Skip to main content
NeuronFeed
CATEGORY

AI Coding startups (2026)

From Cursor's $50B mark to Cognition's $25B Devin talks — the category that pulled SaaS multiples sideways

65 ai coding startups tracked, with the largest concentration in US. Total tracked funding: $17.1B.

Tracked
65
Total Raised
$17.1B
Countries
7
Active Deals
6

Editor's picks

6

Top by score

View all 65 →

Funding by year — AI Coding

2021 → 2026
$20M
’21
$192M
’22
$426.8M
’23
$2.3B
’24
$13.3B
’25
$1.9B
’26

Market overview

When Cognition closed its acquisition of Windsurf in July 2025, the move shifted the center of gravity in AI coding from autocomplete to autonomy. Devin gained an editor surface; Windsurf gained an agent that finishes pull requests. By September, Cognition was repriced to $10.2B; by April 2026, the company was reportedly in talks at $25B. Cursor answered with its own escalation — talks at $50B pre-money in April 2026, against a $2B ARR run-rate. The 40 startups in this slice account for $14.5B in cumulative funding, and Cursor's $7.57B Series D alone is more than half of it.

Where the benchmarks landed

Claude Opus 4.5 and Opus 4.7 sit on top of the Aider Polyglot leaderboard at 89.4% — the canonical multi-language coding eval that runs 225 Exercism problems across C++, Go, Java, JavaScript, Python, and Rust. SWE-bench Verified shows similar standings: Claude Opus 4.7 (Adaptive) at 87.6%, GPT-5.3 Codex at 85%. On the open-weight side, GLM-5 hit 77.8% on SWE-bench Verified — close enough that BYO-model agents like Cline and OpenCode are now meaningfully competitive.

Who is shipping and at what price

Cursor, Cognition (Devin), Replit (Agent, $9B Series F), Poolside ($500M Series B at $3B), Augment Code ($252M Series B), and Sourcegraph (Cody, $228M Series D) are the funded independents. Tabnine ($55M Series B at $1.5B) holds the enterprise-private niche. The free-or-bundled tier is Microsoft GitHub Copilot, Google Gemini Code Assist, Amazon Q Developer, Amazon Kiro, and JetBrains AI Assistant — distribution-first and roughly zero-cost if you already pay the parent cloud. Cline (open-source, $4M seed at $110M) and OpenCode are the BYO-model wedge.

What changes the picture from here

Frontier coding models commoditize fast — every six weeks brings a new Aider Polyglot leader, and per-token prices keep falling. The independent thesis depends on the editor surface compounding around the model rather than under it. Cursor's reported $2B ARR is the proof point. Devin's autonomous-agent pricing — billed against task throughput, not seats — is the alternative. The question for the next 12 months is whether GitHub Copilot's free tier with frontier-grade quality compresses everything in the middle.

Key trends 2026

  • Aider Polyglot is the eval that matters. Claude Opus 4.5 leads at 89.4%, GPT-5 (high) at 88.0%, Gemini 2.5 Pro at 82.2%. The 225-problem multi-language benchmark replaced HumanEval as the reference for who can ship working code, not just pass tests.
  • Open-weight coders closed the gap. GLM-5 hit 77.8% on SWE-bench Verified; DeepSeek V3.2-Speciale and Qwen3-Coder are competitive enough that Cline and OpenCode now route to them by default.
  • Capital is bimodal. Cursor's $7.57B Series D and Cognition's $400M at $10.2B (now $25B in talks) define the top end. Cline raised $4M seed at $110M, and OpenCode runs on zero. There is almost nothing in the middle.
  • Autonomous-agent pricing is real. Devin and Factory ($150M Series C at $1.5B) charge against task completion, not seats — a pricing inversion that matters once benchmark scores cross a threshold enterprises trust.

Benchmarks vs global

Tracked startups
40
densely funded category
Total funding
$14.5B
Cursor Series D = >50% of total
Aider Polyglot leader
89.4%
Claude Opus 4.5, May 2026
SWE-bench Verified leader
87.6%
Claude Opus 4.7 Adaptive, May 2026

Top countries

By startup count

Stage breakdown

Latest round type
  • Seed 25
  • Series A 10
  • Corporate 7
  • Series B 5
  • Series D 3
  • Pre-Seed 3
  • Series C 2
  • Growth 2

Top investors backing AI Coding

See all →

FAQ

Frequently asked

What's the Aider Polyglot leaderboard saying about coding models in 2026?
Claude Opus 4.5 leads at 89.4% (Anthropic-reported), GPT-5 (high) at 88.0%, Gemini 2.5 Pro Preview at 82.2%, and o3 at 81.3%. The benchmark runs 225 Exercism problems across six languages with a two-attempt format. Open-weight models like GLM-5 and DeepSeek V3.2-Speciale are within striking distance, which is why BYO-model agents like Cline have stayed competitive on a $4M seed.
What does Cognition's Windsurf acquisition mean for the category?
Closing in July 2025, the deal gave Cognition the Windsurf editor — roughly $82M ARR and 350+ enterprise customers — under Devin's autonomous-agent backend. Combined enterprise ARR rose 30%+ post-close. Cognition repriced to $10.2B in September 2025 and entered $25B talks in April 2026. Codeium's $2.97B in cumulative funding got partially absorbed in the Windsurf transaction.
Which AI coding tool has the highest revenue per user?
Cursor's reported $2B ARR run-rate is the public benchmark, implying $500-1000+ per active user. Cognition's Devin is priced higher per task. The hyperscaler assistants (GitHub Copilot, Gemini Code Assist, Amazon Q Developer) sit at $19-40/month list, which is structurally lower because they cross-subsidize from cloud margins.
Are open-source coding agents actually competitive?
Cline ($4M seed, $110M valuation) and OpenCode (zero raised) hit usable scores on SWE-bench Verified by routing to GLM-5, DeepSeek V3.2-Speciale, and Qwen3-Coder. The agent harness is open; the model is the user's choice. That decoupling is real pressure on Cursor's all-in-one thesis, though Cursor's editor still wins on UX integration.
What kills a coding-agent company in 2026?
Three risks. First, margin compression from per-token price drops eroding any markup. Second, distribution loss when Microsoft, Google, or AWS ships free equivalents — GitHub Copilot, Gemini Code Assist, Amazon Q. Third, benchmark stagnation: if a startup's agent does not beat the latest open-weight baseline on Aider Polyglot or SWE-bench Verified, enterprise buyers will not justify the line item.

Recent rounds in AI Coding

All rounds →
Date Startup Round Amount
May 2026 Cognition Series D $1B
May 2026 Blitzy Series B $200M
Apr 2026 Factory Series C $150M
Mar 2026 Qodo Series B $70M
Mar 2026 Replit Series F $400M
Feb 2026 Entire Seed $60M
Jan 2026 Syntropy Seed $500K
Dec 2025 LatentForce Seed $1.7M

All AI Coding startups

Page 2

LatentForce

IN est. 2024

Agentic AI for large-scale enterprise code migration and modernization

Raised
$1.7M
Stage
Seed
57

Morph

US est. 2025

Fast Apply infrastructure that merges AI code edits at 10,500 tokens/sec

Raised
$500K
Stage
Seed
57

Kodesage

GB est. 2024

On-premise AI that modernizes legacy enterprise software like COBOL and Oracle Forms

Raised
$6.6M
Stage
Seed
57

Poolside

Verified
US est. 2023

AI platform for software engineering

Raised
$500M
Stage
GROWTH
56

Manicule

US est. 2025

AI-native documentation studio for developer tool companies.

Raised
$500K
Stage
Seed
55

Refortifai

US est. 2026

Reasoning models and RL environments for electronic design automation and chip design workflows.

Raised
$500K
Stage
Seed
55

Replicas

US est. 2026

Background coding agents that delegate engineering work to sandboxed AI in Slack, Linear, and GitHub.

Raised
$500K
Stage
Seed
55

Cline

US est. 2024

Open-source autonomous AI coding agent for VS Code — bring your own model

Raised
$27M
Stage
S-A
53

GitHub Copilot

US est. 2021

The world's most adopted AI code assistant — 1M+ developers inside GitHub

Stage
CORPORATE
52

Gemini CLI

US est. 2025

Google's free open-source terminal AI coding agent — 1M tokens, no cost

Stage
CORPORATE
52

Runtime

US est. 2026

Sandboxed coding agents for everyone on your team — with guardrails and observability.

Raised
$500K
Stage
Seed
52

Superlog

US est. 2026

AI-native observability that installs itself and ships pull requests to fix the bugs it finds.

Raised
$500K
Stage
Seed
52

Limrun

US est. 2025

Cloud XCode, iOS and Android simulators that let any coding agent build mobile apps without a Mac.

Raised
$500K
Stage
Seed
52

Corelayer

US est. 2025

AI on-call engineer that debugs alerts, logs, and data pipelines in regulated industries.

Raised
$500K
Stage
Seed
52

Kilo Code

US est. 2024

Structured AI coding modes for VS Code — architect, code, and review separately

Raised
$8M
Stage
Seed
51

Tabnine

IL est. 2013

Private AI code assistant for enterprise developers.

Raised
$55M
Stage
S-B
50

JetBrains AI Assistant

CZ est. 2024

AI inside IntelliJ, PyCharm, WebStorm — code without leaving your JetBrains IDE

Stage
CORPORATE
50

Sweep

US est. 2023

The fastest AI coding assistant for JetBrains IDEs.

Raised
$5.5M
Stage
Pre-S
50

Indexable

US est. 2026

Sandbox infrastructure for AI agents with millisecond full-environment forks.

Raised
$500K
Stage
Seed
50

Deep Interactions

US est. 2025

A collaborative AI app builder that lets entire teams ship production AI products together in an afternoon.

Raised
$500K
Stage
Seed
50

Zed

US est. 2021

The multiplayer code editor — fast, native, and AI-first from the ground up

Raised
$16.6M
Stage
S-A
48

Phind

US est. 2022

AI search engine for developers

Raised
$10.4M
Stage
S-A
47

Bloop

GB est. 2021

Understand any codebase instantly — semantic code search powered by AI

Raised
$11M
Stage
S-A
47

Continue

US est. 2023

Quality control for your software factory with AI checks on every pull request.

Raised
$3M
Stage
Seed
47