—signaling potential complexity or suitability concerns.
Serverless computing (61%) and event-driven architectures (72%) are gaining traction, reflecting demand for scalable and loosely coupled systems.

Adoption is more cautious for advanced or specialized technologies:

Only 29% currently use service mesh, with over half (52%) not planning to adopt—highlighting complexity as a likely barrier.
Cloud-native continuous delivery tools (e.g., ArgoCD, Flux) are in use by 50%, with another 19% planning to adopt—pointing to growing momentum behind GitOps practices.

Key Insight: Core cloud-native patterns are now mainstream, while more complex components like service mesh and CD tools are still emerging, with adoption driven by organizational maturity and operational readiness.

Top Architectural Challenges in Modern Systems

Survey data highlights the top architectural challenges technical leaders are facing today:

Managing complexity is the most critical concern:
Managing dependencies and reducing coupling in complex systems was selected by 54% of respondents, making it the most cited challenge.

Design trade-offs between scalability and simplicity are a major focus:
50% reported difficulty balancing simplicity with scalability, and another 50% emphasized aligning architecture decisions with business goals and timelines.

Adaptability and modernization are ongoing priorities:
41% are focused on maintaining flexibility for future changes and technology adoption, while 38% are transitioning from monolithic systems to microservices or other modern architectures.
Integration of legacy systems remains a notable hurdle for 37%.

Operational resilience and cloud strategy are secondary concerns:
Ensuring high availability and reliability was cited by 27%, while designing for multi-cloud or hybrid cloud environments lags behind at 17%.

Key Insight:
Architectural leaders are primarily grappling with managing system complexity, making pragmatic design decisions, and future-proofing their systems—while operational and multi-cloud challenges, though important, are comparatively lower priorities.

Challenges in Managing Databases and Data Engineering Efforts

Survey respondents highlighted the following concerns and challenges associated with managing databases:

The top concern is managing costs for storage and database infrastructure (62%), followed closely by optimizing query performance and reducing latency (55%).
Ensuring data consistency and accuracy across distributed systems (42%) and integrating and maintaining data pipelines across diverse systems (40%) reflect the complexity of operating in modern, multi-environment architectures.
Scaling databases to handle increasing data volume and user traffic (36%) and adopting and scaling new technologies like AI/ML for data processing (36%) are also prominent challenges.
Ensuring robust data security and preventing breaches was selected by 34% of respondents.

Overall, the data shows that organizations are balancing cost management, performance optimization, system complexity, and technology adoption as they scale their data operations.

Database and Data Engineering Technology Adoption

On the database adoption front, InfoQ survey respondents showed strong current usage of foundational technologies, with growing interest in emerging solutions.

Key-Value Stores (e.g., Redis, Amazon DynamoDB) are the most widely used, with 76% currently using them. An additional 7% plan to adopt within the next 12 months.
NoSQL Databases (e.g., MongoDB, Cassandra, DynamoDB) show similarly high adoption, with 72% currently using and another 7% planning to adopt in the next year.
Streaming Data Platforms (e.g., Apache Kafka) are currently used by 60%, with 10% planning to adopt within 12 months and 5% beyond that.
Cloud-Native Databases (e.g., Amazon Aurora, Google Spanner) are currently used by 59%, with 7% looking to adopt within the next year.
Data Analytics Platforms (e.g., Databricks, Apache Spark, Tableau) are used by 57% today, and show the highest short-term growth potential, with 13% planning adoption within 12 months and 7% beyond.

In contrast, several specialized database types show significantly lower current adoption but are gaining attention:

Document-Oriented Databases are used by 26%, with 15% planning adoption in the future.
Graph Databases are used by only 21%, but 11% plan to adopt within 12 months and 8% beyond.
Vector Databases, a newer category (e.g., Pinecone, Weaviate), have 19% current usage, with 14% planning near-term adoption—highlighting early momentum in support of AI/ML workloads.

Overall, the data suggests continued investment in established platforms alongside growing interest in technologies enabling real-time processing, AI/ML, and advanced analytics.

Challenges in Delivering High-Quality Software

Respondents identified several key challenges in delivering high-quality software at their organizations.

The most significant issue was managing technical debt while innovating on new features (69%), highlighting the ongoing tension between maintaining existing systems and driving forward progress.

Half of the respondents also pointed to ensuring effective testing and minimizing bugs in production (50%) as a major hurdle, while keeping up with rapidly evolving tools, frameworks, and technologies (46%) remains a persistent challenge.

Other commonly reported difficulties included:

Integrating legacy systems with modern technologies (41%)
Managing cloud infrastructure costs and complexity (39%)
Achieving consistent deployment frequency and reliability (35%)

Additionally, adopting and scaling DevOps practices across teams (31%) and scaling and optimizing software to handle increased user demand (26%) were seen as notable obstacles.

Overall, the data reflects that while innovation is a top priority, technical debt, system integration, and operational complexity continue to place significant demands on software teams.

Conclusion

In 2025, senior software leaders are focused on modernizing legacy systems, managing technical debt, and figuring out how to use AI effectively. While cloud-based AI platforms are becoming the norm, many teams still struggle with the skills, tools, and processes needed to make AI work in the real world. Core cloud-native technologies like containers and microservices are widely adopted, but newer, more complex tools are seeing slower uptake. Across the board, teams are balancing the need to innovate quickly with the realities of cost control, system complexity, and a constant need to learn and adapt.

Audience Demographics

The survey also included specific demographic questions on:

Professional Focus / Role in Organization
Company Size
Seniority (e.g. Do you manage/supervise others?)

If you are a software vendor that would like to see the complete results from this survey (including data from the ‘audience demographics’ and ‘commercial product adoption’ questions), please email us at sales@infoq.com.

InfoQ Developer Marketing Newsletter (Quarterly)