Taiwan Semiconductor Manufacturing Company, the world's largest contract chipmaker, is developing a new packaging technology aimed at lowering the cost of chips while improving their performance. According to GSMArena, citing sources familiar with the matter, Nvidia's AI chips will be the first to rely on the new technology, which is expected to arrive by 2028.
Chip "packaging" refers to how individual silicon components are assembled and connected into a finished product. As squeezing more transistors onto a single chip gets harder and more expensive, packaging has become a key lever for boosting performance and managing cost — making advances in this area increasingly important to the companies that design cutting-edge processors.
The move comes as TSMC rides a wave of demand for chips that power artificial intelligence. The company reported strong 2025 results, with net income of $54.55 billion, according to the Economic Times. The same report said TSMC achieved over $120 billion in revenue for the year, with earnings boosted by AI-driven chip demand.
Nvidia, whose processors dominate the market for training and running AI systems, relies on TSMC to manufacture its most advanced chips. Being the first customer for the new packaging technology would position Nvidia to benefit early from any cost or performance gains.
The sources cited by GSMArena describe the technology as still in development, with a target of 2028 — meaning the impact is years away rather than imminent.
Why it matters: cheaper, faster chips from the dominant manufacturer could shape the economics of the AI boom, influencing how much it costs to build and run the systems that increasingly underpin technology products.