Encord Raises $60M to Rival Scale AI in Physical AI

Encord Secures $60 Million in Series C to Reshape Physical AI Data Infrastructure and Take On Scale AI

The artificial intelligence industry is no stranger to massive funding rounds, but every once in a while, a company comes along that fundamentally changes how the entire ecosystem thinks about a core problem. Encord, a San Francisco-based AI data infrastructure company, has done exactly that by closing a landmark $60 million Series C funding round — a move that positions it as one of the most ambitious challengers to established data labeling giants like Scale AI. This latest AI funding milestone not only reflects investor confidence in Encord's vision but also signals a much broader transformation happening in the world of physical AI, where the stakes — and the data challenges — are unlike anything the industry has faced before.

The funding round was led by Wellington Management, a globally respected investment firm with deep roots in technology-forward sectors. It also attracted participation from a strong roster of returning and new backers, including Y Combinator, CRV, N47, Crane Venture Partners, Harpoon Ventures, Bright Pixel Capital, and Isomer Capital. With this injection of capital, Encord's total funding now stands at $110 million, and the company's post-Series C valuation has been confirmed at $550 million. For a startup operating in a space that only recently started getting the attention it truly deserves, these numbers are nothing short of remarkable — and the AI funding news surrounding this round has set the stage for an exciting new chapter in how physical AI systems are built and trained.

What Is Physical AI and Why Does It Need New Data Infrastructure?

To understand why Encord's funding round matters so much, it's important to first understand what physical AI actually is and why it presents a uniquely difficult set of challenges that older data tools simply weren't designed to handle. Physical AI refers to artificial intelligence systems that are built to operate in the real world — systems like self-driving cars, autonomous delivery drones, industrial robots, and manufacturing automation platforms. Unlike large language models or image recognition tools that can be trained on text scraped from the internet or standard datasets, physical AI models must be trained on deeply complex, sensor-rich data. This includes video feeds, audio recordings, LiDAR point clouds, telemetry signals, and other multimodal inputs that are generated in real-time by physical machines operating in dynamic environments.

The challenge is enormous. Traditional data management platforms were built in a different era, designed primarily for static datasets and simpler labeling workflows. When a company building autonomous vehicles tries to use these older tools, they quickly discover that they weren't designed to handle petabyte-scale video data with the level of precision and speed that physical AI development demands. Teams end up spending an extraordinary amount of time managing data pipelines, manually annotating footage, and trying to stitch together a coherent training dataset — time that could and should be spent building better models. According to Encord's co-founders, machine learning teams typically spend more than 80% of their time on data-related tasks rather than on model development itself. This imbalance is what Encord was created to fix, and the recent AI funding news confirms that the market has taken notice in a big way.

The Vision Behind Encord: Building the Universal Data Layer for AI

Encord was co-founded by Ulrik Stig Hansen and Eric Landau, both of whom serve as co-CEOs of the company. Their backgrounds in big data systems and deep learning research gave them a front-row view of the exact problem they would go on to solve. Hansen and Landau recognized early on that while the AI industry was obsessively focused on model architecture and algorithmic innovations, the real bottleneck limiting AI performance was sitting quietly in the background — in the quality, curation, and evaluation of training data. Every breakthrough in model design was being held back by the messy, fragmented, and often manually intensive processes teams used to prepare and manage data.

Their response was to build something entirely different: an AI-native data infrastructure platform purpose-built for the unique demands of physical AI. Rather than offering a generic annotation tool with a few extra features bolted on, Encord designed its platform to handle the complete data lifecycle from the ground up. This means everything from initial data capture and organization, to sophisticated labeling workflows, to human feedback alignment and redeployment for retraining — all within a single unified environment. The result is a platform that doesn't just help teams work faster; it fundamentally changes how AI development teams interact with data, turning what was once a chaotic manual process into a streamlined, automated, and continuously improving workflow.

What makes Encord's architecture especially compelling is its data flywheel model. The platform uses existing AI models to help improve the quality of the very data being used to train new models. This creates a self-reinforcing feedback loop where the more data runs through the platform, the smarter and more efficient the data curation and annotation process becomes. This kind of compounding improvement over time is exactly the kind of structural advantage that investors look for when evaluating long-term bets in competitive markets — and it's a major reason why this AI funding round attracted the caliber of investors it did.

How Encord Differentiates Itself From Scale AI and Other Competitors

The competitive landscape for AI data infrastructure is not empty. Companies like Scale AI, Labelbox, and SuperAnnotate have been operating in this general space for years, building large customer bases and substantial revenues. Scale AI, in particular, has become something of a household name in enterprise AI circles, valued at tens of billions of dollars and backed by some of the biggest names in the venture world. So the question naturally arises: what makes Encord different, and why should AI teams choose it over these more established alternatives?

The answer lies in specificity and automation. While Scale AI and similar platforms offer broad, generalist data labeling and management services, Encord has made a deliberate strategic choice to go all-in on the physical AI use case. This focus means that every feature, every workflow, and every design decision is optimized specifically for the kinds of sensor-rich, multimodal data that autonomous systems generate. Encord's platform natively supports video, LiDAR, 3D point clouds, and other complex data types in ways that generalist tools simply don't. Its architecture was designed from the outset to handle petabyte-scale data operations without the performance degradation or workflow fragmentation that teams typically encounter when trying to push older platforms beyond their intended limits.

Beyond the technical architecture, Encord's platform is built around active learning and human-in-the-loop automation systems that continuously improve data quality over time without requiring proportional increases in manual labor. This is particularly valuable for companies in the robotics, automotive, and drone sectors, where data is constantly being generated, updated, and revised as real-world conditions change. Rather than having teams repeatedly re-annotate similar data from scratch, Encord's system learns from past annotations and applies that knowledge intelligently to new datasets — dramatically reducing the time and cost of keeping training data current and high-quality. For the AI world, this represents a meaningful leap forward in how AI funding is being deployed to solve practical infrastructure challenges rather than just building more models.

Growth Metrics, Key Clients, and Global Expansion Plans

The numbers behind Encord's growth are perhaps the most compelling argument for the company's trajectory. Over the past year alone, the volume of data managed on Encord's platform grew from one petabyte to five petabytes — a fivefold increase in a single year. Revenue generated from physical AI customers grew tenfold during that same period. These aren't the kind of growth figures that emerge from modest tailwinds; they reflect a company that has found genuine product-market fit in a rapidly accelerating market and is executing on its vision with real discipline and focus.

Among Encord's most notable clients are Woven by Toyota, the autonomous vehicle subsidiary of one of the world's largest automakers; Zipline, the innovative drone delivery company operating across multiple continents; Skydio, a leader in autonomous drone technology; and AXA Financial, demonstrating that the platform's applications extend beyond pure robotics into enterprise financial AI use cases as well. The diversity of this client base is telling — it shows that Encord's platform is flexible enough to serve organizations operating in very different industries while still being purpose-built enough to outperform generalist alternatives in each of those contexts.

With the fresh capital from this AI funding round now in hand, Encord has laid out a clear roadmap for what comes next. The company plans to accelerate product development significantly, building out new capabilities that keep pace with the rapidly evolving demands of the physical AI market. It is also actively expanding into new international markets, looking to capture a share of the global demand for AI data infrastructure as companies around the world accelerate their own autonomous systems development programs. With a team of over 150 employees spanning offices in the United States and the United Kingdom, representing more than 40 nationalities, Encord already has the multicultural, globally distributed DNA needed to support that international expansion effectively.

The broader context of this AI funding news is also worth noting. Investment in physical AI infrastructure is growing rapidly as the technology matures from experimental prototypes in research labs to large-scale commercial deployments across multiple industries. The companies that will define the next decade of AI development are not just the ones building the most impressive models — they are the ones building the infrastructure that makes those models possible to train, deploy, and improve at scale. Encord's $60 million Series C represents a powerful vote of confidence from the investment community that this is precisely the kind of infrastructure the AI world needs right now.

Why This Funding Round Matters for the Future of AI Development

Stepping back and looking at the bigger picture, Encord's Series C is more than just another AI funding news headline. It represents a maturing recognition within the investment community that the data layer of AI development has been systematically underinvested relative to its importance. For years, the lion's share of AI investment went toward model development, compute infrastructure, and application-layer software. Data management was treated almost as an afterthought — a necessary operational cost rather than a strategic capability. Encord's growth and its ability to raise at a $550 million valuation are part of a broader shift in how investors, enterprises, and AI teams think about data.

Physical AI — the kind that will drive our cars, deliver our packages, and run our factories — cannot be built without high-quality, well-managed, continuously updated training data. The tools that exist to manage that data today are either too general, too manual, or too limited in scale to meet the demands of the next generation of autonomous systems. Encord is building an answer to that challenge, and the market response — both from enterprise customers and from blue-chip investors — suggests it is on the right track.

For the broader AI community, this funding round is a reminder that some of the most important innovations in artificial intelligence aren't happening at the model architecture level. They are happening in the infrastructure layers that most people never see — in the pipelines, annotation platforms, and data curation systems that quietly make all the impressive AI demos and deployments possible. As The AI World continues to track the most significant developments across the global AI ecosystem, Encord's story stands out as a particularly instructive example of where smart capital is flowing and why. It is a story about identifying a critical bottleneck, building a purpose-built solution, and executing with enough conviction to attract world-class investors and enterprise customers alike. That is exactly the kind of AI funding news that deserves close attention from anyone seeking to understand where the future of artificial intelligence is truly being built.