Snowflake Unveils Arctic: An Open Source LLM for Enterprise AI Development

Snowflake, a leading data cloud vendor, took a significant step toward enabling enterprise-grade AI development with the launch of Arctic, a new open source large language model (LLM). The release of Arctic comes just weeks after Snowflake’s rival, Databricks, introduced DBRX, another open source LLM aimed at helping users leverage AI for business decision-making.

The introduction of Arctic also follows closely on the heels of a leadership change at Snowflake. In May 2023, CEO Frank Slootman stepped down, and Sridhar Ramaswamy, formerly of search engine vendor Neeva, took the helm. Ramaswamy’s appointment seemed to signal a shift towards a stronger emphasis on AI, including Generative AI at Snowflake.

Designed for Enterprise-Specific Tasks

Arctic was designed with the specific needs of enterprises in mind. According to Baris Gultekin, Snowflake’s head of product for AI, the LLM excels at business-specific tasks such as generating SQL code and following instructions, making it particularly well-suited for enterprise AI development.

“LLMs in the market do well with world knowledge, but our customers want LLMs to do well with enterprise knowledge,” Gultekin explained during a media event.

This focus on enterprise-specific capabilities sets Arctic apart from other open source LLMs and positions it as a valuable tool for businesses looking to develop AI models and applications tailored to their unique needs.

Secure and Cost-Effective AI Development

One of the key advantages of Arctic for Snowflake customers is the ability to develop AI models within the same secure environment where they store their data. This eliminates the need to move data to an outside entity or import an external LLM into the Snowflake environment, reducing the risk of data breaches.

“Information architectures are complex,” noted David Menninger, an analyst at ISG’s Ventana Research. “Any time you can reduce the number of moving parts through tighter integration, it is a potential improvement for customers. In addition, it is always easier to get support for a single vendor solution than a multivendor solution.”

Furthermore, Arctic’s mixture-of-experts architecture enables performance efficiency, potentially leading to cost savings for enterprises. With the rising costs associated with cloud computing and AI development, this efficiency is critical for businesses looking to control spending while still leveraging the power of AI.

Performance and Productivity Gains

Initial benchmark testing conducted by Snowflake suggests that Arctic is competitive with other open source models, such as DBRX and Llama3 70B, in tasks like generating SQL and other coding, following instructions, performing math, and applying common sense and knowledge. The new model even outperformed others in coding, according to Snowflake’s internal testing.

However, independent testing and early customer success stories will be crucial in demonstrating Arctic’s real-world performance and building confidence among potential users. As Menninger pointed out, “Vendors are playing leapfrog right now, which is great for customers. We are seeing improvements in accuracy, cost of training, and cost of inferencing.”

The Road Ahead for Snowflake’s AI Journey

While the launch of Arctic is a significant milestone for Snowflake, the company still has some ground to cover in its AI journey. The perception remains that Snowflake trails its closest rival, Databricks, in enabling customers to build AI applications.

Snowflake’s fully managed environment for AI development, Cortex, is still in preview and is scheduled to be generally available in the coming weeks. Until Cortex is officially backed by Snowflake for production use, some customers may be hesitant to fully embrace the vendor’s AI offerings.

However, under Ramaswamy’s leadership, Snowflake has been making aggressive moves to bolster its AI capabilities. In addition to the launch of Arctic, the company has partnered with Mistral AI, introduced data clean rooms, and is preparing for the general availability of Cortex. These developments demonstrate Snowflake’s commitment to providing users with the tools they need to develop AI applications and stay competitive in the rapidly evolving AI landscape.

As Doug Henschen, an analyst at Constellation Research, noted, “It’s good to see Snowflake moving quickly under new CEO Sridhar Ramaswamy to catch up in the GenAI race. It’s important for data platforms to cover all bases, including analytics, machine learning, AI, and now GenAI.”

Looking ahead, Snowflake would be wise to prioritize AI governance to help organizations navigate the complex regulatory landscape surrounding AI. By providing guidelines and tools for responsible AI development and deployment, Snowflake can further differentiate itself as a trusted partner for enterprises embarking on their AI journeys.

With the launch of Arctic and the impending general availability of Cortex, Snowflake has taken significant strides towards enabling enterprise-grade AI development. As the company continues to innovate and evolve its AI offerings, it is well-positioned to empower businesses to harness the full potential of AI and drive meaningful business outcomes.

---

If you need assistance understanding how to leverage Generative AI in your marketing, advertising, or public relations campaigns, contact us today. In-person and virtual training workshops are available. Or, schedule a session for a comprehensive AI Transformation strategic roadmap to ensure your marketing team utilizes the right GAI tech stack for your needs.

Read more: Snowflake Unveils Arctic: An Open Source LLM for Enterprise AI Development