TI and IBM were racing each other to market with the next speed of 1394 silicon; S200 (or 200 Megabits per second). TI's first Phy (the chip that determines the speed) was an S100 chip. IBM, seeking competitive advantage, had skipped the S100 speed and went straight to S200, giving them a significant head start over TI.
Debugging of the S200 Phy was not going well at TI. They’d licensed their Phy design from Apple. Phys are notoriously difficult semiconductors to get working right and the Apple designs were getting a bad reputation in the industry. TI abandoned the S200 Phy and hired a new “clean” Phy engineer to start work on S400 Phy which contained no Apple IP. IBM continued to slog away at their own S200 design which they were eventually able to bring to market and pick up some modest business.
Unfortunately for IBM, semiconductor fabs worldwide continued to improve rapidly, making the cost of producing new silicon cheaper every year. Once TI launched the S400 Phy, designed to be produced by the newest semiconductor fabs, it was costing TI less to produce an S400 Phy than it cost IBM to produce the S200. The time involved to reconfigure the S200 Phy to be produced on a new fab (at the cost savings) was prohibitive.
TI was able to offer twice the speed at a lower cost. Camcorders continued to use the TI S100 Phy since that was fast enough (and the cost of qualifying a new chip was onerous). New products selected the TI S400 chip because it was the cheapest Phy on the market. Leaving IBM stuck in the middle with a product that no one wanted.