mVerve Let's talk Let's talk
The mVerve Tech Radar - Q1 2026

What's worth using. What's worth watching. What's not.

The Tech Radar is mVerve's view on the tools, methods, and choices shaping the work we do. Updated every quarter. Written by the engineers actually building with them.

Adopt

We have used these in real work. Default choice.

Trial

Worth a real try on a contained project.

Assess

Watching closely. Try it on a small piece. Do not build the foundation on it yet.

Hold

Avoid for new work. Move off where reasonable.

The radar

Twenty-six entries. One picture.

ADOPT TRIAL ASSESS HOLD LANGUAGES & FRAMEWORKS PLATFORMS TOOLS AI & DATA React TypeScript Python 3.12 Go Tailwind v4 Rust jQuery AWS Kubernetes Snowflake CF Workers Plex (Rockwell) Monolithic ERP Cursor Lenis Datadog Linear Storybook Jira (legacy) Claude Sonnet Llama 3/4 MCP LangChain Agentic WF Edge AI Digital Twins
Fig. 03 · mVerve Tech Radar · Q1 2026

Twenty-six entries across four quadrants.

Languages & Frameworks
React
Adopt - Default UI framework for client work.
TypeScript
Adopt - Mandatory for any non-trivial codebase.
Python 3.12
Adopt - Standard for data, ML, automation.
Go
Adopt - Where throughput matters more than ergonomics.
Tailwind CSS v4
Adopt - Lower CSS surface area, faster iteration.
Rust
Trial - Wasm modules and edge runtimes.
jQuery
Hold - Legacy only. Replace on next refactor.
Platforms
AWS
Adopt - Default cloud for most engagements.
Kubernetes
Adopt - Reach for it only when it's truly justified.
Snowflake
Adopt - Warehouse-of-record for analytics.
Cloudflare Workers
Trial - Edge logic + low-latency APIs.
Plex (Rockwell)
Trial - Wrap it with clean APIs. Do not extend the old core.
Monolithic ERP
Hold - Replace these piece by piece. Do not extend them.
Tools
Cursor
Adopt - Default editor for AI-assisted engineering.
Lenis
Adopt - Smooth scroll on every editorial site we ship.
Datadog
Adopt - Observability across services and infra.
Linear
Adopt - Project tracking that doesn't slow teams down.
Storybook
Trial - Component-driven design systems.
Jira (legacy)
Hold - Migrate teams to Linear when you can.
AI & Data
Claude Sonnet 4.6
Adopt - Default model for production agents.
Llama 3 / 4
Trial - Self-hosted reasoning where data must stay private.
MCP (Model Context Protocol)
Trial - Tool integration standard worth committing to.
LangChain
Trial - Good for prototyping. Rewrite when you go live.
Agentic Workflows
Trial - Going live for operations automation.
Edge AI
Assess - Promising for factory vision and inspection.
Digital Twins
Assess - Use carefully. Most pilots stay pilots.
From the editors

Three shifts driving this edition.

01
AI agents are starting to work in real systems.

We are seeing real traction with MCP and Claude Sonnet for operations automation. The question is no longer "can it work?". It is "can it stay reliable for six months without someone watching it?". Trial for now. Not Adopt yet. The gap is closing fast.

02
GreenOps moves from optional to required.

Carbon-aware scheduling and emissions accounting are showing up in big tenders. Teams that can put a real carbon number on their cloud systems will win bids the others cannot even enter.

03
Edge AI is closer than the debate suggests.

For factory vision and predictive maintenance, running AI on basic hardware near the equipment is now cheap enough and accurate enough. Still Assess for most teams. The tooling is rough. But worth a serious try if vision is on your roadmap.

Let's talk

Ready to engineer your industrial future?

Tell us what you're building and we'll tell you, honestly, how we'd approach it. We read every message and reply within one business day.