JAKARTA - OpenAI is reportedly exploring the use of Google's AI chip, namely the Tensor Processing Units (TPUS), to test their AI models. This step surprised many parties, because so far OpenAI is known to rely heavily on NVIDIA's GPU.
While that doesn't mean OpenAI will completely leave NVIDIA, this decision is seen as an attempt to expand supply chains and reduce dependence on one hardware source.
NVIDIA's success in recent years cannot be separated from the surge in hardware needs for AI development. Their GPU has become a mainstay for many AI companies, including OpenAI. However, a report from The Information states that OpenAI is now starting to look at Google's chips as an alternative.
OpenAI's move to work with Google certainly feels odd, considering that both of them are direct competitors in the development of artificial intelligence. However, Google not only builds AI products, but also develops their own AI chips called TPU. Initially, these chips were only used internally, similar to Apple's series A and M chips that were exclusive to their own products.
Later, Google began to open access to TPUs for other companies such as Apple and Anthropic. Now, reports say that OpenAI may be the next party to take advantage of this technology.
It should be noted that this is the first time OpenAI has used non-NVIDIA chips to develop their AI. TPUS is known to be more efficient in terms of costs than GPUs, so it can be a strategic step to reduce OpenAI's operational expenses.
However, reports also reveal that OpenAI will likely not directly use Google's most advanced hardware, which is still being saved for Google's own internal use.
SEE ALSO:
Avoid Supply Chain Congestion
GPUs and TPUs are the main hardware in the development of artificial intelligence. GPUs are more flexible and can be used for various purposes, while TPUs are specially designed for machine calculations and learning tasks.
The full dependence on NVIDIA could be a separate obstacle for OpenAI, given the limited cycle of hardware development and stock availability. By considering other options such as TPU, OpenAI hopes to avoid potential barriers in the development of AI in the future.
This step does not mean that OpenAI no longer believes in NVIDIA, but rather in a diversification strategy to reduce risk and increase efficiency. Although neither Google nor OpenAI have made official comments, if this collaboration does happen, then this could be a new chapter in the dynamics of the global AI industry where cross-competition collaboration becomes a necessity for technological advances.
The English, Chinese, Japanese, Arabic, and French versions are automatically generated by the AI. So there may still be inaccuracies in translating, please always see Indonesian as our main language. (system supported by DigitalSiber.id)
