Kneron, a full-stack AI company based in San Diego, has raised $49 million (€46.4 million) in strategic funding from investors including Foxconn, the HH-CTBC Partnership (Foxconn Co-GP Fund) and Alltek, bringing its full Series B financing total to $97 million (€91.9 million).

The round was led by Horizons Ventures, with participation across several extensions from Liteon Technology Corp, Adata, Palpilot, Foxconn and the HH-CTBC Partnership (Foxconn Co-GP Fund). Using these funds, Kneron will look to accelerate the deployment of new, advanced, AI solutions – with a focus on delivering nano GPT solutions for the automotive industry.

Kneron itself provides end-to-end integrated hardware and software solutions that enable on-device edge AI inferencing. In 2021, it launched an edge AI chip supporting transformer neural networks – a technology that underpins GPT models. The company is also developing NPU (neural processing unit) chips that offer qualities that cater to the needs of AI computing. Its latest auto-grade NPU chip, the KL730, supports advanced lightweight GPT LLMs, such as nano GPT, to expand the number of options on the market.

In aiding autonomous driving, Kneron will utilize the funds to support their latest AI chips, which apply transformers that are commonly applied to language processing functions to image-based applications. Here, the chips leverage the capabilities of these transformers to better process time-series and consider image data holistically. Kneron has said that this helps to improve contextualization and processing, while increasing the accuracy of image-based applications by a minimum of 30%.

Alongside its fundraising announcement, Kneron also announced plans to expand its partnership with Foxconn to accelerate the deployment of advanced AI for automotive (and further) applications. Here, the companies are working to develop an ultra-lightweight AI chip that operates GPT models off the cloud.