Snowflake leans into AI
What happened
Snowflake reported a strong Q4 driven by AI features and is marketing its AI Data Cloud with partners that include AWS and NVIDIA (finance.yahoo.com) and (startuphub.ai). The positioning makes Snowflake customers natural candidates for GPU‑accelerated inference and model serving integrations.
Why it matters
Snowflake reported fourth-quarter product revenue of $1.23 billion, a 30% year‑over‑year increase, with total fourth‑quarter revenue of $1.28 billion and remaining performance obligations of $9.77 billion. (snowflake.com) The company said more than 9,100 accounts are using Snowflake AI features, and Snowflake Intelligence reached roughly 2,500 accounts in three months, signaling broad adoption inside its installed base. (snowflake.com) Snowflake framed the move as a unified enterprise AI lifecycle with explicit integrations to AWS and NVIDIA Blackwell GPUs in its messaging about the AI Data Cloud. (snowflake.com) AWS’s new EC2 G7e instances, which Snowflake is pointing to as a target for GPU‑accelerated inference, support up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB per GPU and up to 1,600 Gbps of networking bandwidth. (aws.amazon.com) Snowflake also highlighted deep commercial ties to AWS — it said it exceeded $2 billion in AWS Marketplace transactions in 2025 and won multiple AWS partner awards, strengthening the go‑to‑market channel for joint AI offerings. (snowflake.com) This push builds on earlier technical work: Snowflake expanded collaboration with NVIDIA in 2024 to run NVIDIA NeMo and TensorRT in the Data Cloud and to support NVIDIA models via Snowflake Cortex for LLM and inference use cases. (businesswire.com)
Key numbers
- Snowflake reported a strong Q4 driven by AI features and is marketing its AI Data Cloud with partners that include AWS and NVIDIA (finance.yahoo.com) and (startuphub.ai).
- Snowflake reported fourth-quarter product revenue of $1.23 billion, a 30% year‑over‑year increase, with total fourth‑quarter revenue of $1.28 billion and remaining performance obligations of $9.77 billion.
- (snowflake.com) The company said more than 9,100 accounts are using Snowflake AI features, and Snowflake Intelligence reached roughly 2,500 accounts in three months, signaling broad adoption inside its installed base.
- (snowflake.com) AWS’s new EC2 G7e instances, which Snowflake is pointing to as a target for GPU‑accelerated inference, support up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB per GPU and up to 1,600 Gbps of networking bandwidth.
What happens next
- (snowflake.com) AWS’s new EC2 G7e instances, which Snowflake is pointing to as a target for GPU‑accelerated inference, support up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB per GPU and up to 1,600 Gbps of networking bandwidth.
Quick answers
What happened in Snowflake leans into AI?
Snowflake reported a strong Q4 driven by AI features and is marketing its AI Data Cloud with partners that include AWS and NVIDIA (finance.yahoo.com) and (startuphub.ai). The positioning makes Snowflake customers natural candidates for GPU‑accelerated inference and model serving integrations.
Why does Snowflake leans into AI matter?
Snowflake reported fourth-quarter product revenue of $1.23 billion, a 30% year‑over‑year increase, with total fourth‑quarter revenue of $1.28 billion and remaining performance obligations of $9.77 billion. (snowflake.com) The company said more than 9,100 accounts are using Snowflake AI features, and Snowflake Intelligence reached roughly 2,500 accounts in three months, signaling broad adoption inside its installed base. (snowflake.com) Snowflake framed the move as a unified enterprise AI lifecycle with explicit integrations to AWS and NVIDIA Blackwell GPUs in its messaging about the AI Data Cloud. (snowflake.com) AWS’s new EC2 G7e instances, which Snowflake is pointing to as a target for GPU‑accelerated inference, support up to eight NVIDIA RTX PRO 6000 Blackwell Server Edition GPUs with 96 GB per GPU and up to 1,600 Gbps of networking bandwidth. (aws.amazon.com) Snowflake also highlighted deep commercial ties to AWS — it said it exceeded $2 billion in AWS Marketplace transactions in 2025 and won multiple AWS partner awards, strengthening the go‑to‑market channel for joint AI offerings. (snowflake.com) This push builds on earlier technical work: Snowflake expanded collaboration with NVIDIA in 2024 to run NVIDIA NeMo and TensorRT in the Data Cloud and to support NVIDIA models via Snowflake Cortex for LLM and inference use cases. (businesswire.com)
