The global shortage of expensive computer hardware is shaking up the artificial intelligence community. To remedy this, Open AI, the company behind Chatbot ChatGPT is seriously considering the possibility of manufacturing chips in-house. According to a Reuters report published on October 5, Open AI has even evaluated an undisclosed company as a potential acquisition target. This is to support its ambitions in the production of chips dedicated to artificial intelligence.
Limits imposed by the GPU shortage
In a blog post by Raza Habib, OpenAI founder and CEO Sam Altman expressed his frustration at the shortage of graphics processing units (GPUs). According to him and his developers, this situation is slowing down his company's progress. It also hinders short-term projects. "A recurring theme throughout the discussion was that OpenAI is currently facing significant limitations due to GPU shortages.
Why would OpenAI benefit from in-house chip production?
- Independence: having its own chips would mean OpenAI would no longer have to put up with delays due to the global shortage of computer hardware.
- Optimization: by designing and manufacturing chips specially adapted to their needs, OpenAI could considerably improve the performance of its products like ChatGPT.
- Cost control: by internalizing production, OpenAI could reduce the costs generated by intermediaries, and potentially lower the price of its offers.
- Innovation: in-house production also enables us to innovate more rapidly and adapt to changing market needs.
OpenAI follows in the footsteps of Google and Amazon
If OpenAI decides to go ahead with its in-house chip manufacturing project, it will join a select group of technology industry giants such as Google and Amazon. The latter have already taken the step towards producing specialized chips in-house. Here are just a few examples:
Google and its Tensor Processing Units (TPU)
In 2015, Google announced its first ASICs, or tensor processing units (TPUs). Google is creating solutions dedicated to machine learning. Their aim is to accelerate Google's own neural network model operations. Today, the third generation of these chips is in use.
Amazon and its Graviton
For its part, Amazon introduced its first ARM Graviton chip in 2018, intended for its Amazon Web Services (AWS) cloud services. The move comes as part of the e-commerce giant's efforts to reduce its dependence on traditional chip manufacturers.
What impact will this have on the artificial intelligence processor market?
OpenAI's in-house chip production would have an indirect impact on the market for artificial intelligence processors. It could :
- Encourage other companies specializing in artificial intelligence to opt for in-house manufacturing. As a result, demand from traditional chipmakers such as NVIDIA or AMD will decline.
- Increase competition between the various players in the sector. It could also spur innovation in the design of new chip architectures dedicated to artificial intelligence.
- Promote diversification of market offerings. New hardware solutions adapted to specific needs in the field of machine learning and artificial intelligence will emerge.
Conclusion
ChatGPT's creator, OpenAI, is seriously considering the possibility of producing in-house chips dedicated to artificial intelligence. This is a strategic move that would see the company join the ranks of major market players such as Google and Amazon. Stay tuned!
