AI Headlines LogoAI Headlines

Funding | 8/23/2025

Zed Bets $32M on DeltaDB to Disrupt Git

Zed, an AI coding startup, has raised $32 million in a Series B led by Sequoia Capital to advance DeltaDB, a novel operation-based version control system. DeltaDB aims to fuse real-time code edits with the surrounding conversations between humans and AI agents, challenging Git's snapshot-based model and signaling a broader shift toward collaborative, AI-powered development tools.

Zed bets on a new kind of code history

If you’ve ever stared at a Git history and thought, there has to be a better way to explain why a change happened beyond just a commit message, you’re not alone. Zed, an AI-focused coding startup, has just closed a $32 million Series B led by Sequoia Capital to turn a long-standing belief in software development on its head. The money isn’t just a trophy; it’s a nudge toward DeltaDB, a version-control system designed to live where conversations live—alongside the code. In plain terms, DeltaDB aims to track every edit with a level of detail that makes it possible to retrace not just what changed, but why, and in what context that change was discussed with human teammates and AI assistants.

A new rival to Git? Here’s the idea

  • Git has dominated the software world for decades by snapshotting the codebase at discrete moments. DeltaDB flips that model on its head by focusing on operations—every keystroke, every edit, and every chat message tied directly to the code it touched.
  • DeltaDB’s core technology uses Conflict-free Replicated Data Types (CRDTs). That means multiple people—humans, AI agents, or both—can edit the same file at the same time and still stay in sync without clunky conflict resolution rituals.
  • The ambition goes beyond faster merges. Zed wants to embed the dialogue around coding decisions into the version history so future edits and AI suggestions have a built-in map of assumptions, constraints, and rationale.

How it fits with Zed’s existing toolkit

Zed’s real-time editor already supports co-editing, shared cursors, and integrated voice chat. DeltaDB would move the editor’s “multiplayer” vibe into the version-control layer, ensuring that the living conversation—whether it happened in a PR thread or a voice note—becomes part of the code’s living memory. Think of it as a collaborative workspace where your IDE, your teammates, and your AI helper all contribute to a single, navigable history of a project.

How CRDTs enable real-time collaboration

CRDTs are the quiet workhorses that let a roomful of contributors edit a document without tripping over each other. In a distributed, AI-enhanced workflow, DeltaDB’s CRDT foundation means a human typing a change and an AI agent proposing edits won’t produce an ugly merge conflict; instead, both streams converge into a single, consistent history. That’s the “multiplayer” fantasy realized—and it’s exactly what Zed wants to demonstrate with DeltaDB.

Why now? Market context and potential impact

The software development space is increasingly crowded with AI coding assistants and smart IDEs. Analysts have predicted a multi-billion-dollar market by the end of the decade, driven by tools that promise to accelerate coding, reduce context-switching, and lower onboarding costs for new contributors. GitHub Copilot and similar products have already changed how teams write code, and DeltaDB could crystallize that change by tying conversations and decisions to the code from day one.

Sequoia partner Sonya Huang framed the investment as betting on a collaborative future, noting that DeltaDB represents a fundamental shift in how developers will work with AI agents. If DeltaDB succeeds, future code reviews might begin with an AI summarizing the conversation that led to a change and offering context for why a particular path was chosen. If it struggles, it could slow down adoption by complicating workflows or by raising expectations that aren’t yet met by current AI tooling.

The challenges ahead

  • Git’s ecosystem is massive. Breaking in means more than a fancy technical idea; it means interoperability with countless tools, services, and workflows that assume a snapshot-based history.
  • User experience will matter just as much as architecture. Integrating deep, contextual conversations into a version history requires an intuitive UX that helps developers discover the reasoning behind edits without overwhelming them.
  • Real-world AI behavior isn’t always spotless. While the DeltaDB vision is compelling, the tools must prove they can speed up work rather than slow it down with misinterpretations or noisy histories.

What this could mean for developers

If DeltaDB becomes a practical reality, the developer’s workspace could resemble a shared design board along with a code base:

  • Real-time collaboration that doesn’t force a trade-off between speed and context.
  • A code history that tells a story—why a change happened, who participated, and what constraints guided the decision.
  • A more intelligent codebase, where AI agents consult a persistent, contextual history to improve suggestions and edits over time.

But this is still early days. DeltaDB’s success hinges on delivering a tangible UX win, proving that embedded conversations actually enhance code quality, and showing developers a clear path to integrate with the rich Git-compatible ecosystems they already rely on.

Looking ahead

This funding round isn’t just about a new product; it’s a declaration that the future of software development could be defined by collaboration in real time, with code evolving in a fluid stream of operations and conversation. Zed isn’t promising a revolution tomorrow; it’s betting that the groundwork—CRDT-based collaboration, operation-level versioning, and an integrated conversational memory—will become the norms that developers reach for when building the next generation of software.

What to watch

  • Adoption: Will teams pick DeltaDB as a drop-in for Git, or will it be adopted piecemeal for new projects?
  • UX breakthroughs: Can DeltaDB deliver a clean, intuitive experience that makes it easier to understand historical decisions?
  • AI reliability: Will real-time AI edits be accurate, helpful, and safe for production code?

Bottom line: Zed’s $32 million round signals strong faith that the future of coding will be more collaborative, contextual, and AI-assisted. DeltaDB could reshape not just how we version software, but how we think about the reasoning behind every change.

Sources

  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHWTKU8-5dDitDSMar-IIVoRzJ4nwpeRYEGhtSI6dCv5aPKduj3RmAFBDrUVKkByTTQsasH24L4nr6RLmRA1pN4wF2e38II7PFSqyRus3ZhLayk9rm0J0WqoeZcg0h9brwK3pv3kjy0cv_E8P2aRovo6kD8Yp7B15OG_MF7SoDXYYrDeA==
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHzfkrEz30RLuUs6SNZa7H9jia7VIScXPIZs-c0SQuJPAkcF5q71XJmnNhk4Ar-hljzgNu3RQS8h3SyH6ehFBZaHeE4sjDyX7PFSqyRus3ZhLayk9rm0J0WqoeZcg0h9brwK3pv3kjy0cv_E8P2aRovo6kD8Yp7B15OG_MF7SoDXYYrDeA==
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQG3XNtyhVJhxz4rMeTPKfJaPGM7EwCl1X_pWW5_wC2NWH23APYuhIMaUCLozNyDjX4FbIDSgkOV2OI76zd-2hDgl2l4Q4Q9U3Z8f1nZLIdPEhDKL4wc_g-J6eoanecBCKnPrg080JMc6kNBA2I3F0JTPxnxqnnsa_XDe51Xh5xZTM5MyyFpDKE_PXc=
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQHuYRVMSQG6pZfzIMFvjfifWfkYBb2D3SF59s-piICf08CFOhvRolhiPxKH3hUyfMpXdl8qyRJoJLa8TkBJ3jW86AHLZ9b--AjUJr7wfavOokfKRrl9KzKiUMtv9Wqfe7rU29jrckGEXfeoxDdq5sqMataGAWvkhg1eQH9DAyCqy_jXHu49COCHXvj_IfZx0mrbHUmZsxAu120MD6o8Y0oemA==
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE1hieemkZ7vVIwaGqgCWaU2oJqngMxJ5l4DaxpLLWs9R9z1Q1F_8FSrd8ItTslQcMXggAiLfEA1pEyceTJLQpBlgax4hz4Aht9LX5lFBWWNBhWMVb7Fm-oqVRjcRRPB8jF-6rSX-1ae5r3U8jrNO1Tz76YAC5zTjrZz5baMKML8y4pqcYPP8IzF6dynzL_2EkF
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGro1kIvjMML5R1ZucaA8KzK_SzOfCFestpWsbadc3fEOO90nErPd8PvhJnVoNFCqlwpk9vnYgKQY-JkKCr0BCAvz9kY9BFnUpbXHMAmCpmXnPmoZT0uewfy5LqpCl6tVZRr0aJ4-2P6MT2td11ap7P3LyfBBZDi7cqzkzDvW3HLqUhGL_NF71ezz4iMLyR_vokR67nELUmv2-Nw6aBdrSHRVIU7LGTOA0e1rg0XnYID5mp
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF9xpnle4wzXu9TKpDE7ipaS3IOWksq9k4Wgq_wzBnA3oH5okDcEizRDRZ9AfOHIlyHMg0NWcpVCSn5qL-LVqUV5jiC3mKver8-Oozlxn5sf83iABxeRlvzxPJRxOrmJkqmZz_f-DACBJBKC_-rj1gTgJIbNOB83J0C_h_czDihzz7Oopg=
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEMnvLHQdVMk8wXDCtPBxZbAwxudWvUORBRFuawYb-wpGTQfTSVSB1GSk9vLDP-HjT_3PFoxsamxe8TYGyr2qUN-OYo57YlUJMMod1XGCdMx2I37jPlrZzR6OtbwAG3wzqcnNS2JlhIerQmBaDvo1MAsMRCQNn2r6P9TlstIrDCQHCJ
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQGro1kIvjMML5R1ZucaA8KzK_SzOfCFestpWsbadc3fEOO90nErPd8PvhJnVoNFCqlwpk9vnYgKQY-JkKCr0BCAvz9kY9BFnUpbXHMAmCpmXnPmoZT0uewfy5LqpCl6tVZRr0aJ4-2P6MT2td11ap7P3LyfBBZDi7cqzkzDvW3HLqUhGL_NF71ezz4iMLyR_vokR67nELUmv2-Nw6aBdrSHRVIU7LGTOA0e1rg0XnYID5mp
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQF9xpnle4wzXu9TKpDE7ipaS3IOWksq9k4Wgq_wzBnA3oH5okDcEizRDRZ9AfOHIlyHMg0NWcpVCSn5qL-LVqUV5jiC3mKver8-Oozlxn5sf83iABxeRlvzxPJRxOrmJkqmZz_f-DACBJBKC_-rj1gTgJIbNOB83J0C_h_czDihzz7Oopg=
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQE1hieemkZ7vVIwaGqgCWaU2oJqngMxJ5l4DaxpLLWs9R9z1Q1F_8FSrd8ItTslQcMXggAiLfEA1pEyceTJLQpBlgax4hz4Aht9LX5lFBWWNBhWMVb7Fm-oqVRjcRRPB8jF-6rSX-1ae5r3U8jrNO1Tz76YAC5zTjrZz5baMKML8y4pqcYPP8IzF6dynzL_2EkF
  • https://vertexaisearch.cloud.google.com/grounding-api-redirect/AUZIYQEGro1kIvjMML5R1ZucaA8KzK_SzOfCFestpWsbadc3fEOO90nErPd8PvhJnVoNFCqlwpk9vnYgKQY-JkKCr0BCAvz9kY9BFnUpbXHMAmCpmXnPmoZT0uewfy5LqpCl6tVZRr0aJ4-2P6MT2td11ap7P3LyfBBZDi7cqzkzDvW3HLqUhGL_NF71ezz4iMLyR_vokR67nELUmv2-Nw6aBdrSHRVIU7LGTOA0e1rg0XnYID5mp
Back to blog