Why You Might Be Paying For AI You Don't Even Use

You've likely heard of Nvidia – they're the big name behind many of the powerful computer chips that make artificial intelligence, or AI, possible. They've been incredibly successful, and their chips are the workhorses for advanced AI systems used by companies like OpenAI (of ChatGPT fame) and Google.

However, a new trend is emerging. Many of these big tech companies are now trying to build their own specialised AI chips. You might think this would reduce Nvidia's power, as businesses try to cut out the middleman. But it turns out, Nvidia has another trick up its sleeve, and it’s called NVLink. Think of NVLink as a super-fast highway connecting all the different parts of an AI system, allowing them to talk to each other incredibly quickly.

Even if these tech giants design their own chips, they often still rely on Nvidia's NVLink technology to make everything run smoothly. This means they might have to pay Nvidia a licensing fee, much like paying a royalty for using a patented invention. It's like buying a car from one brand, but that car still uses an engine design or a special braking system patented by another company. Nvidia is essentially making sure that a piece of its crucial technology is still part of the puzzle, no matter who builds the main computer chip.

For Australian small businesses and everyday users, this might seem like a distant technical detail. But it matters because it helps explain why advanced AI tools might continue to be expensive to develop and run. If one company holds a key piece of the puzzle, they can influence the cost and direction of AI development globally. This could mean higher costs passed down to the software and services we all use, or it could slow down innovation if companies are tied to one supplier's technology.

It's a reminder that even in the rapidly evolving world of AI, the foundational technologies and the companies that own them continue to play a very significant role in shaping the future.