OpenAI Cuts Model Prices and Expands Samsung Chip Partnership

OpenAI is aggressively expanding its enterprise footprint by slashing costs and deepening its semiconductor supply chain. Chief Financial Officer Sarah Friar announced at the Goldman Sachs Communacopia + Technology Conference that the company has reduced the price of its Luna model by 80% to undercut open-source rivals and Chinese open-weight models. This pricing shift coincides with a 32% increase in enterprise revenue between June and July, bringing the company closer to a balanced revenue split between its consumer and enterprise segments.

The company is also accelerating its hardware ambitions. Friar noted that OpenAI used its own models to finalize the design of its Jalapeno chip in nine months. This custom AI inference chip was co-designed with Broadcom Inc and manufactured by TSMC. To further these efforts, Harrison Kim, General Manager of OpenAI Korea, announced expanded cooperation with Samsung Electronics on joint research and production for next-generation chips.

OpenAI's growth in South Korea has been particularly rapid, with ChatGPT Enterprise users among local institutions increasing 28-fold as of August. This surge follows the September 4 launch of the GPT-6 Astra model. Additionally, Samsung Electronics and SK Hynix have signed letters of intent to provide memory chips for the Stargate data center project, while Samsung remains one of the largest global deployers of ChatGPT.