What Eventual does

Eventual is a US-incorporated private company that builds data infrastructure for AI and Physical AI workloads. Founded by engineers who previously built petabyte-scale data systems for model training in self-driving at Lyft, the company addresses a recurring bottleneck in AI development — research teams reportedly spending 80% of their time on ETL, hard-example mining, and dataloading rather than model training. Eventual serves two primary segments: AI/ML teams requiring large-scale data infrastructure for model training, and Physical AI teams that need purpose-built handling of video, lidar, and high-frequency unstructured sensor data.

Eventual operates a dual-product portfolio. Daft is an open-source data engine purpose-built for GPU workloads, video, lidar, and high-frequency unstructured logs; it runs in production at exabyte scale at Amazon, Mobileye, and Together AI. MultiBase is the commercial next-generation platform built natively for video and high-frequency sensor data, providing deep semantic indexing of perception data beyond embeddings and enabling complex temporal and action-oriented queries in plain English (e.g., "left-arm grasp failures on deformable objects"). Both products operate directly on standard open formats (mp4, jpeg) in storage the customer owns, eliminating custom formats, ETL, and dataloading pipelines.

Eventual employs a product-led growth model anchored by the open-source distribution of Daft through GitHub, with supporting community channels including a Slack workspace, a Substack newsletter, technical blog, and presence on X, LinkedIn, and YouTube. MultiBase is positioned as a subscription-based commercial product sold directly to enterprise Physical AI teams via the eventual.ai website. Pricing, revenue, funding, headcount, and office locations are not disclosed in the available source material; no institutional investors or ownership structure details are provided.

Name

Eventual

Legal name

Eventual Inc.

Website

https://eventualcomputing.com

Company type

Private

Operating status

Operating

Short description

Eventual builds data infrastructure for AI and Physical AI workloads. Its open-source Daft engine runs at exabyte scale in production at Amazon, Mobileye, and Together AI, while MultiBase enables semantic indexing and natural-language querying of video and sensor data.

Ownership category

akta.pro rank

Product category

AI Data Infrastructure

NAICS

Computing Infrastructure Providers, Data Processing, Web Hosting, and Related Services (518), Software Publishers (51321)

SIC

Services-Computer Programming, Data Processing, Etc. (7370), Services-Computer Processing & Data Preparation (7374)

akta.pro primary industry

Data & Analytics Platforms (Data Warehousing, Lakes, Streaming) (HDABAAAF)

akta.pro secondary industries

Enterprise AI Data & Knowledge Platforms (Vector Databases, Knowledge Graphs) (HDAEANAH), Real-Time / Streaming Data Warehousing (HDAEABAF), AI Memory & Storage for Training (HBM/DDR/SSD/Object Storage) (HDAAAAAF)

Keywords

• AI data infrastructure
• Data processing engine
• Physical AI infrastructure
• Unstructured data processing
• Sensor data platform

GTM type

B2B

Offering type

Software

Cost components

Personnel, Technology or R&D, Marketing or Sales, Operations, Infrastructure

Revenue model

1. MultiBase Commercial Product: MultiBase is the commercial product for Physical AI teams, representing the revenue-generating offering built on top of the open-source Daft foundation.
2. Open Source (Daft): Daft is open-source, providing community adoption and integration ecosystem that may lead to enterprise adoption and commercial offerings.

Go-to-market motion1 record

Distribution channels2 records

Marketing channels6 records

Core offering

Eventual builds data infrastructure software for AI and Physical AI workloads. Its flagship open-source product Daft is a distributed data engine that processes GPU workloads, video, lidar, and high-frequency unstructured logs at exabyte scale, deployed in production at Amazon, Mobileye, and Together AI. MultiBase is the commercial next-generation product providing semantic indexing of perception data so Physical AI teams can run complex temporal and action-oriented queries on petabytes of fleet data in plain English and feed curated data to GPUs at line rate.

Product overview

Eventual offers a dual-product portfolio: Daft is an open-source data engine for large-scale AI and ML workloads, while MultiBase is the company's flagship data infrastructure product for Physical AI teams. Daft serves as the foundation (running in production at exabyte scale at Amazon, Mobileye, and Together AI), and MultiBase represents the next generation platform purpose-built for video and high-frequency sensor data processing. Both products operate on standard open formats (mp4, jpeg) without custom formats or ETL requirements.

Differentiator

Problem solved

Functional benefit

Brands

• Daft: Open-source data engine capable of running in production at exabyte scale, deployed at Amazon, Mobileye, and Together AI
• MultiBase

Products and services

• Daft Open-source distributed data engine designed for large-scale AI workloads, processing GPU data, video, lidar, and high-frequency unstructured logs at exabyte scale. Used in production by AI and ML engineering teams for model training data infrastructure.
• MultiBase Commercial data infrastructure platform for Physical AI teams, built natively for video and high-frequency sensor data. Provides deep semantic indexing of perception data beyond embeddings, enabling complex temporal and action-oriented queries on petabytes of fleet data in plain English and feeding curated data to GPUs at line rate to maximize Model FLOP Utilization (MFU).

Quantifiable outcome

• Researchers previously spent 80% of their time wrestling with data workflows instead of training models

Named customers3 records

Segments3 records

Ideal customer profiles3 records

Technology focussed Yes

API detail

Has API

No

API docs

API detail

Core technology

AI maturity

App detail

AI capability7 records

Feature3 records

Partnerships

Three partnerships are on record, tiered major.

• AmazonmajorOthersAmazon runs Daft in production at exabyte scale. According to AWS blog, Amazon conducted an exabyte-scale migration from Apache Spark to Ray on Amazon EC2, which references Daft as the underlying data engine enabling this scale.
• MobileyemajorOthersMobileye runs Daft in production for handling sensor and fleet data at scale, representing a significant production deployment in the autonomous driving sector.
• Together AImajorOthersTogether AI runs Daft in production for AI/ML infrastructure workloads, demonstrating adoption in the generative AI and inference platform space.

Scale indicators3 records

Recent moves5 records

Expansion highlights4 records

Emerging players

• Pinecone: Pinecone is a managed vector database purpose-built for AI retrieval and semantic search. While narrower than Eventual (vectors vs. multimodal perception data), it competes for the same AI data infrastructure budget and overlaps with MultiBase's semantic indexing layer for retrieval-augmented AI workflows.
• LanceDB: LanceDB is an open-source multimodal AI database built on the Lance columnar format for video, images, and embeddings. It targets the same 'AI-native multimodal data lake' niche as Daft/MultiBase and represents a credible emerging open-source competitor with overlap in perception data indexing.
• Weaviate: Weaviate is an open-source vector and hybrid search database used for AI-native applications. Like Pinecone, it sits in the semantic retrieval layer adjacent to MultiBase's perception indexing, and competes for the same AI data platform wallet share.

Broad incumbents

• Scale AI: Scale AI provides data labeling, evaluation, and curation infrastructure for AI, with deep penetration in autonomous driving and Physical AI. It addresses an adjacent step in the AI data lifecycle (annotation/curation) and competes for Physical AI customers like Mobileye that Eventual already serves.
• Amazon Web Services: AWS offers S3, Glue, EMR, and managed Spark/Ray environments that constitute the surrounding data and compute substrate in which Daft runs. AWS is both a major customer (exabyte Daft deployment) and the most powerful potential competitor that could natively replicate Eventual's data engine capabilities.
• Snowflake: Snowflake is a leading cloud data platform pushing into AI workloads via Cortex and Iceberg Tables. It overlaps with Eventual on data warehousing and analytics infrastructure for AI/ML use cases and is actively expanding down-stack toward unstructured and AI-specific data.
• Databricks: Databricks operates the Lakehouse platform that combines data warehousing, ETL, and AI/ML workloads (including acquired MosaicML for model training). It competes with Eventual for AI/ML data infrastructure budgets at much larger scale and is the most likely incumbent to expand into Physical AI/video/lidar perception workloads.
• Google Cloud (BigQuery): Google Cloud's BigQuery and Vertex AI stack covers serverless data warehousing and AI/ML infrastructure that overlaps with Eventual's positioning. As a hyperscaler with deep investment in AI, Google represents both a future partner channel and a platform risk similar to AWS.

Direct peers

• Anyscale: Anyscale commercializes Ray, the distributed compute framework referenced in Amazon's exabyte-scale migration that also features Daft. Anyscale is the closest direct peer in AI-native distributed data and compute infrastructure, and the two frequently co-exist in customer stacks (Ray for orchestration, Daft for data engine).

Others

• DuckDB: DuckDB is an open-source in-process analytical database with strong adoption in the Python data ecosystem. While focused on OLAP rather than exabyte-scale GPU workloads, it represents the broader open-source data-engineering community that Daft competes with for developer mindshare and contributor attention.

Market position

Strengths5 records

Weaknesses5 records

Competitive moat5 records

Key risks7 records

Key highlights7 records

Customer concentration

• Xx.com/daftengine
• GitHubgithub.com/Eventual-Inc/Daft
• YouTubeyoutube.com/@daftengine
• LinkedInlinkedin.com/company/eventualcomputing

Revenue estimate

Valuation estimate

Number of profiles

Funding overview

Funding rounds

Investors

Funding detail is available on the Subscription and Enterprise plan.Contact sales →

M&A

Investments

M&A and investment is available on the Subscription and Enterprise plan.Contact sales →

Frequently asked questions about Eventual