Next-gen vector search platform Marqo raises $19.4m

Melbourne-based (but not much longer) vector search company Marqo, which harnesses the untapped information resource of unstructured data with artificial intelligence (AI), has raised US$12.5 million ($19.4 million) in a Series A round led by California-headquartered Lightspeed Venture Partners.

Blackbird dipped back in to further invest in the company after leading a US$5.2 million ($8 million) seed round for Marqo in August, while Singapore's January Capital participated again as well.

Marqo continues to secure investment from individual entrepreneurs too, with Chronosphere founder and CTO Rob Skillington participating in the Series A after Cohere's co-founders Ivan Zhang and Aidan Gomez took part in the seed round.

Over the past year the company has grown its monthly recurring revenue by almost 200-fold and doubled its head count to more than 25 staff, with milestones including the general release of Marqo's managed cloud service in August. Customers on its books include Temple & Webster (ASX: TPW), Envato Elements, CoreLogic, Redbubble (renamed as Articore Group on the ASX) and LingoLooper.

The latest funding will be used to advance the adoption of its next generation search platform, which unlocks the massive value of unstructured data across business-critical applications like end-user search, retrieval-augmented generation, and more. 

"We founded Marqo because we recognised that vector search was going to be instrumental in realising the full potential of AI in our day-to-day lives, but it is far too complicated for developers and enterprises to deploy," says Marqo CEO Tom Hamer, who co-founded the group with Jesse Clarke.

"We saw a need to invent a platform that not only generated superior vector embeddings but also empowered customers to build advanced search experiences within minutes, not months.

"This funding is validation of our approach and will help us scale up to meet the tremendous demand we’re seeing."

Unlike traditional vector databases and search tools, Marqo’s unique vector search platform handles the entire process from embedding generation to storage and retrieval, enabling seamless implementation of multimodal, multilingual search through a single API (application programming interface).

At the core of the platform is a proprietary inference engine that leverages state-of-the-art machine learning models to convert unstructured data into highly performant vectors that return hyper-relevant search results in real time. 

Marqo offers open-source access to its core code for developers as well as a fully managed cloud service, Marqo Cloud, which is optimised for production and suited to enterprises looking to build innovative search experiences and improve topline. 

"Mastering unstructured data will be the key to success in the AI race," says Jesse Clarke, formerly lead scientist at Amazon Robotics AI and principal scientist at StitchFix and physicist at Stanford and UCL. His fellow co-founder Tom Hamer was previously a software engineer at Amazon Web Services (AWS).

"Our platform transforms this challenge into an opportunity, creating vector embeddings that deliver intuitive and accurate search results with human-like understanding. It opens up tremendous possibilities for the future of search, smarter LLMs, and much more."

Lightspeed partner Hemant Mohapatra says the world of search is fast moving from ‘keyword-based’ to ‘natural language-based’ thanks to wide adoption of products such as ChatGPT.

"Marqo’s mission is to bring this transformational technology to every company in the world through a simple developer API and an enterprise-grade platform offered on-prem and on cloud," says Mohapatra.

"Their early growth has been phenomenal and we are excited to back this amazing team helping consumers to search the way they think."

Temple & Webster's head of AI, Anthony Ziebell, highlights the role Marqo has played in supporting the furniture e-retailer's search capacity to the benefit of customers.

"Discovery fuels our business so it's imperative that we deliver high quality and relevant search experiences to our customers," says Ziebell.

"With Marqo, we were able to deploy advanced vector search quickly and easily and see results instantly. 

"We went from sign-up to production A/B testing in five days and, within the next week, had rolled out a new feature to 100 per cent of our traffic after we saw an improvement in key metrics."

Marqo was initially incorporated by Hamer in London in April 2022, but shortly after he met Clark they established an office in Melbourne and moved back to Australia. Now Marqo will be moving office again, relocating its headquarters to San Francisco, California, where CEO Tom Hamer will also be based.

Marqo clarifies it will maintain its presence in Melbourne and London.