The Modern Tech Stack in 2025: What 3,000 Developers Actually Use (And Why It Matters)

The technology landscape in 2025 looks dramatically different than it did just 18 months ago. AI coding assistants have gone from experimental tools to essential parts of developers' daily workflows. TypeScript has cemented its position as the most-used programming language. And perhaps most surprisingly, tools that were founded just 2-3 years ago are now competing head-to-head with decade-old industry standards.

Based on comprehensive survey data from over 3,000 software engineers, engineering leaders, and technical professionals across companies ranging from 2-person startups to Big Tech giants, we now have unprecedented insight into what developers are actually using in 2025—and more importantly, what they love and hate about their tools.

Whether you're building a team, modernizing your infrastructure, or simply trying to stay current with industry trends, understanding these patterns is crucial. Let's dive into what's really happening in the trenches of software development today.

The AI Revolution: GitHub Copilot vs. The New Guard

Perhaps the most dramatic shift in the developer tooling landscape is the explosion of AI-powered coding assistants. The data reveals that 85% of developers now use at least one AI tool as part of their daily workflow—a staggering adoption rate for technology that was barely on the radar three years ago.

The AI Tool Landscape

GitHub Copilot remains the dominant player, with over 50% of survey respondents using Microsoft's AI coding assistant. This represents remarkable growth: GitHub Copilot crossed 20 million users in July 2025, adding 5 million users in just three months. The tool is now used by 90% of Fortune 100 companies, demonstrating its enterprise dominance.

But the real story is the rise of Cursor, an AI-powered IDE that launched only in 2023. Despite being barely two years old and spending $0 on marketing, Cursor is already the second most-mentioned AI tool and the second most-used IDE overall. The company has reportedly generated over $500 million in annualized recurring revenue and captured approximately 18% market share among paid AI coding tools.

This meteoric rise represents one of the fastest go-to-market successes in developer tools history. As one senior engineer noted in the survey: "Cursor being the third-most loved product is a standout achievement, especially because the product is barely two years old."

The Claude vs ChatGPT Battle

Another surprising finding: Claude is rapidly gaining ground on ChatGPT among developers. While ChatGPT still leads with 803 mentions compared to Claude's 533, this represents a massive shift from just one year ago when ChatGPT had 8x as many users. Anthropic's release of Claude Sonnet 3.5, 3.7, and 4.0—which excel at coding tasks—has clearly resonated with technical audiences.

Claude Code also showed impressive traction despite only being released on May 22, 2025, just days before the survey closed. That it racked up significant mentions while mostly available through a private beta suggests strong early momentum that has only accelerated since.

The Enterprise vs Startup Divide

Tool adoption varies significantly by company size:

Larger companies favor GitHub Copilot: With few exceptions, the bigger the company, the more likely developers use Microsoft's offering, likely due to enterprise contracts and IT policies
Smaller companies embrace alternatives: Cursor, ChatGPT, Claude, and Zed all see higher adoption at startups and small businesses, possibly because of more flexible tool policies
AI tool restrictions increase with scale: The biggest companies (10,000+ employees) are most likely to standardize on a single AI tool, while smaller companies allow developers to use multiple tools

This pattern highlights a crucial insight for technical hiring: if you want to attract AI-native developers, your tool policies matter. As we explored in our analysis of AI engineer hiring trends, new grads and junior developers are the most naturally fluent with AI tools—but only if you let them use what they know.

Programming Languages: TypeScript's Total Dominance

The programming language landscape in 2025 is less about disruption and more about consolidation around proven winners.

TypeScript Takes the Crown

TypeScript is now the most-mentioned programming language, surpassing even JavaScript. This aligns with broader industry trends: over 70% of new React projects use TypeScript, and the TypeScript compiler exceeds 60 million downloads per week—up from 20 million in 2021.

Why the surge? Type safety has moved from "nice to have" to "essential" as applications grow more complex. Teams adopting TypeScript report 30-40% reductions in runtime errors, and the combination of React + TypeScript has become the de facto standard for modern web applications.

Python's Renaissance

Python holds strong as the second most-used language, experiencing what can only be described as a renaissance. It's not just for data scientists anymore—Python has become the "shared language of choice" among software engineers, data scientists, and ML/AI engineers. As AI/ML workloads proliferate, Python's dominance in this space ensures its continued growth.

Mobile Platform Maturity

On mobile, the data shows clear winners:

Swift has 6x as many mentions as Objective-C for iOS development
Kotlin is the overwhelming choice for Android, having been Google's preferred language since 2019

Both languages were announced roughly a decade ago (Swift in 2014, Kotlin's Android adoption in 2017), and they've now fully displaced their predecessors for new development.

The Most Loved Languages

Beyond usage, the survey also measured which languages developers love:

Ruby on Rails: The 5th most popular in usage, but 3rd most loved—developers who use it really appreciate its expressiveness
Elixir: 16th in usage but 10th in love—a strong signal that it delivers excellent developer experience
TypeScript, Python, and Go all show strong positive sentiment

Interestingly, no language received significantly more negative than positive mentions, suggesting that popular languages in 2025 are all reasonably well-designed. The days of being forced to use universally-despised languages seem to be behind us.

Database Dominance: PostgreSQL Quietly Ate the World

If there's one technology that best represents "boring technology that just works," it's PostgreSQL. And in 2025, boring has won.

PostgreSQL's Overwhelming Lead

One in three respondents use PostgreSQL—and no other database technology comes close. PostgreSQL is now used by 49% of developers globally and has been voted the most popular database for two consecutive years in Stack Overflow surveys.

The market data backs this up:

PostgreSQL holds approximately 17% of the overall relational database market
It's the second most-used open-source database after MySQL
51% of organizations report using PostgreSQL more today than a year ago
The open-source database market (led by PostgreSQL) is growing at over 35% annually

Why PostgreSQL? It's open source, free to use, battle-tested, and remarkably extensible. Need full-text search? There's an extension for that. Need vector support for AI embeddings? Pgvector has you covered. Need hypothetical index analysis? HypoPG exists.

The Long Tail of Databases

While PostgreSQL dominates, the survey revealed an astonishing 35 different databases mentioned by at least 6 respondents each. From traditional SQL databases (MySQL, Oracle, SQL Server) to NoSQL options (MongoDB, DynamoDB, Redis), from data warehouses (Snowflake, Databricks, BigQuery) to specialized solutions (ClickHouse for analytics, InfluxDB for time-series), developers have never had more choice.

Vector Databases: The Dog That Didn't Bark

Despite the AI boom, dedicated vector databases like Pinecone and Weaviate showed surprisingly low adoption. Why? Because relational databases added vector support that works well enough. Most LLM applications can simply store embeddings as vectors in PostgreSQL (with the Pgvector extension) or MongoDB (with Atlas Vector Search), eliminating the need for yet another database in the stack.

This pattern reinforces a key lesson: the pragmatic choice for databases isn't "which specialized tool is best?" but rather "what would prevent us from just using PostgreSQL?"

Frontend Frameworks: React's Continued Dominance

The frontend landscape has consolidated around clear winners, with less churn than other categories.

React's Unassailable Position

React continues to dominate frontend development with similar market share to 2022, when 76% of developers used it. The combination of React's component model, vast ecosystem, and backing from Meta has created a durable moat.

Next.js leads meta-frameworks by a wide margin (163 mentions vs 29 for second-place Nuxt), providing server-side rendering, static site generation, and API routes that React alone doesn't offer. Built by Vercel, Next.js has become the default choice for production React applications.

The CSS Revolution: Tailwind's Rise

Tailwind CSS has emerged as the preferred styling solution with 170 mentions, significantly ahead of traditional frameworks like Bootstrap (52 mentions). The utility-first approach has won over developers who appreciate not having to write custom CSS while maintaining design flexibility.

TypeScript + React: The Power Couple

The marriage of TypeScript and React has become industry standard. As noted earlier, over 70% of new React projects use TypeScript, and teams report significant reductions in bugs and improvements in developer productivity. The combination provides:

Type safety across components and props
Better IDE autocomplete and refactoring
Reduced runtime errors
Improved code documentation through types

Build Tools: Speed Matters

Vite has emerged as the build tool of choice (75 mentions), displacing Webpack (20 mentions) through dramatically faster development server startup and hot module replacement. When developers talk about loving Vite, speed is almost always the first thing mentioned.

Testing: Playwright's Ascent

For end-to-end testing, Playwright (36 mentions) has overtaken older tools like Cypress (22 mentions), while Vitest (20 mentions) is rapidly becoming the preferred unit testing framework for Vite projects.

Project Management: The JIRA Problem

JIRA's market position perfectly illustrates the difference between "most used" and "most loved."

JIRA's Paradox

JIRA is simultaneously the most-used tool overall (mentioned by more developers than even VS Code or AWS) and the most-disliked tool by a massive margin. It has more negative mentions than the next four most-disliked tools combined.

The complaints are consistent:

Painfully slow performance
Overwhelming complexity
Cumbersome workflows
Poor user experience

Yet JIRA remains dominant across all company sizes, from tiny startups to enterprises. Why? Because it's often engineering leaders, product managers, or executives—not developers—who select project management tools. And JIRA's integration with the broader Atlassian suite (Confluence, Bitbucket) creates lock-in.

Linear: The Developer-First Alternative

Linear represents the most significant challenger to JIRA's dominance. At tiny companies (50 or fewer employees), Linear is almost as popular as JIRA. The tool is specifically designed for software teams and emphasizes speed and simplicity.

Survey respondents frequently mentioned Linear as what they wish they could use instead of JIRA:

"Jira is just a hostile tool towards actually delivering stuff... I'd prefer to use Linear."

"I dislike how JIRA tool is incredibly non-performant given what it does... if my company would allow it, I'd love to try Linear."

Interestingly, Linear is valued at only $1.25B despite having more developer users than publicly-traded competitors Asana (market cap $3.25B) and Monday.com ($9B), suggesting Linear either has room to grow or is more popular with developers specifically rather than broader business users.

The Azure DevOps Surprise

Azure DevOps shows surprisingly strong adoption (nearly 3x GitHub Issues), particularly at larger companies. As an integrated solution covering project management, CI/CD, and testing, it appeals to enterprise IT organizations already invested in Microsoft's ecosystem.

Infrastructure and DevOps: Kubernetes Everywhere

The infrastructure category reveals how containerization and cloud-native approaches have become standard practice.

The Container Trinity

Three tools dominate modern infrastructure:

Docker (containerization)
Kubernetes (container orchestration)
Terraform (infrastructure-as-code)

These tools appear in roughly 25-30% of responses each, indicating that about a quarter of developers work regularly with containerized, orchestrated, cloud-native infrastructure.

AWS Service Proliferation

The popularity of specific AWS services surprised even seasoned cloud architects. Hundreds of respondents mentioned ECS, EKS, EC2, and Fargate—suggesting that developers on AWS use multiple, specialized services rather than generic compute.

This pattern validates AWS's strategy of offering 240+ services: developers building scalable systems want purpose-built tools, not one-size-fits-all solutions.

CI/CD: GitHub Actions' Unexpected Dominance

GitHub Actions has become the most-used CI/CD tool by a significant margin, despite only launching in 2019. It's surpassed much older competitors like Jenkins, CircleCI, and Travis CI.

Why the rapid adoption? Integration. When you're already using GitHub for version control (which most developers are), using the same platform for CI/CD reduces context switching and simplifies authentication.

Cloud Providers: The Big Three (But Mostly AWS)

While AWS, Azure, and Google Cloud Platform collectively dominate, the split isn't even:

AWS: Used by the majority of respondents
Azure: Strong in enterprises and larger companies
GCP: Distant third but still significant

The survey data suggests AWS is over-represented compared to revenue market share (where Azure holds ~30%), likely because Pragmatic Engineer readers skew toward startups and Big Tech rather than traditional enterprises where Azure thrives.

The PaaS Landscape

Beyond the big three, Vercel clearly leads platform-as-a-service providers, with one-fifth the mentions that GCP receives. Vercel's optimization for Next.js (which they created) and excellent developer experience has carved out a solid niche.

Other notable mentions include Heroku (showing surprising staying power despite age), Render and Railway (modern PaaS startups), and Hetzner (the bootstrapped German IaaS provider known for affordable pricing).

What This Means for Technical Hiring and Team Building

These trends have profound implications for how you should think about hiring and team composition:

1. AI Fluency Is No Longer Optional

With 85% of developers using AI tools, candidates without AI experience are increasingly behind the curve. As we detailed in our analysis of technical interview best practices, modern assessments should evaluate:

How effectively candidates use AI assistants
Their ability to review and improve AI-generated code
Prompt engineering skills for complex problems
Critical thinking to identify when AI leads them astray

At CoderScreen, we're helping companies design assessments that reflect real-world AI-augmented development, not just theoretical algorithm knowledge.

2. Junior Developers Are Your AI Natives

As explored in depth in our piece on AI engineer hiring trends, new grads are often more fluent with AI tools than senior developers. Companies like Databricks are hiring 3x as many recent college graduates specifically because of their AI-native capabilities.

If you've been avoiding junior hires due to concerns about AI replacing their roles, you have it backward: junior developers who grew up with ChatGPT and Copilot can transform your organization's AI adoption.

3. Your Tool Stack Signals Company Culture

Tool choices send powerful signals to potential hires:

Cursor vs GitHub Copilot: Startups favoring cutting-edge tools vs enterprises with standardized contracts
Linear vs JIRA: Developer experience prioritization vs enterprise process requirements
PostgreSQL vs specialized databases: Pragmatic simplicity vs technical sophistication

Talented developers often evaluate companies based on their tech stack. If you're using slow, outdated tools that developers openly dislike, expect hiring challenges.

4. The PostgreSQL Default

Unless you have specific requirements preventing it, starting with PostgreSQL is the pragmatic choice for most applications. Its extensibility means you can add capabilities (full-text search, vector support, time-series) without introducing additional databases.

This matters for hiring because developers increasingly expect to work with modern, well-regarded tools. "We use Oracle and SQL Server" is a harder sell than "We use PostgreSQL with modern extensions."

5. React + TypeScript + Next.js Is Your Safe Bet

For frontend development, the combination of React, TypeScript, and Next.js has become so standard that it's almost boring. But boring can be good: it means:

Large talent pool familiar with the stack
Abundant learning resources and community support
Proven scalability and performance
Strong TypeScript support across the ecosystem

Choosing something more exotic (Svelte, Solid, Qwik) might be technically interesting but significantly shrinks your hiring pool.

Key Takeaways: What Actually Matters

After analyzing responses from 3,000 developers, several patterns emerge:

1. AI Tools Are Here to Stay

GitHub Copilot reaching 20 million users and Cursor generating $500M ARR from near-zero in 18 months proves that AI coding assistants aren't a fad. They're fundamentally changing how software gets built.

Action item: If you're not evaluating AI tool usage in technical interviews, you're assessing for outdated skills.

2. TypeScript Won the JavaScript Wars

With 70%+ of new React projects using TypeScript and 60 million+ weekly downloads, the debate is over. TypeScript's type safety and developer experience improvements make it the clear choice for serious applications.

Action item: If your team is still writing vanilla JavaScript for new projects, you're falling behind industry standards.

3. PostgreSQL Is the Database Default

One in three developers uses PostgreSQL, and no other database comes close. Its combination of reliability, extensibility, and zero licensing costs makes it the pragmatic choice.

Action item: Before selecting a specialized database, ask "Why not PostgreSQL?" rather than "Which database should we use?"

4. Developer Experience Drives Tool Adoption

Cursor's meteoric rise and Linear's growth both demonstrate that developers gravitate toward tools that are fast, intuitive, and pleasant to use—even when more established alternatives exist.

Action item: Evaluate your tool stack from a developer experience perspective. Slow, clunky tools drive away talent.

5. Company Size Determines Tool Choices

Large companies favor enterprise vendors with contracts (GitHub Copilot, JIRA, Azure DevOps), while smaller companies embrace newer, more flexible alternatives (Cursor, Linear, various AI tools).

Action item: Understand how your company size influences which tools make sense. Fighting against natural patterns usually fails.

The Future: Pragmatism Over Novelty

Perhaps the most important insight from this survey is the triumph of pragmatic, battle-tested tools over novel alternatives.

PostgreSQL (38 years old) dominates databases. React (12 years old) dominates frontend. TypeScript (13 years old) dominates typed JavaScript. These aren't the newest, shiniest options—they're the ones that work.

The exceptions—Cursor, Claude, Linear—succeed not by being novel for novelty's sake, but by delivering dramatically better experiences in specific areas: AI-native IDE design, coding-focused LLM capabilities, developer-first project management.

For teams building products and hiring developers, the lesson is clear: optimize for proven, well-supported tools that developers actually enjoy using. The hot new framework or database might make for interesting conference talks, but your team will be more productive and your hiring will be easier with boring, reliable choices.

The modern tech stack in 2025 isn't about having the most cutting-edge technologies. It's about having the right balance of stability, productivity, and developer happiness. And increasingly, with AI tools integrated throughout the development lifecycle, it's about giving developers the tools that let them focus on solving problems rather than fighting their infrastructure.

Want to ensure your technical hiring process reflects modern development practices? Get started with CoderScreen to create assessments that evaluate real-world skills with modern tools, AI assistance, and practical problem-solving—the capabilities that actually matter in 2025.