Artificial Intelligence, Digital Tokens, and an Emerging Crisis at Pootlepress
anthropic chips nvidia openai
| Source: Mastodon | Original article
AI costs may surge as companies face expensive chip bills.
As we reported on April 26, the cost of AI has been a topic of discussion, with Elon Musk dropping fraud claims against OpenAI and Altman ahead of trial. Now, Jamie Marsfield's recent take on the cost of AI highlights that the technology is not free, but rather a costly endeavor with expensive chips running in the background. Marsfield notes that if OpenAI and Anthropic were to charge users the real cost of usage, the "magic" of AI might start to lose its appeal.
This matters because the cost of AI is closely tied to the processing of tokens, which are the building blocks of AI models. As NVIDIA's blog explains, AI models process tokens to learn relationships and unlock capabilities like prediction and generation. The faster tokens can be processed, the faster models can learn and respond. However, this processing power comes at a cost, with companies like DeepSeek billing based on the total number of input and output tokens.
What to watch next is how companies like OpenAI and Anthropic will balance the cost of AI with the need to make their services appealing to users. As the demand for AI continues to grow, the question of who will bear the cost of these expensive chips and token processing will become increasingly important. Will users be willing to pay the true cost of AI, or will companies find ways to subsidize or reduce these costs? The answer to this question will have significant implications for the future of the AI industry.
Sources
Back to AIPULSEN