In the early days of artificial intelligence (AI), the trend of developing its own LLMs became a trend in the industry. This was followed by chatbots, writing bots, image-generating models, video-generating models, programming models and then world models for robotics. If all this is done, the next step is to develop its own chips with Anthropic also doing so in the near future according to a Reuters report.
Anthropic is evaluating building its own AI chips with no official commitment yet. At this point, they may continue to buy chips from existing suppliers. The idea of developing its own chips is due to the issue of GPU and AI chip shortages. This move is in line with industry trends with Meta and OpenAI also expressing their intention to develop their own chips.
Demand for Anthropic services like Claude has surged this year driven by the introduction of AI agents like OpenClaw and the better capabilities of their models compared to competitors. Anthropic's advantage now is that their models are used by more professional customers who are willing to pay high subscription prices.
But it's like a double-edged sword. The increase in customers requires more data centers to bear the burden of processing power. Costs must be incurred and cost reductions must be made to ensure the company's operations are sustainable.
