As soon as seen purely as a knowledge warehousing powerhouse, Snowflake is present process a serious reinvention. At its Snowflake Summit 2025, underway in San Francisco, the corporate unveiled a sweeping set of AI merchandise that reimagine how knowledge is ingested, processed, and become intelligence, all inside one unified platform.
Conventional ETL (Extract, Rework, and Load) processes usually contain integrating a number of separate instruments, akin to Talend, Informatica for knowledge integration, Airflow for orchestration, and Spark for processing, to construct complicated knowledge pipelines. These instruments are usually mixed to deal with extraction, transformation, and loading duties, which might result in complexity, increased prices, and upkeep overhead.
Then again, Snowflake’s Openflow, a brand new multimodal ingestion service powered by Apache NiFi, helps enterprises pull in knowledge from various sources and codecs into Snowflake’s AI Information Cloud.
“That is the productisation of an acquisition we made a number of months in the past of an organization referred to as Datavolo. Openflow is a managed service that helps organisations each extract knowledge from a wide range of sources and be capable of course of it,” mentioned Christian Kleinerman, EVP of product at Snowflake, in a media briefing.
Openflow permits prospects to maneuver knowledge from the place it’s created to the place it’s wanted, supporting each batch and streaming modes. It options a whole lot of pre-built connectors and processors, and affords extensibility to construct customized connectors. The service helps Snowflake’s Convey Your Personal Cloud deployment mannequin and is now typically accessible on AWS.
Furthermore, the platform removes the present bottlenecks in knowledge engineering, together with inflexible pipelines, fragmented stacks, and sluggish ingestion. Openflow helps each structured and unstructured knowledge and integrates with sources like Field, Google Adverts, Oracle, Salesforce Information Cloud, Workday, and Microsoft SharePoint.
“Most of our prospects are fascinated about loading knowledge into Snowflake or making it accessible to Snowflake,” mentioned Kleinerman. He additional added that their purpose is to simplify knowledge motion and processing from anybody supply to another vacation spot.
With Openflow, Snowflake can also be extending its knowledge engineering capabilities. Prospects will quickly be capable of run dbt Tasks natively in Snowflake with assist for options like in-line AI code help and Git integration.
The aptitude will probably be accessible inside Snowflake Workspaces, a brand new file-based improvement setting. These initiatives will finally be powered by dbt Fusion.
Snowflake additionally introduced expanded assist for Apache Iceberg tables, which permits organisations to construct a related lakehouse view and entry semi-structured knowledge utilizing Snowflake’s engine. New optimisations for file dimension and partitions are anticipated to enhance efficiency and management.
Snowpipe Streaming, now in public preview, provides assist for high-throughput, low-latency knowledge ingest, with knowledge changing into queryable inside 5 to 10 seconds. This additional improves Openflow’s potential to handle near-real-time knowledge streams.
From Information to Motion
Moreover, Snowflake has introduced new agentic AI choices at its annual person convention, together with two improvements referred to as Snowflake Intelligence and Information Science Agent.
Snowflake Intelligence, launching quickly in public preview, permits non-technical customers to question and act on structured and unstructured knowledge by way of pure language prompts.
The product is powered by Cortex Brokers and LLMs from OpenAI and Anthropic, and runs instantly inside prospects’ Snowflake environments, inheriting safety and governance controls.
“Snowflake Intelligence breaks down these boundaries by democratising the flexibility to extract significant intelligence from an organisation’s complete enterprise knowledge property — structured and unstructured knowledge alike,” mentioned Baris Gultekin, head of AI at Snowflake.
Snowflake Intelligence additionally incorporates third-party content material by way of Cortex Data Extensions, together with CB Insights, Packt, Stack Overflow, The Related Press, and USA TODAY.
Then again, Information Science Agent automates core machine studying duties utilizing Claude from Anthropic. These duties embody knowledge preparation, characteristic engineering, and mannequin coaching. The agent offers verified ML pipeline code and permits customers to iterate by way of strategies or follow-ups.
“We’re leveraging AI to assist prospects create machine studying pipelines, writing code, validating it, and finally automating the end-to-end ML lifecycle,” mentioned Kleinerman.
The corporate claims the agent reduces the time spent on debugging and experimentation, permitting knowledge scientists to prioritise higher-impact work.
These launches are a part of Snowflake’s broader push to allow enterprise AI use circumstances. For analytics, Snowflake has additionally launched AISQL,which extends its SQL language to incorporate AI operations as easy operate calls.
“The purpose of that is to convey the ability of AI to analysts and personas which can be usually snug with database know-how,” Kleinerman defined. This contains processing textual content for sentiment evaluation and classification, and supporting multimodal knowledge like PDFs, audio, and pictures.
Analysts can now enrich tables with chat transcripts, correlate sensor knowledge with pictures, and merge structured knowledge with sources like social media sentiment—multi function interface.
The software integrates with sources like Field, Google Drive, Workday, and Zendesk utilizing Snowflake Openflow and helps pure language conversations that return insights, generate visualisations, and floor enterprise information.
The corporate additionally launched SnowConvert AI, an agent that automates knowledge migrations from platforms akin to Oracle, Teradata, and Google BigQuery. It reduces the necessity for guide code rewriting and validation, and accelerates database, BI, and ETL migration processes by two to a few occasions.
“SnowConvert AI allows organisations to shortly and simply transfer from legacy knowledge warehouses… whereas staying supported and with out disrupting important workflows,” the corporate mentioned.
With these launches, Snowflake is shifting past the standard knowledge warehouse, positioning itself as a full-stack AI platform for enterprises, spanning ingestion, processing, and clever automation.
The publish Snowflake Simply Killed the Information Pipeline as We Know It appeared first on Analytics India Journal.
