IBM is adopting a strategy move by looking into using its in-house created artificial intelligence processors to lower the operating expenses of its recently introduced “watsonx” cloud service. The “Artificial Intelligence Unit” chip may be included into the company’s cloud service, according to Mukesh Khare, General Manager of IBM Semiconductors, who spoke with attendees at a semiconductor conference in San Francisco.
With this move, IBM is attempting to capitalize on the rising demand for generative AI tools that can create content that is similar to human writing. It comes after their ground-breaking AI system, Watson, had a difficult time gaining momentum in the market for more than a decade, in part because of exorbitant costs. With regard to their outstanding power efficiency, Khare emphasized that employing their own unique chips might be a game-changer.
Despite announcing the chip’s existence back in October, IBM stayed mum on the maker and the planned uses. According to Khare, the device was created by Samsung Electronics working with IBM on semiconductor research. IBM wants to determine whether or not Watsonx can include it.
Khare claimed that hundreds of prototype chips are currently in use, despite the fact that IBM hasn’t specified a precise timeframe for making the device available to cloud customers. IBM is currently pursuing in-house design of AI processors with other tech behemoths like Google and Amazon.
Though Nvidia, whose chips dominate the market for AI training, is not directly IBM’s competitor, Khare highlighted that this is not the company’s intention. As referred to by industry insiders as “inference,” their chip is instead designed for cost-effectiveness. Khare thinks their chip will have the most influence on inference, which entails using an already trained AI system to make judgments in the real world.
In summary, IBM’s move into deploying its AI processors for cloud services might change the sector by providing a practical answer for real-world AI applications, marking a hopeful turn in the company’s AI path.
IBM Explores In-House AI Chips to Cut Cloud Costs
