Voltron data cooperated with Accenture to solve one of the largest headaches in artificial intelligence

Since artificial intelligence pays an unprecedented demand for data processing, Mountain View Startup provides a solution to one of the least challenges discussed but most important: transferring and transferring huge data groups quickly enough to keep up with it.

Voltron dataThat announced a strategic partnership with Tone Today, a GPU analysis engine that can help institutions overcome the preparation of the bottleneck data that hinders artificial intelligence initiatives. The primary product of the company, TheusInstitutions enable the PEABYTE data processing using graphics processing units (GPU) instead of traditional computer processors (CPU).

“Everyone focuses on new cheerful things that you can touch and feel, but the basis of the data set below it will be a key,” said Michael Abbott, who leads banking and the practice of capital markets at Acceneture, in an exclusive interview with Venturebeat. “To make Amnesty International, you have to transfer data quickly and quickly you haven’t had to.”

Building for Amnesty International Tsunami: Why will you not interrupt traditional data processing?

The partnership comes as companies that rush to adopt artificial intelligence Data infrastructure Not equipped to deal with the size and speed of the required data. This challenge is expected to intensify when artificial intelligence agents become more prevalent in institutions operations.

“It is possible that the agents will write more SQL inquiries than humans on a very short horizon,” said Rodrigo Aramburo, CTO and co -founder of Voltron Data and co -founder. “If CIOS and CTOS already say it spends a lot on data and cloud infrastructure analyzes, and the demand is about to be higher, then we need a step at the bottom of the cost of running these queries.”

Unlike the traditional databases sellers who re -updated GPU support on the current systems, Voltron data built its engine from A to Z to accelerate the GPU. “What most companies did when they tried to do GPU is that they will embrace graphics processing units on a current system.” “By building from A to Z … We are able to get 10x, 20x and 100x depending on a specific work profile.”

From 1,400 servers to 14: The first adoption sees exciting results

Company jobs Theus As a complement to existing platforms such as Snowflake and Databrics, and take advantage of the APache Arrow frame for effective data. “It is truly accelerated for all these databases, instead of competition,” said Abbott. “He still uses the same SQL to get the same answer, but reaching there is much faster and faster in a parallel way.”

Early adoption focused on intensive data industries such as financial services, as cases of use include fraud, risk modeling and organizational compliance. A large retail company has reduced the number of servers from 1400 CPUs to 14 GPU servers only after thisus, according to Aramburu.

Since its launch Nvidia GTC Conference Last March, Voltron Data got about 14 clients from the institution, including two major government agencies. The company plans to issue a “drive test test”, which will allow potential customers to experience GPU on TERABYTE data collections.

Transforming the lack of graphics processing unit into an opportunity

Current Lack of graphics processing unit It was difficult and useful for the Voltron data. While new publishing operations face restrictions on devices, many institutions have an unstable infrastructure of GPU originally purchased for the burdens of artificial intelligence work that can be reused to process data during inactivity periods.

“We have actually seen a blessing in that there are many graphics processing units in the market that were not in the past,” adding that Theosus could work effectively on the oldest GPU generations that may be neglected.

Technology can be of special value for banks that deal with what Apot calls “besieged data” – closed information in formats such as PDF files and documents that may be valuable to tract artificial intelligence but is difficult to access and process widely. “I have seen some data that will appear to Colitron that you are 90 % more effective in transferring data using this technology from standard CPU,” said Abbott. “This is strength.”

Since institutions are struggling with data requirements from artificial intelligence, it is possible that solutions that can accelerate data processing and reduce infrastructure costs become more important. Voltron data can help access to more institutions facing these challenges, while giving Accessure customers access to technology that can significantly improve the performance and efficiency of artificial intelligence initiatives.

Daily visions about business use cases with VB daily

If you want to persuade your boss at work, you have covered VB Daily. We give you the internal journalistic precedence over what companies do with obstetric artificial intelligence, from organizational transformations to practical publishing operations, so that you can share visions of the maximum return on investment.

Read with us privacy policy

Thanks for subscribing. Check more VB news bulletins here.

An error occurred.