Microsoft spends approximately 47% of its capital expenditure on Nvidia’s AI chips

As of Nvidia’s 2025 fiscal fourth-quarter results, Microsoft is estimated to be the biggest spender on its AI chips, with roughly 47% of its capital expenditures directly related to AI chips. The estimate accounts for nearly 19% of Nvidia’s revenue on an annual basis.
Nvidia’s stock price hit a high of $153 at the beginning of the year, but it has been dropping dramatically recently due to investor worries. As the first quarter results are looming, analysts believe spending on Nvidia’s AI chips for data center revenue will continue to slow down.
Big Tech companies drive Nvidia’s AI chip sales
On Sunday, a Bloomberg report estimated on an annual equivalent basis that Nvidia’s fellow members of the Magnificent Seven Group, including Microsoft, Amazon, Meta, Tesla, Alphabet, and Google, are the biggest customers for its AI chips. According to the report, Microsoft is the most significant driver of Nvidia’s revenue, followed by Amazon, Meta, and Google’s parent Alphabet (GOOG).

25% of Meta’s capital expenditure goes to Nvidia, and the company accounts for just over 9% of Nvidia’s annual revenue. According to the report, the results reflected direct spending on Nvidia. Indirectly, Big Techs props up Nvidia, for instance, Microsoft spends a lot on renting data center capacity from cloud provider CoreWeave (CRWV). Core Weave spends billions on Nvidia’s AI chips to run those data centers.
Microsoft accounted for 72% of CRWV revenue in the last fiscal year’s financial results. According to DA Davidson analyst Gil Luria, Nvidia relies on a handful of large technology companies for about half its revenue. He added that the revenue had grown dramatically in recent years, but it appears to moderate Nvidia’s growth.
According to Nvidia’s 2022 fourth-quarter earnings report, Microsoft spent less than 1% of its capital on Nvidia and similarly accounted for less than 1% of Nvidia’s revenue on an annualized basis.
Luria warned that the expenditure may slow as customers increasingly use the chips they developed in-house. According to Luria, Nvidia chips have a considerable advantage in AI pre-training, and the cost of the custom chips is competitive.
Google, Microsoft, Meta, and Amazon have developed custom chips to target specific AI workloads, making the chips more efficient for each company’s needs.
Microsoft could moderate its spending on AI hardware
Rival chip maker like Broadcom (AVGO) are developing custom chips for their customers, whereas Nvidia’s GPUs are for more general-purpose AI computing. Investors have raised concerns about how Big Tech firms could monetize the technology and generate a return on their investment.
Analysts believe spending on Nvidia’s AI chips for data center revenue will continue to slow down. According to Bloomberg consensus data, Nvidia’s data center revenue is expected to jump more than 74% to $39 billion in the first quarter of the following fiscal year report. In the first quarter of Nvidia’s fiscal year 2025, data center revenue rose by 427%.
Microsoft confirmed its $80 billion spending plan to build AI data centers, with 80% of the expenditure going to the US. Meta raised its 2025 expenditure outlook, predicting that spending will fall between $64 billion and $72 billion, up from the previous range of $60 billion and $65 billion.
AMD won over Nvidia’s top customers with its MI300X data center GPU AI chip and is planning to ship its MI350 series, which is based on a new architecture called CDNA (compute DNA). CDNA delivers 35 times more performance than the MI300X, so it is a legitimate threat to Nvidia’s AI chips.
Nvidia’s stock reached a high of 153.13 on January 7, but then it dropped, which was caused by a series of investor worries. Investors were worried about the sustainability of AI data center demand from hyperscalers after two years of heavy investment triggered by the launch of OpenAI’s ChatGPT in November 2022.
China’s DeepSeek AI model reportedly required less advanced processors to build, threatening Nvidia’s sales. Nvidia is trading at 131.29, with a change of 1.16% today.
Cryptopolitan Academy: Want to grow your money in 2025? Learn how to do it with DeFi in our upcoming webclass. Save Your Spot
Microsoft spends approximately 47% of its capital expenditure on Nvidia’s AI chips

As of Nvidia’s 2025 fiscal fourth-quarter results, Microsoft is estimated to be the biggest spender on its AI chips, with roughly 47% of its capital expenditures directly related to AI chips. The estimate accounts for nearly 19% of Nvidia’s revenue on an annual basis.
Nvidia’s stock price hit a high of $153 at the beginning of the year, but it has been dropping dramatically recently due to investor worries. As the first quarter results are looming, analysts believe spending on Nvidia’s AI chips for data center revenue will continue to slow down.
Big Tech companies drive Nvidia’s AI chip sales
On Sunday, a Bloomberg report estimated on an annual equivalent basis that Nvidia’s fellow members of the Magnificent Seven Group, including Microsoft, Amazon, Meta, Tesla, Alphabet, and Google, are the biggest customers for its AI chips. According to the report, Microsoft is the most significant driver of Nvidia’s revenue, followed by Amazon, Meta, and Google’s parent Alphabet (GOOG).

25% of Meta’s capital expenditure goes to Nvidia, and the company accounts for just over 9% of Nvidia’s annual revenue. According to the report, the results reflected direct spending on Nvidia. Indirectly, Big Techs props up Nvidia, for instance, Microsoft spends a lot on renting data center capacity from cloud provider CoreWeave (CRWV). Core Weave spends billions on Nvidia’s AI chips to run those data centers.
Microsoft accounted for 72% of CRWV revenue in the last fiscal year’s financial results. According to DA Davidson analyst Gil Luria, Nvidia relies on a handful of large technology companies for about half its revenue. He added that the revenue had grown dramatically in recent years, but it appears to moderate Nvidia’s growth.
According to Nvidia’s 2022 fourth-quarter earnings report, Microsoft spent less than 1% of its capital on Nvidia and similarly accounted for less than 1% of Nvidia’s revenue on an annualized basis.
Luria warned that the expenditure may slow as customers increasingly use the chips they developed in-house. According to Luria, Nvidia chips have a considerable advantage in AI pre-training, and the cost of the custom chips is competitive.
Google, Microsoft, Meta, and Amazon have developed custom chips to target specific AI workloads, making the chips more efficient for each company’s needs.
Microsoft could moderate its spending on AI hardware
Rival chip maker like Broadcom (AVGO) are developing custom chips for their customers, whereas Nvidia’s GPUs are for more general-purpose AI computing. Investors have raised concerns about how Big Tech firms could monetize the technology and generate a return on their investment.
Analysts believe spending on Nvidia’s AI chips for data center revenue will continue to slow down. According to Bloomberg consensus data, Nvidia’s data center revenue is expected to jump more than 74% to $39 billion in the first quarter of the following fiscal year report. In the first quarter of Nvidia’s fiscal year 2025, data center revenue rose by 427%.
Microsoft confirmed its $80 billion spending plan to build AI data centers, with 80% of the expenditure going to the US. Meta raised its 2025 expenditure outlook, predicting that spending will fall between $64 billion and $72 billion, up from the previous range of $60 billion and $65 billion.
AMD won over Nvidia’s top customers with its MI300X data center GPU AI chip and is planning to ship its MI350 series, which is based on a new architecture called CDNA (compute DNA). CDNA delivers 35 times more performance than the MI300X, so it is a legitimate threat to Nvidia’s AI chips.
Nvidia’s stock reached a high of 153.13 on January 7, but then it dropped, which was caused by a series of investor worries. Investors were worried about the sustainability of AI data center demand from hyperscalers after two years of heavy investment triggered by the launch of OpenAI’s ChatGPT in November 2022.
China’s DeepSeek AI model reportedly required less advanced processors to build, threatening Nvidia’s sales. Nvidia is trading at 131.29, with a change of 1.16% today.
Cryptopolitan Academy: Want to grow your money in 2025? Learn how to do it with DeFi in our upcoming webclass. Save Your Spot