by Key Takeaways
- By developing energy-efficient analog chips as GPU substitutes, Sagence AI is enhancing AI sustainability and performance.
- For more effective AI operations, Sagence’s novel chip design employs analog processing rather than digital processing.
- By avoiding the requirement for data transfer, these analog devices may speed up task completion.
A new chip manufacturing company called Sagence AI is developing a cutting-edge analog chip that will improve AI capabilities.
This novel analog chip, in contrast to conventional chips, will save the requirement for data transfer, which could speed up AI model task execution.
Analog chips will lessen the need for high power consumption
Sagence AI’s ambitions to transform the chip business with its cutting-edge analog chips are highlighted in a recent TechCrunch piece. The company’s founder, Vishal Sarin, hopes that these analog chips can encourage sustainable energy consumption while overcoming the constraints of conventional digital hardware.
The goal of this novel strategy is to tackle some of the major issues that AI is now experiencing, like graphic processing units (GPUs).
The majority of AI models are now powered by GPUs; however, their high energy consumption is a major drawback. A Goldman Sachs study suggests that by 2030, data centers’ increasing reliance on GPUs might result in a 160% increase in power consumption.
Additionally, digital chips are inherently inefficient. They utilize more energy and slow down computations since they require a lot of components and move data between memory and processors all the time.
Sarin contends that this trend is unsustainable, which is why Sagence AI analog chips use significantly less energy and are made to handle AI tasks without the continuous data transfers that digital chips require.
With this ability, analog circuits may be able to process data more quickly and densely, providing a viable substitute for current technology.
Analog Chips: A Well-Known But Developing Idea
In the tech industry, analog chips are not new. As part of its study to improve the effectiveness and performance of AI models on fresh data, the well-known computer company IBM offered analog chips last year.
Like Sagence’s vision, IBM’s innovation was motivated by the conviction that if AI progress kept depending only on digital hardware, it would stall.
