Weaviate—The AI Native Vector Database
About Weaviate
Weaviate is an Amsterdam-based software company that has created the popular open source Weaviate vector database and Weaviate Cloud. Weaviate makes it easy to store, search, and share data objects and vector embeddings from your favorite ML models, and scale seamlessly into billions of data objects.
Are you interested in getting started with Weaviate? Use this link.
Tell us your challenge. We're here to help.
Contact usStartup software company Weaviate turns to Google Cloud to build its cloud-native vector search business, using credits from the Google for Startups Cloud Program to get started.
Google Cloud results
- Uses Google Cloud for Startups credits to develop Weaviate
- Teams are now able to work collaboratively in different timezones, in ways they couldn't do previously
- Leverage the ease of use of Google Cloud to build complex, scalable systems with a global team
- Creating more efficient processes with autoscaling on Google Kubernetes Engine
Building a vector database from scratch on Google Cloud
Data is king for any business looking to stay ahead, but this is only true if that data can be analyzed to create actionable insights. With structured quantifiable data that is often stored on data warehouses and SQL databases, this is fairly easy to achieve. It is much harder to effectively analyze unstructured data, such as photographs, emails, audio recordings, technical manuals, and other complex business data.
Dutch software company Weaviate.io created Weaviate to solve this problem. An open-source, AI-native vector database, Weaviate uses machine-learning models to store and make sense of business data on a deeper level than existing databases can offer. This makes it easier for developers, from startups to enterprises, to create a new wave of AI applications ranging from custom-made search and recommendation systems to ChatGPT plugins.
"AI developers who work with data have a search problem that needs solving," explains Bob van Luijt, the co-founder and CEO of Weaviate. "When Google changed from PageRank to RankBrain, and started using machine learning to improve its search results for the first time, it was a big moment. For me, it was a huge spark of inspiration. I knew that I wanted to commoditize this and give this power to other businesses as well."
A dream powered by Google for Startups Cloud Program
Just like many startups, Weaviate began with an innovative idea and limited funds. Bob van Luijt had one advantage over others just starting out on pursuing a digital dream, however, he was already a Google Developer Expert. Not only did this give him the right skillset, it also helped him network.
"I met a lot of people at Google through the Google Developer Experts program, so when I embarked on this journey with Weaviate, they helped connect me with people who could help us get started," he says. "We were lucky enough to access the Google for Startups Cloud Program, which gave us credits to use against Google Cloud products. This was so helpful in getting us going."
Having previously worked with Google Cloud, building the Weaviate AI-native vector database on Google Cloud was the obvious choice for van Luijt, and the Google Cloud credits really helped him to get started. The credits offer eligible startup businesses up to $100,000 to spend on Google Cloud solutions, as well as access to dedicated tech and business support.
"We are a success story, thanks to Google Cloud credits," says van Luijt. "They are so important for new startups. We didn't have any funding in the beginning, we started out with nothing. These credits helped us to get started and helped us to prove our worth to investors. These credits helped us to get started and as of April 2023, less than two years later, the Weaviate open source database has been downloaded over 2 million times."
"We are a success story, thanks to Google Cloud credits. They are so important for new startups. We didn't have any funding in the beginning, we started out with nothing. These credits helped us to get started and helped us to prove our worth to investors."
—Bob van Luijt, CEO, WeaviateBuilding from the ground up on Google Cloud
Weaviate relies on two key Google Cloud products: Google Kubernetes Engine (GKE) and Compute Engine. As Weaviate runs on Kubernetes clusters, it made sense to use GKE because it makes it so much easier to deploy, scale and manage Kubernetes. "We depend on Kubernetes," says van Luijt. "I can't think of any other platform that's more efficient to run Kubernetes on than Google Cloud. The developer experience is so easy to use. From integrating our pipelines to setting up a demo to doing deep testing, everything just works really well. It is so important for us."
Ease of use is a continuing theme with the company's use of Compute Engine too. As with all machine learning models, Weaviate needs access to GPUs, and while GPU acceleration is key for this process, so is efficiency. By offering easy access to the GPUs within the company's Kubernetes clusters, Compute Engine makes it much easier to create more efficient processes through effective autoscaling. By ensuring that GPUs don't run when they are not in use, Weaviate can be more efficient with pricing for its customers.
"I can't think of any other platform that's more efficient to run Kubernetes on than Google Cloud. The developer experience is so easy to use. From integrating our pipelines to setting up a demo to doing deep testing, everything just works really well. It is so important for us."
—Bob van Luijt, CEO, WeaviateEasy collaboration across time zones
Google Workspace has also played a role in the company's success. As a fully remote workforce working in an asynchronous manner, Weaviate relies on colleagues being able to collaborate simultaneously wherever they are.
"Using solutions such as Sheets, Docs and Drive has been invaluable to us as a company, they allow us to work and stay connected in different timezones," says van Luijt. "Even when customers request information from us, we can just send them a Google Doc link. We work entirely in and from the cloud."
As far as the impact Google Cloud has had on Weaviate as a business, it's the backbone that the business is built upon; not just from a technology perspective, but a human one too.
"As a startup, if we didn't have these Google Cloud tools, what we do, and how we work would not be possible," says van Luijt.
"And, our relationship with our contacts at Google Cloud is invaluable too. Google Cloud has the attitude of 'if you are successful, we are successful.' That's refreshing. Google Cloud was welcoming from the beginning. It knows it's a journey to success and has always wanted to help us get there."
"Google has the attitude of 'if you are successful, we are successful.' That's refreshing. Google was welcoming from the beginning. It knows it's a journey to success and has always wanted to help us get there."
—Bob van Luijt, CEO, WeaviateExpanding potential with new capabilities
Looking to the future, van Luijt is excited about the possibilities of new cloud and AI capabilities from Google that will enable Weaviate to continue delivering innovative vector data management solutions to AI developers. "The great thing about Google Cloud is that as it evolves and gets better over time, it allows us to bring more exciting products to our users and customers too. It has enabled us to recently launch Weaviate as a fully managed cloud-native vector database cloud. That's really important for AI developers who want a quick on-ramp to vector search without having to host and configure database clusters by hand.
"There are a few other innovations we have coming soon, thanks to our relationship with Google Cloud. I need to keep that under the radar for now, but it's safe to say, the things we have in the pipeline with Google Cloud are super exciting indeed."
Tell us your challenge. We're here to help.
Contact us
About Weaviate
Weaviate is an Amsterdam-based software company that has created the popular open source Weaviate vector database and Weaviate Cloud. Weaviate makes it easy to store, search, and share data objects and vector embeddings from your favorite ML models, and scale seamlessly into billions of data objects.
Are you interested in getting started with Weaviate? Use this link.