Taiwan’s Foxconn, the contract manufacturer that assembles Apple’s iPhones, has built its own AI. Called FoxBrain, the company says the large language model was trained in just four weeks with help from Nvidia, using 120 of that company’s H100 chips. FoxBrain has reasoning and mathematical skills and can analyze data and generate code. Initially built for in-house use, Foxconn says it intends to open source the model and hopes it will become a collaborative tool for its partners and enable advancements in manufacturing techniques and supply-chain management.

The Hon Hai Research Institute (backed by Foxconn parent Hon Hai Technology Group) said in an announcement that FoxBrain represents a “milestone in the development of Taiwan’s AI technology with a more efficient and lower-cost model training method” that can be completed in a very short amount of time.

“It is Taiwan’s first large language model with reasoning capabilities that is optimized for traditional Chinese and Taiwanese language styles,” according to Reuters.

FoxBrain is based on Meta Platforms’ Llama 3.1 architecture with 70B parameters, per the announcement, which also includes benchmark test results (indicating that “in most categories among TMMLU+ test dataset, it outperforms Llama-3-Taiwan-70B of the same scale, particularly excelling in mathematics and logical reasoning”).

“The company said its model’s performance was slightly behind some models of China’s DeepSeek but was approaching world-class levels,” reports The Wall Street Journal, which calls Foxconn “the world’s largest contract electronics maker.”

Despite its size and scale, Foxconn is “facing challenges in its core electronics manufacturing business due to industry shifts and declining profitability,” and as a result “has been diversifying into areas such as AI and electric vehicles,” notes WSJ.

Reuters points out that in addition to its contract with Apple, Foxconn also produces artificial intelligence servers for Nvidia.

Nvidia provided support through technical consultation and use of its Taiwan-based Taipei-1 Supercomputer, with model pre-training completed using Nvidia NeMo.

FoxBrain “will become an important engine to drive the upgrade of Foxconn’s three major platforms: smart manufacturing, smart EV and smart city,” Foxconn explains.

Foxconn plans to discuss additional details about FoxBrain at Nvidia’s GTC 2025 developer conference, March 17-21 in San Jose, California, where an estimated 25,000 IRL and 300,000 virtual attendees are expected to participate, according to VentureBeat.