The growing demand for generative AI apps has led to a necessity for increasingly larger databases to store the associated data (e.g. model training data). These databases are likely to be resource-intensive from a hardware perspective and, depending on the algorithms used to orchestrate them, can have high latency. Companies are sometimes forced to make trade-offs between database cost, performance, and accuracy.
But it doesn’t should be that way, says Ohad Levi, CEO and co-founder of Hyperspace. Hyperspace uses “domain-specific computing” to hurry up two specific database tasks: lexical searches and vector searches. Lexical searches are a variety of keyword-based search that appears for exact matches in a database, while vector searches bear in mind the semantic meaning and context of the search query.
Levi claims that the instances of Hyperspace, that are a mixture of FPGA and GPUs can enable searches as much as ten times faster than traditional, non-accelerated databases.
“Our product helps firms coping with data search at scale, especially in AI and generative AI applications,” Levi told TechCrunch. “Unstructured data is overtaking traditional search capabilities. Data search solutions must support lexical and vector search datasets to satisfy current market demands.”
Before founding Hyperspace, Levi was an optimization engineer at Intel after which a product marketing manager at HP. He says he was frustrated by the constraints of the old search solutions that worked for the massive tech firms, so he teamed up with former Intel design consultant Max Nigri to found Hyperspace.
Hyperspace doesn’t sell its instances. Instead, it sells access to managed database software that runs on those instances (currently hosted on AWS). Hyperspace's databases can handle various forms of structured and unstructured data, including video, images, and text, and are priced by size and query volume.
“Hyperspace is a cloud-native managed database that operates as a software-as-a-service model and is billed per usage,” explained Levi. “Our team can develop customized AI infrastructure solutions to assist firms solve their search problems.”
Hyperspace's performance gains are impressive once they're true. Levi says the corporate's instances also deliver five times the throughput at 50% lower cost than a typical database. (These are averages; in a selected comparison point, Levi claims Hyperspace is usually faster than Elastic.) But can Hyperspace persuade firms to make use of a brand new database platform when there are such a lot of established providers—like Azure, AWS and Google Cloud—to select from?
Levi says yes, claiming that Hyperspace is already seeing initial customer gains. The Tel Aviv-based company has signed deals with firms within the fraud prevention and e-commerce sectors, including Forter, nSure and Renovai, and has tripled its annual recurring revenue and total contract volume within the last yr.
Hyperspace also recently closed a $9.5 million seed funding round led by MizMaa with participation from JVP and toDay Ventures. Levi says the cash will probably be used to expand Hyperspace's database offering to “hundreds” of instances and launch a free starter plan.
“Hyperspace has a complete recent set of revolutionary products that can advance the search market and support the needs of our large enterprises and small and medium-sized customers,” said Levi. “We see no headwinds. Every generative AI system is a search system, and search is getting harder than ever. The need for higher AI infrastructure is growing each day, and with more data, the necessity for higher search applications becomes more apparent.”
