Where there is AI, there is data, and where there is data, there is data storage.

Schumpeter said that innovation is to "establish a new production function", that is, to form a "new combination" of production factors and production conditions that has never been seen before. From this perspective, the combination of data, a new production factor, and the production conditions of artificial intelligence technology is a complete innovation, and the intelligent changes it will bring will be so surging.

In 2023, with the rise of large AI models, we seem to have seen a passionate picture of the AI ​​era accelerating. According to the New Generation Artificial Intelligence Development Research Center of the Ministry of Science and Technology of China, the previously released "China Artificial Intelligence Large Model Map Research Report" shows that China has released 79 large models with a scale of more than 1 billion parameters, and the war of hundreds of models has become a foregone conclusion.

The popularity of large models has once again ignited the flame of the AI ​​​​era, but at the same time it has also brought us new revelations: data is the "fuel" of artificial intelligence. The intensity and heat of this flame in the AI ​​​​era completely depend on Whether the value of the data can be released.

In the AI ​​era, the logical relationship between data storage and artificial intelligence must also be a mutually reinforcing and spiral upward model. This also means that the development of the AI ​​era will also drive the development of the data storage industry.

01

The value enhancement of data to AI

As we all know, the significance of data to artificial intelligence is just like electric cars need batteries. If there is not enough storage data, the value that artificial intelligence can achieve will be extremely limited.

In the AI ​​industry, there has always been a consensus of "garbage in, garbage out". That is to say, if there is no high-quality data input, no matter how advanced the algorithm or how huge the calculation is, No effort can produce high-quality results. Therefore, the main reason that determines the height of AI intelligence depends on the quality of data.

Of course, in addition to the quality of the data, the magnitude of the data also determines the height of what a large AI model can achieve.

Because the expressive ability of a model based on small-scale data is limited by the data size, it can only perform coarse-grained simulation and prediction, which is no longer applicable in situations where accuracy requirements are relatively high. If you want to further improve the accuracy of the model, you need to use massive data to generate relevant models.

This shows that the scale of data also determines the value of AI intelligence. Therefore, both the quality of the data and the order of magnitude of the data illustrate the role of data in artificial intelligence, which has become increasingly prominent with the deepening of AI applications.

This is easy to understand. When AI systems have more and better quality data, they can better predict future trends and generate more value.

For example, Tesla uses massive amounts of data to train its powerful artificial intelligence driving model, bringing an extraordinary experience to global users; while the Internet platform uses a large amount of user data to analyze with artificial intelligence, which can Customizing digital advertising based on user portraits is expected to bring global digital advertising revenue of up to US$679.8 billion in 2023.

These cases all prove the importance of data in increasing the value of artificial intelligence and even business model innovation.

02

The AI ​​era drives the rapid development of data storage

This logic, if understood in reverse, is also true: the popularity of artificial intelligence has produced an even larger amount of data, which poses more challenges to the storage and processing of data

With the surge of global digitalization, data centers are being constructed at a geometric speed. A report by Schroders in 2023 shows that the power consumption of data centers will rapidly increase from 17 gigawatts in 2022 to 17 gigawatts in 2030. 35 GW, which means that the total number of data centers is expected to double in the next eight years.

Correspondingly, there is a surge in demand for data storage. "Fortune Business Insights" predicts that the global data storage market is expected to grow from US$247.32 billion in 2023 to US$777.98 billion in 2030. The market size is almost It will be tripled.

The comparison of these two data shows that the growth rate of data storage is much higher than the growth rate of the data center. We can also read two details from it:

First, precisely because of the arrival of the AI ​​era, new requirements have been put forward for new data centers: data storage capabilities have become the focus of construction; second, the demand for data storage expansion has become the main driving force for data center construction.

From this, it is not difficult for us to draw a new conclusion: the development of the AI ​​era will inevitably promote the rapid development of the data storage field. Core manufacturers in the data storage field will have a more promising future in the market, especially owners of core technologies represented by Seagate, who will capture the largest share of business growth amid the surge in market demand.

03

HDD’s unique position in the AI ​​wave

In fact, there has been a common misconception in the industry in the past few years: it is believed that HDDs will be completely replaced by SSDs. But in fact, data center cloud service providers require a large number of high-density and large-capacity HDDs for cloud storage. Realize that the emergence of this market demand has made the growth rate of data center HDD products no less than that of SSD.

In fact, whether it is the big data demand represented by HDD or the fast data demand represented by SSD, the scale of demand is constantly expanding. Especially with the advent of the AI ​​era, the demand for HDDs is increasing day by day.

The "Digital World—From Edge to Core" white paper sponsored by Seagate Technology and released by International Data Corporation (IDC) predicts that cloud data centers are becoming new enterprise data repositories. IDC predicts that by 2025, 49% of the world's stored data will be in public cloud environments. Because the artificial intelligence conversation mainly focuses on processors and cloud storage, while cloud storage relies more on HDD. Data created by artificial intelligence will require more HDDs to store in the future.

We can’t imagine that the hard drive industry has developed so fast in the past 45 years. In the 1980s, a 5.25-inch hard drive could only store 5 million bytes of data, and the latest technology in 2023 , Seagate has provided new products to some customers in July, with storage capacity of up to 30TB per block.

It is the leap in data storage technology that creates more imagination space for the industry. However, the demand for data storage due to the wave of artificial intelligence is growing at an order of magnitude. Perhaps the hard drives that are readily available today will become "scarce" products in the next few months.

Forbes predicted in a recent article that shipments of hard drive products will grow by 900% from 2020 to 2028. This means that if cloud service providers cannot purchase sufficient data storage space, they may even be unable to meet the needs of artificial intelligence growth.

Judging from the current market structure of mechanical hard drives, this is a market with high market concentration. Seagate, Western Digital and Toshiba dominate the world, with Seagate occupying the top spot. Their product innovation capabilities determine the development pace of the entire field of data storage to a certain extent, and will further influence the pace of upgrades and evolution in the AI ​​era.

Where there is artificial intelligence, there is data. Where there is data, there is data storage. As the AI ​​era is about to arrive, we also need to give data storage represented by mechanical hard drives a correct value positioning.

The above is the detailed content of Where there is AI, there is data, and where there is data, there is data storage.. For more information, please follow other related articles on the PHP Chinese website!