Snowflake’s current quarter numbers have made Wall Street happy. Snowflake showed a stronger-than-expected sales forecast for the current quarter, indicating that its new generative AI-focused products are driving faster growth.
Snowflake has announced that its product revenue is projected to be between $805 million and $810 million for the current quarter ending in July, up 34% year-over-year. This surpasses forecasts from analysts who projected a revenue of $787.5 million. Additionally, the company increased its annual product sales projection to $3.3 billion from $3.25 billion.
Snowflake CEO Sridhar Ramaswamy said, “Our AI products, now generally available, are generating strong customer interest. They will help our customers deliver effective and efficient AI-powered experiences faster than ever.” Ever since Ramaswamy joined Snowflake after its acquisition of his company, Neeva, generative AI has been one of its biggest focuses.
All About Generative AI
Snowflake had been negotiating to acquire generative AI startup Reka AI for over $1 billion, but the discussions ended without an agreement. In April, Snowflake introduced its own LLM suite called Arctic and now allows customers to utilise third-party AI models, such as Mistral and Meta, on their data within the company’s platform, dubbed Snowflake Cortex.
But apart from the Reka AI deal that fell through, Snowflake announced a definitive agreement to acquire TruEra, an AI startup specialising in tools for testing, debugging, and monitoring machine learning models and large language model applications in production.
Snowflake announced plans to “acquire certain technology assets and hire key employees” from the AI-focused startup, which raised $25 million in 2022.
On the acquisition, TruEra co-founder, president, and chief scientist Anupam Datta said, “We are looking forward to this next phase in our journey with the Snowflake team with whom we share a commitment to delivering effective & trustworthy generative AI and predictive ML at scale.”
This acquisition marks Snowflake’s sixth significant investment to enhance the capabilities of its data cloud and its third major initiative in the data observability space. Prior to this, Snowflake had invested in two monitoring solutions companies – Observe and Metaplane.
In a blog post, Snowflake highlighted that the TrueEra acquisition will help the company ensure more accuracy and trustworthiness with the data used for training AI models.
TruEra has been instrumental in solving the black box problem with AI and the team behind the startup are experts in RAG-based solutions. Following the acquisition, all three co-founders will be joining Snowflake alongside the TrueEra team.
This AI-focused approach is enabling Snowflake to better compete with others in the field, such as Databricks, which offers similar services and, interestingly, has employed a similar acquiring strategy ever since it acquired MosaicML.
In an exclusive interview with AIM, Snowflake head of AI Baris Gultekin said that he had worked with Ramaswamy for over 20 years at Google, calling him an incredible leader. “Sridhar brings incredible depth in AI as well as data systems. He has managed super large-scale data systems and AI systems at Google,” Gultekin said.
Gultekin further said that Snowflake is developing LLMs at a very affordable price, prioritising the security of their customers’ data. “Despite using a 17x less compute budget, Arctic is on par with Llama 3 70B in language understanding and reasoning while surpassing enterprise metrics,” said Gultekin.
The Microsoft Fabric and NVIDIA Spread
In addition, Microsoft announced an expanded partnership with Snowflake, aiming to deliver a seamless data experience for customers. As part of this, Microsoft Fabric‘s OneLake will now support Apache Iceberg and facilitate bi-directional data access between Snowflake and Fabric.
OneLake, a unified, SaaS-based open data foundation, was launched by Microsoft with the introduction of Fabric. The foundation underscores the company’s commitment to open standards. The support for Iceberg, alongside Delta Lake in Microsoft Fabric OneLake, further enhances this commitment.
In essence, Snowflake can store data in Iceberg’s format in OneLake. Data written by either Snowflake or Fabric will be accessible in both Iceberg and Delta Lake formats through XTable translation in OneLake. Snowflake can read any Fabric data artefact in OneLake, whether stored physically or virtually, through shortcuts.
And that’s not all.
In a recent interview, Ramaswamy revealed that the cloud data company plans to deepen its collaboration with AI powerhouse NVIDIA. “We collaborated with NVIDIA on a number of fronts – our foundation model Arctic was, unsurprisingly, done on top of NVIDIA chips. There’s a lot to come, and Jensen’s, of course, a visionary when it comes to AI,” Ramaswamy said.
Snowflake is expected to make a lot more announcements at its Data Cloud Summit this June. As Ramaswamy said, “Our product pipeline, especially in AI, has been in overdrive. The era of enterprise AI is here, right here at Snowflake.”