Nvidia, headquartered in Santa Clara, California, has increased its share buyback authorization by $150 billion as growing demand for artificial intelligence training and inference boosts its cash generation.
The additional approval surpasses the share repurchase authorization of Apple, elevating Nvidia's total stock buyback capacity to $235 billion. The semiconductor company expects to utilize the total authorization through fiscal year 2028.
The approval of a $110 billion amount in 2024 represents the largest increase ever recorded in a share buyback program. Corporate buyback programs allow publicly traded companies to repurchase their own shares on the open market, reducing the total supply of stock outstanding and returning capital to shareholders.
Shares in Nvidia rose 1.2% in pre-market trading following the news. As of Friday's close on September 25, the company's stock had accumulated a gain of more than 20% this year. Nvidia is a major developer of graphics processing units and specialized microchips that power data centers and artificial intelligence applications globally.
Valuation and market expectations
Nvidia shares were trading at approximately 16.5 times projected earnings for the next 12 months, according to data compiled by London Stock Exchange Group. That level marks the company's lowest valuation multiple since January 2015 and stands well below its 15-year average multiple of 30.
Some analysts suggested that the compressed earnings multiple could signal a slowdown in expected profit growth. Price-to-earnings ratios measure how much investors are willing to pay for each dollar of forecasted profit. London Stock Exchange Group, which compiled the data, tracks global financial analytics and market metrics.
The buyback expansion follows a period of reduced repurchase activity. The company experienced a slowdown of approximately 50% in its share buybacks between July and September 23.
Revenue forecasts and startup investments
Last month, Nvidia forecast revenue growth of about 70% for fiscal year 2028. The projection helped reassure investors who had questioned the sustainability of the surge in artificial intelligence spending after years of explosive expansion.
Demand for Nvidia's hardware has been driven by twin AI computational tasks. AI training involves feeding vast datasets into machine learning models to build capabilities, while inference refers to executing operational models to process real-time requests. Alongside hardware production, Nvidia has also been investing in artificial intelligence startups and cloud service providers.
Those financial investments have attracted attention from some investors over whether the funding indirectly supports market demand for Nvidia's own chips. Nvidia concluded its July quarter holding $22.44 billion in cash and cash equivalents.
