Roboflow co-founders Joseph Nelson and Brad Dwyer

At GV, we have long been excited for AI to transform real-world industries and improve quality of life. Computer vision is a technology with that potential, capable of reinventing the way we understand, interact with and control the environment around us. Today, machine vision powers a slew of use cases, including vehicle autonomy, medical diagnostics, surgery robots, vertical farming, public safety, retail operations, worker safety and manufacturing quality.

Historically the tooling for creating computer vision applications and deploying them into the physical world has been nascent, complex and fragmented, requiring uncommon expertise across data preparation, model development, training infrastructure, deployment and hardware integration. This has been a barrier to adoption, restricting access to the few organizations with the technical talent or deep pockets to spend on outsourced development.

Multiple forces are converging to catalyze a Cambrian explosion moment for the computer vision market.

In our enduring pursuit of a company and team that could democratize access to this technology, we were fortunate to finally meet Joseph and Brad at Roboflow, who are building an essential platform and ecosystem in this space. We are thrilled to announce that GV has led Roboflow’s $40M Series B.

Roboflow provides an end-to-end collaborative platform for developers to build computer vision applications, encompassing multiple products that span the development lifecycle, including:

• Universe, the world’s largest collection of open computer vision datasets and models, counting 500M+ images, 500K+ datasets and 150K+ pre-finetuned models.
• Annotate, an AI-assisted data annotation tool to automate the labeling pipeline.
• Train, offering hosted training and multimodel finetuning with integrated modeling testing, experimentation and evaluation.
• Deploy, enabling managed deployment via Roboflow’s inference APIs, on-premise, in-VPC or at the edge.
• Workflows, an AI orchestration engine that enables enterprises to create visual agents, chain together models, build complex applications, and deploy them to the edge.

Across their collection of open source and commercial products, Roboflow has over 1M users and 25K organizations including over half of the Fortune 100. Their users represent every conceivable use case and vertical, serving mission-critical workflows across healthcare, manufacturing, public safety, retail, urban planning, farming, conservation and many more. And this is just the beginning.

Multiple forces are converging to catalyze a Cambrian explosion moment for the computer vision market. Multimodal foundation models are enabling a step function improvement in detection and analysis capabilities off the shelf, dramatically growing the number and complexity of addressable problem spaces for computer vision. Demand for AI capabilities on-device is growing, driving the need for tooling to support edge deployments to power intelligence in the physical world. Labor pressures in manufacturing, agriculture, healthcare and retail are necessitating greater automation and employee safety. Roboflow’s platform has matured over the past few years in partnership with its many users and enterprise customers, and they are perfectly positioned to leverage and accelerate these trends.

At GV, we have always believed in the power of open source communities as foundations for enduring technology platforms. Over the years, we’ve been fortunate to partner with some generational teams (Gitlab, Vercel, Snyk, Cockroach, Cloudera to name a few), and see in Roboflow many of the same ingredients: a vibrant developer community, a platform enabling an end to end workflow, a fanatical focus on engineering excellence and, of course, a track record of real value for enterprises.

We are delighted that Roboflow chose to partner with GV as they pursue their mission of “making the world programmable.” Congratulations team Roboflow!