Collaboration for business AI solutions: Dell AI Factory with NVIDIA

What is Hardwarewartung 24 thinking?

Dell and NVIDIA are undoubtedly leading hardware manufacturers that develop products that are popular in the data centre environment. However, we at Hardwarerwartung 24 doubt that many customers trust a pure hardware manufacturer to have in-depth knowledge in complex business processes.

So why should a customer choose Dell when it comes to optimising business processes with AI? Despite Dell and NVIDIA’s seemingly comprehensive end-to-end AI solution, these products may not be flexible enough to fulfil specific customer needs or industry-specific requirements. AI is mainly intended to enable the automation of complex and individual customer processes, an area in which neither Dell nor NVIDIA offer specialised expertise.

Instead of in-depth process knowledge, typical for a hardware manufacturer, a platform is presented that is only relevant for a very small group of customers.

For whom is the Dell AI Factory with NVIDIA actually useful?

Hyperscalers:

Companies such as Amazon, Microsoft and Google are more likely to build their infrastructures on self-built servers instead of relying on expensive hardware from Dell.

The science community:

For scientific applications, the hardware, especially the NVIDIA GB200 AI chips for training models, could be well suited. However, the high cost of branded products such as Dell could be an obstacle, as cheaper alternatives such as Supermicro are available.

Start-Ups:

The platform could be particularly useful for start-ups working on the development of large language models (LLM). Their venture capital funding would enable them to afford the higher costs. However, for start-ups that want to only operate AI applications, platforms that specialise in inference, such as Groq, are more effective and cheaper.

Large Enterprise:

Dell’s cooperation with NVIDIA appears to be focussed on large enterprises, offering high-end hardware and integrated solutions specifically tailored to their needs. Nevertheless, there are significant challenges that call into question whether such companies will actually benefit from having their own AI platform.

For this hypothesis to be true, we must assume that these companies will develop their own AI applications rather than simply utilising existing AI applications. An assumption that may not be accurate for the following reasons.

The development and implementation of AI applications requires highly skilled AI developers who are not only technically skilled but also have a deep understanding of innovative AI technologies. These talents, however, are difficult to attract for companies that are not traditionally based in the Technology or AI sector.

Large technology companies such as Google, Amazon, Facebook and Apple offer not only financially attractive packages, but also an environment that is geared towards innovation and technological development. These companies invest heavily in research and development and offer resources and a corporate culture that are tailored to the needs of AI professionals.

In contrast, traditional large companies such as VW, T-Systems or Exxon might offer significant financial resources, but they often lack the dynamic and innovative work environment that AI experts are looking for. Such organisations are often locked into their structures and processes, leaving less room for creative freedom and ground-breaking innovation – both of which are crucial factors in attracting leading AI talent.

In addition, setting up an AI infrastructure and developing AI applications in-house is an enormous challenge that requires extensive investment in technology and human resources development. This can be too much of a hurdle for many large companies that are primarily active in other business areas.

Ultimately, choosing to rely on a platform such as Dell and NVIDIA may not be the optimal solution for many of these companies, especially if they choose to use existing AI applications rather than develop them themselves. In this case, a hardware platform designed for inference would be better suitable for hosting the applications in-house.

Bottom line

This announcement appears to be an attempt by traditional hardware manufacturers to position themselves in the dynamic world of artificial intelligence. They are trying to reinforce the narrative that, despite the current hype around AI and the massive changes, they are not the losers, but are better positioned than the tech industry would suggest thanks to their valuable customer access.

However, we must also acknowledge that Dell, as one of the few OEM hardware manufacturers, has repeatedly managed to benefit from major technological changes that seemed to be to its disadvantage. A good example is the big cloud trend of the last ten years, which, contrary to initial assumptions, has done them more good than harm because they made the right strategic decisions.

We are excited to see where the journey will take us this time and will of course continue to offer maintenance services for Dell hardware. If it catches on, also for the new Dell AI Factory.