The heart and driving force of every AI-based technology are special chips whose computing power is capable of handling such advanced tools and models. The better and more efficient such a chip is, the better such AI works and is able to achieve more satisfactory synthetic and operational results. This is quite a simple and rather logical principle.
Currently, the leader in this segment is, of course, NVIDIA, a company until recently associated mainly with the production of graphics chips for computers and laptops. On the wave of the AI craze, this company has grown to gigantic proportions, leaving its competition far behind. For this reason, the “Greens” supply almost the entire AI market with their chips, leaving “crumbs” for their competitors.
However, the latest reports suggest that OpenAI may decide to create its own technology in this segment. An aspiring start-up not only has the ambition to do something incredibly daring, but has even found the right partners to do it.
OpenAI establishes cooperation with technological giants
As reported Reuters OpenAI is establishing increasingly closer cooperation with three industry giants – TSMC, Broadcom and AMD. As a result, an original, independently developed chip for artificial intelligence is to be created. These are very ambitious plans that may significantly change the way the development of systems for driving language models is currently perceived.
According to information, the decision made by the OpenAI authorities is to be a strategic step towards becoming independent from NVIDIA and significantly reducing costs.. An additional aspect here is the desire to achieve even better performance. These plans clearly show that the company is trying to meet increasingly complex plans and stay ahead in terms of innovation.
The issue of the supply of such systems and the above-mentioned dependency also seems to be important in this situation. NVIDIA, as the undisputed leader in the production of AI driving systems, serves virtually the entire market, which means that even such well-known and large companies as OpenAI must take into account extended waiting times for the ordered systems.
Will three technological giants be a savior for OpenAI?
The three specific entities mentioned earlier were not chosen by Altman and company completely at random. We are talking about technological giants that are supposed to provide not only knowledge and experience, but also the certainty that it will be possible to meet the company’s technological needs.
TSMC, the world leader in semiconductor production, is to provide processing capacity that will protect OpenAI against potential stagnation. Broadcom will be responsible for designing chips and developing them in accordance with the company’s needs. Thanks to this, the creators of ChatGPT will be able to count on relatively constant access to systems that will power their technology in a very specific way and at a level that no other manufacturer will be able to achieve.
A somewhat surprising addition to this rate is, of course, AMD. NVIDIA’s rival in the field of graphics chip production was included in this committee for a reason. Open AI plans to use MI300X chips via the Azure platform from Microsoft. In this way, Advanced Micro Devices hopes to capture at least part of the market and increase revenues to USD 4.5 billion by the end of 2024.
Ambitious plans and cautious moves
Whichever way you look at it, the intentions of Altman and company seem extremely intriguing. From the perspective of market observers, a small revolution can be expected if the first proprietary chips actually appear around 2026. Such a break from NVIDIA’s shackles may give other AI companies a signal that it is worth looking for a technological partner elsewhere.
With all this, there is also a business perspective. Creating your own chips is a huge saving, and the appearance of a new player on the market will increase competitiveness and make prices more attractive. This may be a chance for completely new development opportunities for many companies (which cannot boast of such media recognition as the main character of this text).
On the other hand, OpenAI must maintain at least good relations with NVIDIA for some time. After all, disrupting any relationship with this particular entity could have disastrous consequences for the company. Nevertheless, observing Altman’s further ambitions may be extremely interesting.