SQream Technologies Ltd., the developer of a high-speed database that runs on graphics processing units, has raised $45 million in new funding.

Tel Aviv-based SQream announced the Series C funding round this morning. The round was led by World Trade Ventures with participation from several new and returning backers. Since its 2010 launch, SQream has raised more than $160 million in outside capital. 

A key difference between central processing units and GPUs is that the latter chips have more cores. The cutting-edge Sierra Forest server CPUs that Intel Corp. plans to launch next year will feature up to 144 cores. Nvidia Corp.’s flagship H100 graphics card, in turn, includes more than 18,000.

Each core is a miniature computing module that can carry out calculations independently from the other modules. Because a graphics card includes more cores than a CPU, it can carry out more calculations at once. SQream has developed a database, SQreamDB, that harnesses GPUs’ ability to carry out a large number of computations in parallel to speed up processing.

Before companies can analyze their business information for insights, they have to move that information from the systems where it’s usually stored into a centralized database. The task often takes a significant amount of time. According to SQream, SQreamDB can reduce information ingestion times by up to 90% and lower costs in the process.

The database includes tools that can be used to prepare newly ingested records for analysis. According to SQream, the tools make it possible to filter erroneous information and summarize the remaining data. Additionally, SQreamDB can perform denormalization, a process through which the contents of a database are reorganized in a manner that speeds up queries.

The software also optimizes data processing in other ways. It decouples compute from storage, which means companies can add storage capacity to a SQreamDB environment in small increments and avoid provisioning too much hardware. For added measure, the database compresses information to save space. 

Information stored in SQreamDB can be analyzed with standard SQL queries. Customers also have the option to run analyses using third-party business intelligence, or BI, platforms. SQreamDB provides integrations with Tableau and several other popular BI products.

“As generative AI shines a light on the importance of leveraging AI and ML within enterprises, as well as the value of GPUs as part of the analytics process, we have seen interest in our technology skyrocket,” said co-founder and Chief Executive Officer Ami Gal.

The company offers SQreamDB alongside a second product called SQream Blue. It’s a data lakehouse focused on a narrower set of use cases. In particular, the platform is designed to help companies organize their raw data into a form that can be more easily used to train artificial intelligence models and run analyses.

In 2021, SQream added a third product to its portfolio by buying a startup called Panoply for a reported $60 million to $70 million. The product it obtained through the deal is an ETL, or extract transform/load/platform. It can be used to collect data from multiple systems into a centralized environment for analysis.

SQream will use its latest $45 million funding round to further extend its feature set. As part of the effort, the company will prioritize the development of new AI capabilities. SQream also plans to grow its workforce in North America and expand its partner ecosystem.

Image: SQreamDB