Simplyblock Provides Postgres Git-Style Branching

Why It Matters

Vela accelerates AI and ML development cycles while slashing storage costs, giving enterprises full control over their PostgreSQL environments. It addresses the scalability and latency limits of managed database services, a critical need as AI workloads grow.

Key Takeaways

• Vela creates instant Postgres branches via copy‑on‑write
• Eliminates snapshot‑restore delays and storage cost overhead
• Runs on local NVMe with RDMA for AI workloads
• Cloud‑agnostic, Kubernetes‑native, self‑hostable architecture
• Provides Git‑style version control for database schemas and data

Pulse Analysis

PostgreSQL has become the de‑facto data layer for modern AI applications, prompting cloud providers and startups alike to build serverless, AI‑ready variants. Companies such as Databricks, Neon, and Snowflake have invested heavily in PostgreSQL extensions to meet the demand for low‑latency, high‑throughput data access. In this context, Simplyblock’s Vela arrives as a timely innovation, offering developers the ability to spin up fully functional database branches in seconds, a capability traditionally reserved for heavyweight, on‑premise environments.

At the heart of Vela is a copy‑on‑write storage engine that lives on NVMe‑backed block devices, eliminating the need for costly full‑copy clones. Integrated with Kubernetes, Vela automates provisioning, autoscaling, and snapshot management through a single API, while supporting RDMA over Converged Ethernet (RoCE) for ultra‑fast data paths. This architecture directly tackles the IOPS throttling and latency spikes that plague managed PostgreSQL services during large‑scale model rollouts, delivering predictable performance and granular cost control.

For enterprises, Vela’s Git‑style branching introduces familiar version‑control semantics to database development, enabling diffing, merging, and rebasing of schema and data changes. The hosted sandbox lowers the barrier to entry, allowing teams to experiment without infrastructure commitments before moving to self‑hosted, cloud‑agnostic deployments. As AI workloads continue to scale, tools that combine developer agility with storage efficiency—like Vela—are poised to become essential components of the AI data stack.

Simplyblock provides Postgres Git-style branching

By Simplyblock · Published February 11 2026

Fast cloud‑native block storage provider Simplyblock has built its Vela software to provide instant Postgres database branches with no data copying and running on its block storage.

Having created its Kubernetes‑integrated block storage, the company has moved up stack to produce system application software that uses it. Simplyblock CEO and co‑founder Rob Pankow thinks Postgres is becoming the AI database. For example, Databricks is developing Lakebase, its serverless Postgres database built for AI agents. It acquired Neon for $1 billion in May 2025 to get serverless Postgres tech. Snowflake acquired Crunchy to get its Postgres software; “the AI‑ready, enterprise‑grade and developer‑friendly PostgreSQL database to the AI Data Cloud.” Pankow reckons Postgres has become the anchor of modern AI development and wants to ride that train.

Rob Pankow.

He said:

“Vela brings the speed of serverless, the structure of version control, and the resilience of modern storage into one unified database platform. By combining orchestration, virtualization, and high‑performance storage under one API‑driven architecture, we’re giving teams the tools to run Postgres like a modern platform. Not just a database.”

Pankow says the next question is where and how to run it.

Managed Postgres instances have problems. Pankow said:

“I speak with engineering teams every week. They all describe similar experiences. They try to scale a managed Postgres instance during a model rollout. They hit IOPS limits. They hit throttling windows. They see latency spikes at the exact moment they need predictability. They also see cost blowups because the only way to remain safe is to overprovision every environment. These problems accumulate slowly at first. Then they become unmanageable once AI workloads reach production scale.”

A managed database means a one‑size‑fits‑all solution. AI workloads demand extremely fast storage and predictable performance. They also require large and frequent database clones for testing and experimentation. Managed databases struggle in both areas. The internal storage layers of managed systems create unavoidable bottlenecks. The cloning mechanisms depend on snapshot‑restore cycles or full‑blown physical copies. Both approaches are slow and expensive, especially at scale.

Developers should run Postgres in their own cloud accounts, co‑locating data and GPUs. Vela lets you deploy Postgres on the same instance as your storage, using the speed and performance of local NVMe devices attached to the instance. It provides resilience and scalability, as well as copy‑on‑write functionality, typically not available with local storage, deployed and managed in the user’s own cloud instance with local NVMe devices.

A database branch is a clone minus the inherent database copying disadvantages, using copy‑on‑write instead. As we understand it, a branch builds on the clone concept but adds Git‑like semantics and lifecycle. It’s a named, versioned line of development derived from a parent, usually “main” or another branch, and branches are organized in a tree or hierarchy structure. You can diff changes between branches, merge or promote changes back to the parent, rebase, or resolve conflicts.

Vela is built from the ground up to be self‑hostable and cloud‑agnostic, with deep integration into Kubernetes. It introduces Git‑style branching and instant environment creation. Vela manages autoscaling, snapshots, and lifecycle orchestration natively, and includes integrated authentication, row‑level security, a self‑service UI, and observability in the control plane. For high‑performance it supports RDMA networking via RoCE alongside NVMe/TCP.

Simplyblock has launched a hosted Vela Sandbox to enable developers to experience the branching model and spin up full environments in minutes with no setup. Once they are ready to scale, Vela runs in their own infrastructure—on Kubernetes, bare metal, or virtual machines—preserving data sovereignty, security, and operational control.

A Vela Beta is available with documentation, guides, and the hosted sandbox here.