There is no shortage of demand in AI as enterprises see return on investments by increasing productivity, reducing cost and automating more things even amid tough macroeconomic conditions, said Baris Gultekin, head of AI at Snowflake.
This also comes at the back increasing return on investments their customers are seeing. According to a recent report from Snowflake, for every dollar spent on AI, there is a 45% return on the investment.
“We are seeing a lot of productivity gains. Ability to do things that used to take a long time is now possible. Certain things that used to require a lot of expertise are now easily accessible to others. So the ROI is clearly measured by our customers. Everyone is very cost conscious, and then they are very clearly seeing the returns on their investments,” Gultekin said.
India market and services sector
When it comes to the Indian market, Gultekin said while the firm does not have specific India-focused products, multimodal and multilingual data benefits their products as well, he said. The company is also doing proof of concepts (POCs) with customers and working with them to build their applications to improve the return on investment on AI.
Snowflake is also partnering with companies in the services industry. “For instance, we have our own systems integrator partners that we work very closely with. They’re all building solutions for joint customers. Ultimately, AI is moving so fast, and everyone is trying to run as fast as possible. Partnering with others who can bring expertise to a wide range of customers, is something we’d like to do, and we do,” he said.
Trends
Financial services are one the biggest customers for Snowflake, where they analyse large amounts of data. The company is also seeing a lot of demand from insurance and healthcare, where significant paperwork is involved and can be analysed to bring in efficiency. “We are seeing demand in manufacturing and retail across the board. This is actually interesting because it’s not really concentrated on a single industry,” Gultekin said.
Challenges
There are a few challenges for the enterprise sector. “Many companies are building products, and they need to make sure that they’re able to evaluate them and observe them before they can get them ready for large-scale production,” Gultekin said.
Another big challenge is data. “AI is data hungry. Our customers need to be able to break down these data silos, to make sure that they’re able to govern that access,” Gultekin noted.
Though the costs are coming down, GPUs are expensive, as companies host LLM models. “We are optimising our usage and utilisation of these GPUs. Couple of weeks ago, we released an optimised version of Llama 3.1, where we were able to reduce the costs by 70% or so. We reflected that reduction back to the customer. So being able to take advantage of the open-source ecosystem and optimise how these models are used, allows us to reduce the cost to us, and then reflect the benefits back to the customer,” Gultekin said.
This also comes at the back increasing return on investments their customers are seeing. According to a recent report from Snowflake, for every dollar spent on AI, there is a 45% return on the investment.
“We are seeing a lot of productivity gains. Ability to do things that used to take a long time is now possible. Certain things that used to require a lot of expertise are now easily accessible to others. So the ROI is clearly measured by our customers. Everyone is very cost conscious, and then they are very clearly seeing the returns on their investments,” Gultekin said.
India market and services sector
When it comes to the Indian market, Gultekin said while the firm does not have specific India-focused products, multimodal and multilingual data benefits their products as well, he said. The company is also doing proof of concepts (POCs) with customers and working with them to build their applications to improve the return on investment on AI.
Snowflake is also partnering with companies in the services industry. “For instance, we have our own systems integrator partners that we work very closely with. They’re all building solutions for joint customers. Ultimately, AI is moving so fast, and everyone is trying to run as fast as possible. Partnering with others who can bring expertise to a wide range of customers, is something we’d like to do, and we do,” he said.
Trends
Financial services are one the biggest customers for Snowflake, where they analyse large amounts of data. The company is also seeing a lot of demand from insurance and healthcare, where significant paperwork is involved and can be analysed to bring in efficiency. “We are seeing demand in manufacturing and retail across the board. This is actually interesting because it’s not really concentrated on a single industry,” Gultekin said.
Challenges
There are a few challenges for the enterprise sector. “Many companies are building products, and they need to make sure that they’re able to evaluate them and observe them before they can get them ready for large-scale production,” Gultekin said.
Another big challenge is data. “AI is data hungry. Our customers need to be able to break down these data silos, to make sure that they’re able to govern that access,” Gultekin noted.
Though the costs are coming down, GPUs are expensive, as companies host LLM models. “We are optimising our usage and utilisation of these GPUs. Couple of weeks ago, we released an optimised version of Llama 3.1, where we were able to reduce the costs by 70% or so. We reflected that reduction back to the customer. So being able to take advantage of the open-source ecosystem and optimise how these models are used, allows us to reduce the cost to us, and then reflect the benefits back to the customer,” Gultekin said.
