Unlocking full capabilities across enterprise AI with a unified data platform

VAST Data recently announced a slew of groundbreaking announcements that promise to revolutionize the landscape of enterprise AI.

When it comes to developing and running AI programs, organizations want to have a seamless operation with minimal complexities. While most organizations would take the hybrid cloud approach in their data management, for developers, it meant having access to platforms that can support their data integration as well as generate the desired results within a controlled ecosystem.

To successfully develop and run AI use cases, programs need to run in a data ecosystem on the cloud that is capable of pushing the boundaries of data analytics but also keep to regulatory and compliance requirements. As businesses continue to have exponential data growth, their biggest challenge goes down to understanding their data and classifying them based on their importance.

Put simply, complex data infrastructure poses significant challenges for organizations to scale AI to effectively process and extracts insights from their massive datasets. As businesses look to deploy RAG and improve their focus on LLMs, the need to have a platform that can classify structured and unstructured data sets, while ensuring there are minimal complexities is imperative.

VAST Data unveiled the latest version for the VAST DATA Platform about a year ago. The VAST Data Platform 5.0 for AWS is already capable of lowering monthly cloud costs. But there are still more opportunities to improve data management.

According to Sunil Chavan, Vice President, Asia Pacific and Japan at VAST Data, there is a lot of data migration and data gathering process that businesses need to do to get into the AI bandwagon.

“If you take a standard enterprise or digital native company that has been around for about ten years, they will easily have about 30 to 40 petabytes of data running on different clouds. However, a lot of these cloud hyperscalers are not able to fulfil the current strategy desired. Hence, organizations are taking their data out and moving it into a new AI cloud and to train it into an AI learning model,” said Chavan.

Chavan pointed out that this is where the challenge begins for organizations. Classifying which data needs to be trained, which data needs to go towards inferencing and such. Typically, an organization would need to at three or four solutions or rely on a generative AI application to solve this.

This in turn will lead to extra costs for the organization. While there are solutions that are capable of doing this, Chavan highlighted that moving 10 or 20 petabytes of data to a single ETL (extract, transform and load) platform running on the cloud can easily lead to very high costs for the organization.

“There is also the GPU cost as most businesses don’t have GPUs and are just getting their data together at an astronomical fee. You’re taking that data out and moving it to a learning model, paying that cost, and pushing that data back again into the same process. The data ingestion cost that holds the entire cycle is really costing organizations a bomb,” explained Chavan.

Access to a unified platform

To deal with these challenges, VAST Data recently unveiled the VAST InsightEngine with NVIDIA. The solution is a world first in securely ingesting, processing and retrieving all enterprise data in real-time. It is also the first application workflow that runs on the VAST Data Platform.

Operating with NVIDIA NIM microservices, the platform embeds incoming data using advanced models powered by NVIDA accelerate computing. These vector and graph embeddings are then stored in the VAST DataBase within milliseconds after the data is captured. This ensures any new file, object, table or streaming data is instantly ready for advanced AI retrieval and inference options.

“We are seeing a lot of challenges in organizations, especially when they want to use the cloud, to bring the data together, prepare the data and get into the learning model process, followed by some inferencing on that. Right now, while cloud providers are focused on helping customers develop learning models and such, we are helping them push the boundaries by having a data pipeline from data ingestion to inferencing and then to a rag model on top of that. This helps enterprises learn and get more value out the cloud service providers as well,” said Chavan.

To streamline this, VAST Data has also launched Cosmos, an initiative designed to transform how organizations build and advance AI, by creating a comprehensive and supportive environment to nurture innovation, collaboration, growth, and economic prosperity.

Built by AI practitioners for AI practitioners, Cosmos basically aims to streamline AI adoption for its members by offering a comprehensive, interconnected ecosystem that facilitates conversation, use cases, and provides learning opportunities through labs, vendor showcases, and general AI research news.

“It’s not just about us. It's about connecting the dots and bringing people together so that enterprise benefit from a faster, developed enterprise AI ecosystem at a reasonable price,” mentioned Chavan.

And this is where it gets interesting as well. The cloud adoption in Asia is moving towards regulated sovereign cloud or a dedicated cloud for a specific use case. Chaven also highlighted that this is an increase in area of focus for cloud service providers in Asia.

“It is an area that is really growing and is a big focus for VAST Data because there are certain strategic advantages we bring for any sovereign data, sovereign AI cloud from our technology point of view and a lot of customers we have in that regard. We are talking to governments in the region about this” added Chavan.

With VAST Data also focusing on healthcare and education, the COSMOS platform as well as Insight Engine can help businesses in their AI journey as its focus on building the AI foundation required.

Building the partner ecosystem

Supporting the development and success for businesses in AI implementation plans is the role partners play. VAST Data is seeing a new partner ecosystem developing along with NVIDIA, as the company works very closely with the NVIDIA partner ecosystem.

At its recent summit, VAST DATA also announced a collaboration with Equinix to provide dedicated, secure AI Infrastructure Solution validated for NVIDIA DGX SuperPOD. This collaboration provides the infrastructure foundation of an AI development platform to help customers develop and implement generative AI applications by modernizing and leveraging enterprise data for AI workloads – without compromising on data security or infrastructure ease of use.

Separately, VAST Data also announced that it is expanding its partnership with Cisco to deliver End-to-End, Full Stack AI Infrastructure. Through this growing partnership, Cisco plans to offer the VAST Data Platform natively on select Cisco Unified Computing System (UCS) servers as an integrated solution via Cisco’s global sales team and network of channel partners around the world.

Looking at VAST Data’s plans for the region, Chavan stated that they will most likely be launching and running more projects on AI data centers and clod providers in the region for the next six to nine months.

“Our focus for the fiscal year of 2025 is to really get into the data insights and enterprise AI thought process with our partners and bring that AI closer to enterprise customers. We want to provide that ecosystem,” Chavan concluded.