The demand for Nvidia’s processors was greatly surpassed by its quarterly revenue prediction on Wednesday, and the company also announced it will repurchase $25 billion worth of stock, sending its shares surging after hours.

A rise in generative AI technologies that can read and write in human-like ways and are almost completely driven by Nvidia’s chips has shown no signs of slowing down, as seen by Nvidia’s prediction beating estimates by billions of dollars.

The additional $25 billion in share repurchases that Nvidia announced on Wednesday come as its stock has already tripled this year, making it the first chip firm to ever be valued at a trillion dollars. Investors believe Nvidia will be the primary beneficiary of the AI boom, and they bet on this outcome.

Analysts predict that demand for Nvidia’s highly sought-after AI chips is at least 50% higher than supply and that the imbalance will continue for the upcoming quarters.

Jensen Huang, the CEO of Nvidia, stated in a statement that “companies worldwide are transitioning from general-purpose to accelerated computing and generative AI.”

Nvidia, a company based in Santa Clara, California, saw a 9.6% increase in share price in after-bell trading, setting a record.

However, the company’s management claim that its complete AI systems, not just its chips, were the main driver of the quarter’s rise. Nvidia makes full AI computers with memory chips from other suppliers and tens of thousands of other pieces, while being known for its graphics processing units (GPUs), which are known for their graphics.

Microsoft increased 1.9%, Meta Platforms (META.O) up 2.1%, and Palantir Technologies (PLTR.N) increased 4.6% in extended trading on Wednesday as a result of the news about Nvidia.

According to Daniel Ives, analyst at Wedbush Securities, the results were a “‘drop the mic’ moment that will have a ripple effect for the tech space for the rest of the year.”

All companies, from AI start-ups to significant cloud service providers like Microsoft (MSFT.O), are vying for more Nvidia chips. Due to businesses there placing urgent orders to stockpile chips before any additional U.S. export restrictions take effect, demand from China is also at an all-time high.

Colette Kress, the company’s chief financial officer, told analysts on a conference call that should the U.S. impose new export restrictions on sales of AI chips to China, it would not immediately affect the company’s financial performance. Such restrictions would “result in a permanent loss of an opportunity for the U.S. industry to compete and lead in one of the world’s largest markets.”

A third-quarter revenue estimate of roughly $16 billion, plus or minus 2%, was made by the corporation. The average estimate of the analysts surveyed by Refinitiv was $12.61 billion.

Compared to projections of $11.22 billion, adjusted sales for the second quarter came in at $13.51 billion.

According to Refinitiv statistics, revenue at the company’s data center division increased 141% to $10.32 billion in the three months ending July 30, exceeding analyst projections of $7.69 billion by more than $2 billion.

Jacob Bourne, senior analyst at Insider Intelligence, noted that the company’s Q2 earnings “underscore its dominant position in harnessing the AI momentum.” Yet as demand for Nvidia’s chips grows worldwide, it is crucial to overcome supply chain obstacles to increase manufacturing.

In order to assure supply, Nvidia is paying a lot of money. The company reported a 53% increase in inventory commitments from the prior quarter to $11.15 billion, mostly due to the long-term supply requirements for its data center chips.

Refinitiv forecasts that analysts predict Nvidia’s data center sector sales to increase to as high as $40 billion for its fiscal 2025, led by Nvidia’s leadership in AI processors and other associated technologies like the software to use those chips to power products like ChatGPT.

While analysts anticipate that rival Advanced Micro Devices’ (AMD.O) major AI processor will take some market share away from Nvidia next year, they also feel that Nvidia’s software has a significant advantage over ROCm, a CUDA rival.

The semiconductor industry has suffered recently due to lackluster sales of chips meant for personal computers and data centers. However, AI is a bright area, with startups and cloud computing companies alike purchasing Nvidia and other companies’ AI-related processors, including Broadcom and Marvell Technology.

Analysts anticipate that expenditure on AI will continue to rise at the expense of other conventional server hardware.

According to Refinitiv statistics, revenue at Nvidia’s gaming division increased to $2.49 billion, exceeding analyst projections of $2.4 billion.

In the second quarter, the firm earned $2.70 per share after items, beating projections of $2.09, according Refinitiv data.

Nvidia anticipates its adjusted gross margin for the current third quarter to be 72.5%, plus or minus 50 basis points. According to statistics from Refinitiv, analysts expect the gross margin to be 70.4% on average.