HP acquires supercomputer maker Cray, why is there a 27% premium?

In most people's minds, HPC (High Performance Computing) is a highly professional and relatively quiet field. People usually build supercomputers to hit the Top 500 list, and then package and sell these behemoths to universities and research institutions.

This was true for a long time. However, with the development of big data, the geometric growth of data volume has greatly expanded the application space of HPC. In this process, the liquidity of the HPC market has also gradually increased. A recent acquisition in the industry is a good example of this.

In late May, Hewlett Packard Enterprise (HPE) bought supercomputer maker Cray for $1.3 billion, or $27 per share, a 27% premium over Cray's average price over the month.

Let’s start with this acquisition to get a glimpse of the changes in the HPC market.

If you can't be stronger, buy the stronger one.

Both parties in this acquisition are names worth remembering in the HPC market. Currently, the top five supercomputer suppliers in the world are Lenovo, Inspur, Sugon, Cray and HPE.

HPE is Hewlett Packard Enterprise, one of the two companies spun off from HP in 2014. HPE focuses on hardware and service businesses in the enterprise market, including the construction of supercomputers. Cray is also an HPC supplier. In the list of the world's top 500 supercomputers, the top ten supercomputers are all built by Cray.

But the difference between the two is that Cray is much more focused on the HPC business than HPE. In most cases, HPE's main service targets are still the computing needs of enterprise customers, not just HPC. Although HPE is one of the top five supercomputer suppliers, it is not among the top ten. Its share in the TOP500 is only 9%, and it focuses more on the commercial HPC field.

Cray is very focused on building supercomputers with extremely high computing power for use in government and academic research institutions. For example, Cray is currently working with AMD to build the world's first ExaScale-class supercomputer for Oak Ridge National Laboratory with the support of the US government. This contract is worth $600 million. Last year, Japan also approached Cray to collaborate on building a supercomputer for nuclear fusion research.

However, Cray's profitability is not ideal because it focuses on serving orders from governments and research institutions, which are expensive, have low sales volumes, and long sales cycles. According to the financial report, Cray's revenue in the first quarter of 2019 was US$72 million, down from US$80 million in the same period last year, and its losses further expanded to US$29 million. In other words, Cray's business model is closer to Parisian handmade couture, while HPE is a global fast-selling brand.

Under such severe revenue situation, it is not surprising that Cray was sold to HPE, which has a richer revenue structure.

Using 30% premium for tens of millions of losses:

Business logic under the quintet

HPE's acquisition of Cray seems to have a simple business logic - HPE focuses on the commercial HPC market, and Cray specializes in government and research institution orders, so the two parties can complement each other very well. The acquisition of a loss-making heavy business model company at a premium of nearly 30% per share is probably more than just business complementarity.

HPE's acquisition of Cray is naturally based on its belief that it can copy and paste Cray's more powerful technical strength in HPC into its own mid-range enterprise market, just like a luxury goods group acquiring a handicraft workshop, with the ultimate goal of packaging IP and trademarks into products on the assembly line.

HPE's move is believed to be influenced by the "quintet" in the HPC market.

The so-called quintet in the HPC market is a combination of AI-HPC-IOT-big data-5G. This model, which seems to pile up all the popular technologies together, is not without reason. With the mutual drive between AI technology and IoT applications, the demand for data volume and computing will only continue to soar, and after 5G greatly widens the data transmission pipeline, it also provides room for further expansion of data storage. In short, in the long run, the demand for computing power in the industrial and commercial fields is bullish in the long run, which also means that the HPC capabilities currently used in basic scientific research, earthquake simulation and other fields will step out of the ivory tower and enter the commercial market.

Among the quintet, AI technology, which is currently the most widely used, has played an absolute leveraging force. AI's demand for computing has made NVIDIA successful, and AI's demand for computing is likely to be unlimited. At present, many companies in the HPC industry have sensed this change and have made relevant arrangements or even transformations. For example, Lenovo proposed the concept of "intelligent supercomputing", Inspur also emphasized the support for AI scenario applications in its new product i48, and even Cray has established a special AI department. In recent HPC-related conferences such as ASC and ISC, the emergence of AI has also become inevitable.

HPE, after a long period of dormancy in the commercial HPC market, was the first to sense the demand for more efficient HPC among enterprise customers under the influence of Quintet, and therefore launched this acquisition.

After breaking the theory that supercomputing is useless

In it, we see the changing landscape of the HPC market under the quintet.

AI's demand for computing power has overturned the widely circulated "supercomputer uselessness theory" in one fell swoop. For a long time, HPC, especially the supercomputers ranked high in the Top 500 list, were relatively far from commercial applications, and most of them remained at the government and academic research level. It is easy to question what the role of HPC is? Has the competition for supercomputer rankings become a strategic reserve at the political level? With the Quintet pushing computing power demand to HPC from an industrial perspective, the difference between "useless" and "useful" has begun to reverse.

This also brings up the next question for the HPC market: self-driven and externally driven changes in production models.

Take Cray as an example. Its production and manufacturing of supercomputers are basically self-driven. It continuously breaks through through its own technological advancement, or cooperates with research institutions and governments to customize production according to their requirements. It can be said that such a production model is somewhat alienated from the real industrial market.

When HPC enters the industrial market side by side with AI, 5G, IoT, big data and other technologies, it must accept external drivers from the industrial market and understand the general needs of the industry. This transformation is not an easy task for HPC, especially since industrial demand has a strong dynamic feature, and HPC's high unit price and long sales cycle model are heavy.

The change in the demand for HPC in the field of autonomous driving is a good example. When autonomous driving first came into contact with HPC, most of them required densely built GPU computing nodes to train the deep learning models that constitute autonomous driving. But soon this demand changed, and autonomous driving customers began to focus on the storage needs of massive data. The reason is that with the development of data supervision laws, the compliance storage of data has become increasingly important.

To capture this dynamic in time, HPC companies are bound to have a deep tracking and attention to the industry, especially the other four partners in the quintet. Therefore, we can see that HPC companies such as Cray, which focus on the university and government markets, are splitting and merging with companies with more industry experience such as HPE. In 2016, HPE also acquired HPC manufacturer SGI. It is not difficult to imagine that there will be more and more mergers or acquisitions in the HPC industry in the future.

The merger and acquisition of the No. 4 and No. 5 figures is a big deal in any field. The butterfly's wings have been flapping for a long time, and now we are facing the waves on the ocean. The quintet is playing one by one. In the process of new technologies coming to the industry in groups, internal boiling and digestion are also indispensable.