Snowflake has pledged $6 billion to AWS over a five-year period, with Graviton chips being a key focus.
The five-year agreement represents a commitment that is 2.4 times larger than Snowflake’s 2023 AWS contract and comes as shares rise 38% following a Q1 earnings surprise. The significance of the Graviton component is crucial from a strategic perspective.
Snowflake has entered into a five-year, $6 billion agreement with Amazon Web Services, which both companies are promoting as the most significant expansion of their 11-year partnership to date.
The deal, revealed on Tuesday, includes commitments to utilize AWS Graviton, the cloud provider’s custom Arm-based CPU series, and to create stronger product integrations for what Snowflake refers to as “agentic enterprise” workloads.
The increase in Snowflake’s AWS investment highlights the scale of the announcement. The company pledged $1.2 billion during its 2020 IPO, which grew to $2.5 billion in a renewal in 2023. The new $6 billion commitment is approximately five times the amount from 2020 and 2.4 times the 2023 figure.
This expansion aligns with Snowflake’s growth: the company reported Q1 fiscal-2027 earnings on Wednesday that significantly exceeded estimates, resulting in a 38% jump in shares tied to both the earnings report and the AWS deal.
The Graviton aspect deserves special attention. AWS Graviton, currently in its fourth generation, represents Amazon’s proprietary Arm-server processor line intended to replace x86 chips from Intel and AMD in AWS data centers, offering enhanced price-performance.
Snowflake’s commitment to running its data-cloud workloads on Graviton at a large scale is a substantial endorsement of the Arm-server concept, which has been quietly transforming cloud-infrastructure economics over the past five years.
This also serves as a relevant data point for Amazon, particularly in light of recent news that ByteDance is developing its own Arm and RISC-V CPUs to alleviate pricing pressures from Intel and AMD.
The shift towards bespoke Arm-server silicon led by hyperscalers has clearly become a key narrative in data-center CPUs.
Specifically for AWS, the Snowflake deal is part of a series of major commitments to AI infrastructure. Anthropic has pledged significant AWS spending over multiple years; earlier this year, OpenAI signed a substantial agreement with AWS that positions it as a competitor to Microsoft Azure; and Meta has been noticeably increasing its AWS resources for inference workloads. Snowflake is the latest addition to this lineup and holds the title for the largest non-foundation-model commitment.
AWS’s ability to accommodate an additional $6 billion of demand over five years amidst an already high-demand data-center pipeline indicates how quickly Amazon is scaling up its capacity.
The strategic backdrop for Snowflake lies in its focus on the agentic-AI narrative, which the company has been actively endorsing.
Snowflake’s argument, akin to that of other large enterprise data platform vendors today, is that AI agents will rely mainly on trusted and governed enterprise data already stored in customers’ cloud-warehouse environments rather than external training datasets.
Creating this future necessitates significantly more computational integration with the underlying cloud provider; it requires direct access to AWS’s native AI tools (such as Bedrock, SageMaker, and the Q assistant) and greater synergy in marketplace and go-to-market strategies.
An unresolved issue remains the comparison with Databricks. Snowflake’s most prominent competitor is closely tied to Azure through its 2023 Microsoft partnership and has been aggressively promoting its multi-cloud capabilities.
Snowflake’s substantial commitment to AWS, particularly with the explicit focus on Graviton, indicates a different strategic direction: to choose a larger hyperscaler partner, secure customer acquisition through the AWS Marketplace, and embrace the implicit single-cloud orientation that comes with that choice.
The effectiveness of this strategy compared to Databricks’s diversified approach will unfold over the coming years.
Sales through the AWS Marketplace for Snowflake doubled year-over-year to $2 billion in 2025, suggesting that the integration strategy is already yielding commercial results.
Neither company has specified which generation of Graviton Snowflake is committing to. Snowflake's CEO, Sridhar Ramaswamy, mentioned that more details will be shared at the upcoming AWS re:Invent conference later this year.
Other articles
A Google engineer utilized internal search data to place a $2.7 million bet on Polymarket. He has been charged by the DOJ.
The Department of Justice has accused Google information-security engineer Michele Spagnuolo of utilizing internal search-trend data to earn $1.2 million on Polymarket’s Year-in-Search market.
The founder of JD.com promises to safeguard 900,000 jobs from AI, yet his approach to warehousing suggests a different direction.
Liu Qiangdong, the founder of JD.com, promised in an internal speech to safeguard the company’s 900,000 employees from being replaced by AI. This commitment feels somewhat contradictory given JD's significant reliance on automation and his own conceptualization of the ‘unmanned era.’
Orbital Industries secures $50 million in Series B funding for AI-designed data center equipment.
Orbital Industries, previously known as Orbital Materials, has secured $50 million in a Series B funding round, led by Plural, with participation from Nvidia’s NVentures. The funding is intended for their AI-designed, PFAS-free GPU cooling fluid and modular data center infrastructure.
Gold-Backed Stablecoins Aim to Make Cryptocurrency More Relatable
New digital assets are prompting well-known concerns regarding trust, regulation, and transparency of reserves. Gold has historically been linked with stability in volatile market conditions. Stablecoins were developed to mitigate the volatility associated with cryptocurrency assets. An increasing number of initiatives are now merging these concepts by establishing digital tokens that are connected to physical gold reserves. [...]
The Pentagon has awarded Dell a $9.7 billion contract to unify Microsoft licenses throughout the military.
The Pentagon has granted Dell a five-year contract worth $9.7 billion to unify Microsoft software licensing for the US military and intelligence agencies, which is expected to result in annual savings of $422 million.
ByteDance is developing its own CPUs based on Arm and RISC-V architectures to support its AI infrastructure.
ByteDance is creating its own data center CPUs through parallel efforts on both Arm and RISC-V paths, as Intel and AMD raise prices by 10-35% each quarter and US export regulations become stricter.
Snowflake has pledged $6 billion to AWS over a five-year period, with Graviton chips being a key focus.
Snowflake has pledged $6 billion to AWS over a five-year period, which is 2.4 times greater than its agreement in 2023, focusing on AWS Graviton chips. Following this announcement and a better-than-expected Q1 earnings report, shares surged by 38%.