Two small but telling stories from the technology world Friday capture the current state of the AI race: a push to make powerful AI models run on ordinary computers, and continued strain on the chips that make AI possible in the first place.
Meta launched a new open-weight AI model built specifically to run on Macs, according to TechCrunch. Open-weight models are AI systems whose underlying parameters are published for anyone to download and run themselves, rather than being locked behind a company's own servers and paid access. Meta has positioned itself as the leading advocate of this open approach among major AI developers, in contrast to rivals who keep their most capable models closed. A version tuned to run well on Apple's own hardware matters because it lets developers and hobbyists experiment with capable AI locally, without needing to pay for cloud computing time or send their data to an outside server.
On the hardware side, Nvidia's RTX 5090, currently the company's flagship consumer graphics card and a favorite among gamers and AI hobbyists who use it for local processing, reportedly disappeared from the shelves and website of the retailer Micro Center, fueling speculation of a production halt, according to a report from Tom's Hardware carried by Yahoo Tech. Nvidia's chips sit at the center of the broader AI boom, powering everything from consumer gaming rigs to the massive data centers that train large language models, which has made any sign of a supply disruption closely watched by investors as well as consumers.
The two stories are connected by a common thread: demand for AI compute, whether for giant data centers or a single enthusiast's desktop, continues to outstrip what the industry can comfortably supply, even as the software side of the business works to make powerful models more accessible and less dependent on expensive hardware.
Watch for Nvidia's own comment on the RTX 5090 supply situation, and for how developers respond to Meta's new Mac-focused model in the days ahead.
What does 'open-weight' mean for an AI model?
It means the company has published the model's underlying parameters publicly, so developers can download and run it on their own hardware instead of only accessing it through the company's paid service.
Why does a graphics card shortage matter beyond gaming?
Nvidia's high-end graphics cards are also used by AI developers and hobbyists for local processing, so supply constraints touch both the gaming and AI communities.