- AI PCs emerge as a viable option for running local AI without unpredictable costs
- One-time PC cost eliminates the need to shell out for cloud token fees
- Further research reaffirms the growing popularity of smaller models
According to new data from Gartner, now could be a good time to buy AI PCs as cloud computing faces many challenges in a rapidly changing business world.
Data center construction is lagging behind demand as slack chains come under strain and local communities oppose new projects, meaning metered computing could end up costing some companies more than they anticipated.
By moving some of their AI processing locally, businesses could avoid some of these additional monthly costs with the one-time purchase of a more powerful PC as part of their regular refresh cycles.
AI PCs present an ideal hybrid computing model
Although adoption of AI PCs started quite slowly with businesses struggling to understand the benefits, they are now seen as a backup option in the cloud rather than a primary benefit in their own right.
With increasing use of AI and unpredictable token consumption hitting businesses hard, forecasting monthly costs is a major new challenge many are facing.
Small language and reasoning models, including models specially trained for individual business use cases, ultimately require fewer resources than major boundary models, allowing them to be run locally as part of a broader hybrid approach.
Gartner predicts that speech and chat, text generation, image and sound generation, and more may soon migrate to workers’ PCs, with only the most intensive tasks routed through hyperscaler data centers.
By 2029, the company’s researchers predict that about a third (30%) of companies could use AI PCs to reduce the costs of AI tokens in the cloud. By 2030, 70% of enterprise PCs could run some GenAI tasks locally.
Omdia researchers have also noticed a shift in the use of AI models, with small and medium-sized models proving popular, with domain-specific tasks not requiring the full extent of computing.
“Older GPUs retain their value and remain in use as they continue to provide a cost-effective option for inference and disaggregation of small and medium-sized models,” said Alexander Harrowell, senior principal analyst for Advanced Computing.
Via The register
Follow TechRadar on Google News And add us as your favorite source to get our news, reviews and expert opinions in your feeds.
