Why AI Companies Are Building Their Own Brains

You might have heard about AI models like ChatGPT or Claude, which are becoming really powerful at writing, summarising, and even creating images. But what you might not realise is that these sophisticated AI systems need incredibly powerful computer hardware – essentially, the 'brains' of the operation – to work.

Traditionally, companies like Anthropic (who make Claude) buy these specialised computer chips from other big tech manufacturers. Think of it like a car company buying its engines from a separate engine builder. However, the latest news suggests Anthropic is following in the footsteps of tech giants like Google and OpenAI by starting to design its own custom AI chips. This is a big move and it signals a shift in the AI world.

Why are they doing this? Well, these AI models are getting so advanced that off-the-shelf chips aren't always perfect for their unique needs. By designing their own, Anthropic can tailor the hardware precisely to their AI software, making it more efficient, faster, and potentially more powerful. It’s about getting the most out of their AI models, rather than fitting them into existing hardware.

For us everyday Aussies, this isn't something that will impact our day-to-day immediately. But in the long run, if AI systems become even more efficient and capable because of these custom chips, it could lead to better and more affordable AI tools for small businesses and individuals. It means the AI we use in the future might be a bit smarter and quicker because the companies behind them are building everything from the ground up to work together seamlessly.