Vector databases are the unsung heroes of the AI landscape, providing the foundation for storing and retrieving unstructured data such as images, videos, and text. They empower individuals and systems to efficiently search through disorganized content. These databases are critical for large language models like GPT-4, enabling real-time indexing and search as new data is generated or updated. This functionality supports personalization, recommendations, sentiment analysis, and more.
The rise in generative AI has sparked the growth of several vector database startups, attracting substantial attention and funding. Pinecone and Weaviate secured $100 million and $50 million respectively in April, while Chroma and Qdrant, newer players in the field, received $18 million and $7.5 million respectively in seed financing during the same month. Last year, Zilliz, the developer behind the Milvus open-source vector database, raised $60 million in funding.
In the midst of this surge, Marqo, an Australian startup, strives to differentiate itself by adopting a comprehensive “end-to-end” strategy for vector search.
Pioneering Solutions to Data Chaos
Marqo, co-founded by Jesse Clark, former lead machine learning scientist at Amazon’s Seattle robotics unit, and Tom Hamer, previously a database software engineer at Amazon Web Services (AWS) in Sydney, addresses the problem of unstructured data. This type of data accounts for up to 90% of all generated data, highlighting the need for efficient tools to understand and handle it.
Marqo sets itself apart with its comprehensive suite of vector search capabilities that are readily available. With built-in vector generation, storage, and retrieval, there’s no need for external tools like OpenAI or Hugging Face. Marqo offers a seamless experience through a single API, giving it a significant edge over existing solutions.
Tom Hamer, co-founder and CEO of Marqo, highlights the significance of Marqo’s end-to-end system that seamlessly integrates all these components, addressing a significant challenge for developers.
Enhancing Search Quality
Generating relevant, accurate, and up-to-date search results is crucial for effective search systems. Marqo tackles this challenge with its inherent capabilities. As Hamer explains, “Marqo’s continuous learning technology enables search to automatically enhance itself through user engagement… a key aspect for ecommerce and other end-user search scenarios.”
Marqo’s progress is evident from its funding trajectory. After successfully raising £660,000 ($840,000) in pre-seed funding last year, it has now announced an additional $4.4 million in seed funding to enhance its commercial endeavors. This funding will facilitate the launch of a new cloud service that complements its existing open-source project.
Empowering Developers with Open Source
Marqo follows in the footsteps of its rivals by embracing an open-source approach. This empowers developers to customize and assess the product according to their requirements. By fostering engagement, soliciting recommendations, and encouraging contributions, this approach paves the way for a superior outcome.
According to Tom Hamer, the utilization of an open-source foundation for Marqo enabled us to establish a strong feedback loop with users and iterate rapidly. Moreover, open source serves as an excellent channel for customer acquisition.
Marqo Cloud offers a solution for open-source projects that eliminates the need for extensive resources. With Marqo Cloud, users can rely on a cloud platform that takes care of infrastructure, maintenance, and operations, guaranteeing top-notch performance and efficiency.
Global Ambitions
Marqo, while being based in Australia, has set up a parent company in the U.K., where its initial investor, Creator Fund, is situated. Additionally, the startup has a small office in London and intends to expand its European operations to encompass sales, marketing, and customer support.
Marqo recently raised seed funding with Australian VC firm Blackbird Ventures leading the round. Creator Fund, who also invested in the pre-seed round, participated as well. Notably, January Capital and the co-founders of Cohere, Ivan Zhang and Aidan Gomez, also made additional investments.
