BigQuery is 5x more than the snowflake and the base on the basis: what does Google do to make it even better

Google Cloud declared a A significant number of new features it Then Google Cloud Event last week, with at least 229 new messages.

Buried in this mountain by news that includes New AI chips and agent AI opportunities as well as Database updateEsGoogle Cloud also made some major movements with his BigQuery Data warehouse service. Among the new opportunities is BigQuery Unified Mevernance, which helps organizations discover, understand and trust their data assets. Management tools help to pay attention to key barriers to accept AI by providing data quality, accessibility and reliability.

Bets are huge for Google as they take rivals in the corporation data space.

BigQuery has been on the market since 2011 and has increased significantly in recent years, both in terms of opportunities and on the consumer base. Obviously, BigQuery is also a big business for Google Cloud. During Google Cloud, it was then discovered for the first time how big the business is. According to Google, BigQuery had five times more than the number of customers both on the snowflake and on the base.

“This is the first year in which we were allowed to actually publish customer statistics, which was delightful to me,” said Yasmin Ahmad, Managing Director of Data Analysis on Google Cloud, to VentureBeat. “Databricks and Snowflake, they are the only other type of corporate storage platforms on the market. We have five times more customers than each of them.”

How Google Improves BigQuery to Acceptance of the Enterprise

While Google now claims that there is a more consumer consumer base than its rivals, it also does not remove the leg from the gas. In recent months, and more specifically on Google Cloud Next, Hyperscaler has announced a number of new opportunities to improve businesses.

A key challenge for Enprise AI is access to the correct data that complies with the Business Service Level Agreements (SLAS). According to Gartner Research, quoted by Google, organizations that do not allow and support their AI use cases through AI, ready for data, will see over 60% of the projects for AI not to run business Slas and be abandoned.

This challenge stems from three permanent problems that struck the data management of the enterprise:

1. Fragmented data silos
2. Fast -changing requirements
3. Misguising organizational data crops where teams do not share a common language around the data.

Google’s unified management solution is a significant deviation from traditional approaches by incorporating management opportunities directly within the BigQuery platform and does not require individual tools or processes.

Unified Unified Management: Technical Deep Diving

At the heart of Google’s announcement is the Unifiade BigQuery Unified, powered by the new Universal Catalog BigQuery Universal. Unlike traditional catalogs that contain only basic information about the table and columns, the universal catalog integrates three different types of metadata:

1. Physical/technical metadata: Scheme’s definitions, data types and profiling statistics.
2. Business metadata: Conditions for a business dictionary, descriptions and semantic context.
3. Performance: Request models, statistics for use and form -specific technology information such as Apache Aceberg.

This unified approach allows BigQuery to maintain a complete understanding of data assets in the enterprise. What makes the system particularly powerful is how Google has integrated twins, its sophisticated AI model, directly into the control layer through what they call the engine of knowledge.

The knowledge engine actively enhances management by detecting relationships between data sets, enriching metadata with business context and automatic data quality monitoring.

Key capabilities include semantic search with natural language understanding, automated metadata generation, detection of AI relationships, packaging assets, business dictionary, automatic cataloging both structured and unstructured data and automatic detection of anomaly.

Forget about indicators, Enterprise AI is a bigger problem

Google’s strategy goes beyond the competition of the AI ​​model.

“I think there is too much of an industry that just focuses on targeting this individual chart, and in fact, Google thinks overall about the problem,” Ahmad said.

This comprehensive approach addresses the entire life cycle of enterprise data, answering critical questions such as: How do you get trusting? How do you deliver scale? How do you provide management and security?

As innovation in each layer of stack and combines these innovations, Google creates what Ahmad calls a real -time data flywheel, where, as soon as the data is captured, regardless of the type or shape or where it is stored, there is immediate generation of metadata, generic and quality.

This was said that the models matter. Ahmad explained that with the advent of thinking models like the Gemini 2.0, there is a huge unlock for Google data platforms.

“A year ago, when you asked Genai to answer a business question, everything that got a little more complicated, you’ll actually have to break it in many steps,” she said. “Suddenly, with the thinking model, he can come up with a plan … You don’t have to encode a firm way to build a plan. He knows how to build plans.”

As a result, she said that you can now easily have a data engineering agent to build a pipeline, which is three steps or 10 steps. Integration with Google AI capabilities transforms what is possible with Enterprise Data.

Influence in the real world: How businesses benefit

Levi Strauss & Company It offers a convincing example of how unified data management can transform business operations. The 172-year-old company uses Google’s data management opportunities as it is shifting from the fact that it is mainly a wholesale business to turning a brand directly to users. In a Google Cloud Next Vinay Narayana session, which manages data and AI platform Engineering in Levi’s, she details the case of using his organization.

“We strive to allow our business analysts to have access to real -time data, which are also accurate,” Narayana said. “Before we started our journey to build a new platform, we found different challenges of users. Our business users did not know where the data lived and if they knew the data source, they did not know who owned it. If they had any access to it, there was no documentation.”

Levi’s builds a Google Cloud data platform that organizes business domain data products, which makes it detected through Analytics Hub (Google Data Market). Each data product is accompanied by detailed documentation, information about genealogy and quality indicators.

The results were impressive: “We are 50 times faster than our Legacy Data Platform and this is at the low end. A significant number of visualizations are 100 times faster,” Narayana said. “We have over 700 users who already use the platform on a daily basis.”

Another example comes from Verizon, who uses Google Management Tools as part of his One Verizon Data initiative to combine Siled data into business units.

“This will be the largest Telco data warehouse in North America, which is moving on BigQuery,” Arvind Rajagopalan, AVP of Data Engineering, Architecture and Products of VerizonHe said during the next Google Cloud session.

The company’s data mansion is massive, including 3,500 users who put approximately 50 million requests, 35,000 data pipelines and over 40 five data.

In the Google Cloud spotlight session, AHMAD also provided many other users. Radisson Hotel Group customizes its advertising on a scale, training BigQuery twin models. The teams survived a 50% increase in productivity, while revenue from AI campaigns increased by over 20%. Gordon Food Service migrates in BigQuery, ensuring that their data are ready for AI and increase the reception of applications aimed at the client by 96%

What is the “big” difference: Study of a competitive landscape

There are numerous suppliers in the Enterprise Data warehouse space, including Databricks, Snowflake, Microsoft with Synapse and Amazon with Redshift. All these suppliers have been developing different forms of AI integrations in recent years.

Databricks has a A comprehensive Lakehouse data platform and has been expansion own AI optionsThanks partly to the acquisition of Mosaic of $ 1.3 billion. Amazon Redshift added support for generative AI in 2023, with Amazon Q help users build requests and get better answers. The snowflake, in turn Partnership with a large language model (LLM) Suppliers, including anthrop.

When pressed in comparisons, specifically with Microsoft’s suggestions, Ahmad claims that Synapse is not a platform for data for enterprise for the types of use cases for which customers use BigQuery.

“I think we have skipped the whole industry because we have worked on all the pieces,” she said. “By the way, we have the best model, this is the best model integrated into the data stack that understands how agents work.”

This integration has led to a quick acceptance of AI capabilities within BigQuery. According to Google, the use of Google AI customers in BigQuery for multimodal analysis has increased 16 times a year.

What does this mean to businesses receiving AI

For businesses that are already struggling with the implementation of AI, Google’s integrated approach to management may offer a more rational path to success than combining separate data management and AI systems.

Ahmad’s claim that Google is a “skipping” competitors in this space will face control as organizations set these new opportunities to work. However, customer examples and technical details suggest that Google has made significant progress in dealing with one of the most challenging aspects of Enterprise’s AI acceptance.

For technicians who evaluate data platforms, key questions will be whether this integrated approach provides sufficient extra value to justify migrating from existing investments in specialized platforms, such as a snowflake or database, and whether Google can maintain its current innovation rate as competitors respond.