Since the launch of ChatGPTthe Artificial intelligence has become a priority for many companies, companies that increasingly allocate an increasing number of resources to train their own AI and for which a large amount of storage space is necessary, storage space that, if we continue at the current pace, It will soon be too small.
That is what can be deduced from the survey that the manufacturer of storage units Seagate commissioned Recon Analytics. According to this report, storage requirements related to Artificial Intelligence will double in the next 3 years.
More AI means more storage space
According to this study, carried out in companies around the world, 72% of the companies surveyed affirm that they already use, in one way or another, Artificial Intelligence. The remaining 28% affirm that they are working on adopting it in the next 3 years.
Currently, 65% of companies that use the cloud for data management generated by AI, and it is estimated that it will grow up to 69% in the next 3 years. Of those using more than 100 PB of storage, 87% use cloud or physical storage disks to create training checkpoints.
Creating control points is fundamental in the training of any Artificial Intelligence, and as the language models and the data necessary to train them grow, the storage capacities, not only of the servers but also of the companies, are continually growing and will continue to do so. in the next 3 years.
Checkpoints, which take up a large amount of storage space, are not removed quickly, but are kept between 6 and 12 months by 29% of respondents and more than a year by 32% of respondents. %. This is because these training data allow us to improve the precision of the AI models.
To store all this information, 61% of companies have chosen to use information systems. cloud storage to satisfy your storage demand without worrying. 56% choose to use an advanced data management system that combines cloud and local storage while 55% have been forced to upgrade the servers they use locally.
To try to reduce the space they occupy, 49% use different compression techniques to reduce the volume of data generated by AI as they are scalable and efficient solutions that adapt to the exponential growth of storage.
Just as any user should make it a rule to make a backup of all their data, companies that work with AI, if they do, or at least 80% of those surveyed, especially those that work locally and not in the cloud where these platforms are in charge of replicating the data to avoid before any incident.
In the coming years, companies that still work locally are expected to move to the cloud, which means that large data centers will have a much greater storage need than today.
