AI Express, a secretive Silicon Valley artificial intelligence startup, has raised over $11 million in seed funding to bring the ability of AI to perhaps the most important challenge in enterprise computing today: reining in runaway cloud costs. The funding features a seed round led by Daniel Gross and Nat Friedman and a pre-seed round led by Matt Turck at FirstMark, with participation from industry leaders.
The company, which has emerged from obscurity today, has developed technology that uses advanced language models and machine learning to routinely optimize code and reduce cloud computing costs by as much as 80%. The first product focuses on optimizing SQL queries for Snowflakethe favored cloud data warehousing platform.
“The opportunity is just enormous,” said Ben Lerner, founder and CEO of Espresso AI, in an exclusive interview with VentureBeat. “Snowflake alone has annual sales of $2 billion. If you have a look at data warehousing broadly, it's definitely a whole bunch of hundreds of thousands of dollars in revenue for us and billions of dollars in potential savings for our customers.”
A cloud cost crisis is brewing
Moving to the cloud has been a double-edged sword for corporations. While cloud platforms offer unprecedented flexibility and scalability, they’ve also set recent standards recent challenges around cost control and visibility. Many businesses today are fighting unexpectedly high bills and have difficulty predicting and managing their expenses.
Data warehousing is a selected pain point. As corporations consolidate data silos and launch recent analytics and machine learning initiatives, data warehouses have change into a number of the largest consumers of cloud resources. However, optimizing these workloads for cost and performance is notoriously difficult.
“What we hear from users on a regular basis is that Snowflake is their second largest promotional item after AWS,” Lerner told VentureBeat. “And once you go to a Snowflake event, they really concentrate on two things: cost and performance.”
AI to the rescue
Espresso AI's solution is to harness the ability of enormous language models (LLMs), the underlying technology behind viral sensations like ChatGPT, to deal with the issue of code optimization. By training these models to deeply understand SQL queries and database architectures, Espresso AI has built a platform that may routinely refactor queries to make them more efficient.
Here's how it really works: Espresso AI integrates into an organization's existing Snowflake setup and constantly analyzes the queries executed against the information warehouse. Using a mixture of natural language processing, program synthesis, and reinforcement learning, it identifies optimization opportunities and quickly rewrites queries to enhance performance and minimize compute usage.
“The reason that is so powerful is because in lots of existing applications that you must have a human within the loop to examine the accuracy,” Lerner explained. “When you optimize code, you already know what you wish it to do – just faster. And so we will routinely check whether the optimized code is correct.”
Setup is designed to be easy and rise up and running in under 10 minutes by simply changing a single connection string. “It’s as easy as changing a URL,” Lerner said. “You point your BI and analytics tools to the Espresso endpoint as an alternative of on to Snowflake, and we handle the remainder.”
Ready for growth
Espresso AI has seen strong early traction with several enterprise customers using its platform to optimize production snowflake workloads. The company plans to make use of its funds to speed up product development and market launch.
While Snowflake is the initial focus, Espresso AI's technology is extensible to any SQL data warehouse. Support for platforms like Datastones is on the short-term roadmap. Longer term, the corporate plans to make use of its AI optimization engine to speed up computing performance across the whole stack, from data preprocessing to model training.
“It's hard to quantify what the world will appear to be when computers run 100 times faster,” Lerner said. “Everything will occur quickly. We will give you the chance to do more research and more machine learning. There are a variety of limitations in data processing today.”
Of course, achieving 100x speedup is simpler said than done. While Espresso AI has demonstrated impressive leads to early customer deployments, achieving performance improvements at scale would require significant research breakthroughs. The company also has to fend off competition from cloud providers themselves, who’re investing heavily in cost management and optimization features.
But if Espresso AI can achieve even a fraction of its founding vision, the impact might be profound. With corporations spending more $600 billion per 12 months for cloud and on-prem computingThe market opportunities for AI-supported efficiency improvements are enormous.
In a time of belt tightening and digital transformation, technologies that may deliver significant cost savings without sacrificing performance will find an enthusiastic audience amongst CIOs. By bringing the ability of AI to the unsexy but essential area of code optimization, Espresso AI could also be creating something truly transformative.
If a cup of coffee is the value to pay for holding cloud costs, expect so much more IT executives to line up for a sip of espresso AI.
