The State of Engineering
& Product in 2026
Survey results from FirstMark’s CTO & CPO Guilds, showing what top technology companies are building, buying, and betting on in 2026
About this survey
The FirstMark CTO & CPO Survey captures what today's technical executives are thinking and how they're operating across people, processes, metrics, tools, and platform shifts.
This edition was open exclusively to members of FirstMark's Product & CTO Guilds — an invite-only community of technical leaders from the FirstMark portfolio and the broader unicorn ecosystem.
We hope you find the results useful, and welcome your input: email us at community@firstmark.com with questions or data points you'd like included in future surveys.
Want to participate and get early access?
CTO
SURVEY
SURVEY
20
26
CTOs from the world's leading companies participated, including:
Table of Contents
LLM Infrastructure for 2026Copilots and Agentic workflowsAI Productivity: Operational Infrastructure or Hype?AI Launches Saturate the MarketThe Surge of Specialized ToolsMulti‑Cloud as a Sales MandateObservability UnbundledHybrid Workplace Sentiment & Impact on Speed and ExecutionThe insights that follow are not a summary of every data point in this report. They reflect the most meaningful shifts we see across product, engineering, and technology leadership today. By comparing this year’s responses with last year’s, clear patterns emerge in how teams are building products, buying tools, deploying AI, and structuring their organizations.
Rather than cataloging tools or preferences, each insight pulls together multiple signals into a single takeaway. The charts that accompany them are intentionally focused, highlighting what has changed and why it matters. Taken together, these insights show how leading teams are adapting to growing system complexity, faster AI adoption, and changing expectations around productivity, tooling, and work models.
LLM infrastructure is going multi-vendor.
As the LLM API market matures rapidly, we've seen multi-vendor adoption become the norm, with 58%+ of respondents using 2+ model APIs across their org. There has also been major movement in vendor adoption: Gemini rose to #2, Anthropic fell to #3, and Bedrock jumped from #10 to #4. The signal is clear: procurement, governance and cloud alginment now drive decisions, not just model capability. CTOs are moving from single-vendor bets to portfolio strategies, optimizing for cost, latency and reliable fallbacks. To stay flexible as rankings keep changing, build an abstraction layer to avoid lock-in and negotiate contracts with churn in mind.
Key Takeaway: The "best model" changes quarterly. Optimize for the platform and design for portability, fallbacks, and governance.
2024
2025
OpenAI
Google Gemini
Anthropic
AWS Bedrock
Azure
Other
None
Groq
Copilot adoption flipped: Cursor and Claude Code now lead.
Developer AI tooling is changing faster than most organizations can govern. Cursor rose from #4 to #1, Claude Code climbed from #9 to #2, and GitHub Copilot fell from #1 to #3. The takeaway isn’t which tool wins—it’s that developer workflows are shifting toward more integrated, agentic environments. Treat copilots as a workflow layer; evaluate systemically, standardize responsibly, and govern securely. Run a quarterly copilot review across key teams, set an approved tool policy (security, IP, governance), and track impact via cycle time, PR throughput, and defect rate.
Key Takeaway: Copilots are a fast-moving workflow layer. Govern adoption and prove ROI with dev metrics.
2024
2025
Cursor
Claude Code
GitHub Copilot
OpenAI Codex
Other
None
AI productivity gains are real. Expectations are catching up.
AI-driven productively gains are accelerating: leaders reporting meaningful engineering improvements jumped from 64% to 73% and fewer believe AI's impact has been overstated, a sign that teams calibrate based on actual usage, not hype. AI is now operational infrastructure. The shift is from experimentation to intentional leverage: redesign workflows with guardrails and clear measurement, instrument productivity (cycle time, review time, incident load), and reorient org design toward higher level roles (e.g. IC/staff+, platform, and ML).
Key Takeaway: Move from AI experiments to measured, guardrailed workflow redesign.
2024
2025
Has AI boosted your engineering productivity?
No
Unsure
Yes
The Impact of AI Has Been...
Underestimated
About Right
Overestimated
Shipping AI is now table stakes as nearly every team is building for it.
AI shipping has hit saturation: teams launching AI products jumped from 64% to 88%, and the share planning to launch within 12–18 months rose from 75% to 93%. Presence is no longer a differentiator; depth is. The edge now comes from end-to-end automation, data advantage and measurable business outcomes. Focus on workflow automation over standalone “AI features,” build an internal evaluation and trust layer (quality, safety, governance), and tie AI to ROI metrics like margin, retention, and cycle time.
Key Takeaway: AI is baseline, and companies must now differentiate with automation depth and measurable ROI.
2024
2025
No plans
Planning to Launch AI
Product This Year
Product This Year
AI Product Launched
The tool stack is expanding and teams are in Buy Mode.
After years of consolidation pressure, teams are now planning to add more tools. Specialized tooling is increasingly a lever for velocity, automation, and resilience. The new bottleneck isn't tool cost, it's integration. Teams are trading sprawl for speed, which raises the bar on orchestration across workflows, identity, and logging. To keep the stack productive, build a governance loop (evaluation → rollout → review → deprecate) and treat workflow cohesion as a core productivity initiative.
Key Takeaway: Tool sprawl is the trade for speed. Win by orchestrating and governing the stack.
2024
2025
Less
More
The Same
Multi-cloud is a sales requirement, not a redundancy strategy.
Multi-cloud adoption isn’t rising because teams fear outages. It’s rising because customers expect it. Cloud portability is now a go-to-market feature shaped by procurement, security, and buyer preference. AWS remains the leader, but multi-cloud is standard because customers ask for it, not because teams are optimizing for redundancy. Treat cloud portability as a product feature: focus on the segments that care most, and build commercial messaging around “deploy in your cloud.”
Key Takeaway: Multi-cloud wins deals. Build and sell “deploy in your cloud.”
Amazon Web Services (AWS)
Google Cloud Platform (GCP)
Microsoft Azure
Oracle
Cloudflare
Lambda
Other
Observability is fragmenting and most teams now rely on multiple tools.
Datadog remains the leader, but teams are layering tools like Sentry, AWS CloudWatch, Grafana, and OpenTelemetry to cover gaps. This reflects rising system complexity and a shift from “choose a vendor” to “design an observability architecture.” Winning teams aren’t finding the perfect tool; they’re standardizing telemetry and workflows across many. To reduce fragmentation, standardize where possible, optimize data retention and routing to control costs, and build a unified incident workflow that spans all observability tools.
Key Takeaway: Standardize telemetry and incident workflows across a multi-tool observability stack.
Datadog
Sentry
AWS CloudWatch
Grafana
Open Telemetry
Prometheus
Elasticsearch
New Relic
Other
Hybrid is the default and confidence in remote productivity is collapsing.
Sentiment shifted sharply: belief that engineers are more productive remotely fell from 55% to 11%. But the result isn't a full return to office; hybrid remains dominant, with most orgs coming in 3 days a week. Leaders increasingly see in-person time as necessary for cohesion, speed, and high-quality execution. Make hybrid the priority is making it intentional: define what in office time is for, optimize remote time for deep work and async execution, and set predictable collaboration rhythms for planning, reviews, and onboarding.
Key Takeaway: Hybrid is here to stay (for now). Make in-office time purposeful and remote time execution-focused.
In-Office
Distributed
Hybrid
Hybrid Breakdown:
Days in-office per week
1
2
3
4
Enter your email for instant access to the full report.
View interactive charts, filter data by organization size, financing round, and revenue scale, and more...
Thank you! Please check your email for access to the full report.
Oops! Something went wrong while submitting the form.