Snowflake, a Data Cloud company, is planning to acquire Reka AI for more than $1 billion in order to develop the world's leading AI ecosystem. The acquisition aims to strengthen Snowflake's position in the AI market and enhance its large language models (LLMs). Reka AI, which has received funding from Snowflake and other tech companies, specializes in LLMs. Snowflake's new CEO, Sridhar Ramaswamy, is an AI expert, and the company has also invested in AI startup Metaplane. The acquisition is expected to further Snowflake's commitment to AI research and development. Snowflake will release its fiscal 2025 first quarter financial results on May 22. [6c5570e1]
In addition to the acquisition of Reka AI, Snowflake has recently launched Arctic, an open enterprise-grade large language model (LLM). Arctic utilizes a unique Mixture-of-Experts (MoE) architecture and has outperformed industry benchmarks in SQL code generation and instruction following. Snowflake is making Arctic's weights available under the Apache 2.0 license to encourage collaboration and innovation. Arctic is part of Snowflake's Arctic model family, which includes text-embedding models optimized for retrieval use cases. Arctic activates only a fraction of its 480 billion parameters at a time, using significantly fewer parameters compared to other LLMs in the market. Snowflake Cortex will integrate Arctic to accelerate the development of production-grade AI applications. The launch of Arctic and the acquisition of Reka AI demonstrate Snowflake's commitment to democratizing AI and driving innovation across industries. [47ae7c56]
AtScale, a leading provider of semantic layer solutions, has announced the launch of its Semantic Layer Platform on Snowflake Marketplace. The platform offers consistent and governed semantic data definitions for self-service BI and AI initiatives within Snowflake accounts. AtScale and Snowflake are collaborating to enable customers to deliver KPI consistency, enable OLAP-style analytics, and provide low-latency BI performance. The launch of AtScale's Semantic Layer Platform was announced at Snowflake's annual user conference, Snowflake Data Cloud Summit '24. [b2b74a0b]
Hopsworks AB, a Swedish data lakehouse company, is releasing Hopsworks 4.0, which it claims is the industry's first unified platform for building batch, real-time, and large language model (LLM) artificial intelligence systems. The release includes a new feature store, vector database index system, native Python access, and more granular fine-tuning features for LLMs. The platform supports the development and operation of data pipelines at scale, and offers a centralized repository for managing features, experiments, AI models, and data assets. Hopsworks 4.0 also enhances performance and availability, adds vector and similarity search capabilities, and improves support for Kubernetes. The platform provides higher throughput when reading data from the lakehouse compared to platforms like Databricks and Vertex AI. The release is expected to be generally available soon. [ee507153]