Masayoshi Son, the chief executive and billionaire founder of the tech conglomerate SoftBank, believed so strongly in Arm, the British chip design company he bought in 2016 for $32 billion, that he guaranteed some big investors they wouldn’t lose money on the deal.

If anything, Mr. Son told those investors, including Saudi Arabia’s Public Investment Fund, he would be willing to pay them as much as double the acquisition price when they eventually exited, according to four people with knowledge of the negotiations, who requested anonymity to discuss a private arrangement.

Mr. Son recently made good on his guarantee, agreeing to buy out his fellow investors at a $64 billion valuation. But as Arm prepares for its initial public offering, one of the biggest in recent years, Mr. Son’s conviction is being put to the test. The company, which designs chips used in cellphones, including all iPhones, is seeking a market value of $51 billion to $54 billion in its I.P.O. — far below where Mr. Son pegged the company’s value.

Mr. Son’s bullish moves reflect his deeply held belief that Arm — which he recently described as the financial and existential centerpiece of SoftBank — is well positioned to take advantage of the revolution in artificial intelligence. At SoftBank’s annual meeting in June, Mr. Son, now 66, said he had a recent revelation: He wanted to spend the remainder of his life and career as “an architect to build the future of humankind” through artificial intelligence. Arm’s chip designing capabilities, he said, would “play a central role” in accelerating the development of A.I. Mr. Son also sees several other big markets for Arm, such as data centers and automotive.

But his monumental bet on Arm also reflects the urgent need for a big win for SoftBank after years of deals that didn’t live up to their promise.

Mr. Son, who started SoftBank in Tokyo in 1981 after graduating from the University of California, Berkeley, built it into one of the largest technology investors in the world. Early investments in the Chinese e-commerce giant Alibaba and in Japan’s mobile phone industry were critical to SoftBank’s success. However, some of SoftBank’s later investments made through its $100 billion Vision Fund, including WeWork, have been far less rewarding.

Leading up to Arm’s public offering, Mr. Son has been determined to keep as much of the company under SoftBank’s control as he can, according to four people involved in the deal who were not authorized to speak publicly. SoftBank will sell just 10 percent of Arm to investors in the public offering. On average, companies have sold 16 to 29 percent of their shares in I.P.O.s over the past 10 years, according to Dealogic.

Mr. Son originally wanted to sell even less than 10 percent, one of the people said. However, he had taken a large personal loan using his stake in Arm as collateral, and the lenders forced him to sell at least a tenth of the company so they could have access to publicly traded shares — which are easily sellable — in case of a default or a steep plunge in Arm’s value.

Mr. Son has repeatedly told underwriters and potential investors that after this offering, SoftBank has no plans to sell additional shares for several years, according to one person involved in the deal.

Arm’s success is crucial to SoftBank, which in recent years has reeled from a series of troubled bets on technology companies. SoftBank’s Vision Fund, which Mr. Son launched in 2017, raised tens of billions of dollars from big investors, including $50 billion from Saudi Arabia’s sovereign wealth fund.

It also gained notoriety for making money-losing bets on companies that ran into trouble, including WeWork; a robot-making pizza company called Zume; and the Indian hotel chain Oyo. Losses at SoftBank’s Vision Fund unit widened to more than $30 billion in its last fiscal year, which ended in March.

Arm’s public offering comes roughly 18 months after Nvidia, the Silicon Valley chip maker and a customer of Arm, abandoned its offer to buy Arm for $40 billion. The Federal Trade Commission had sued to stop the deal.

While Mr. Son has highlighted how Arm’s chip-design technology could be central to an A.I. revolution, it has yet to produce the kind of growth that would bedazzle investors. The company’s revenue declined slightly for the fiscal year that ended in March, to $2.68 billion from $2.7 billion a year earlier. Profitability dropped as well. Arm had net income of 10 cents per share in the most recent quarter, down from 22 cents the previous year.

Arm’s I.P.O. underwriters pitched the company’s shares to Wall Street investors conservatively, based on what they expected the initial demand to be. But investors who initially appeared to be wary of the chip designer’s valuation and growth potential have become more optimistic after Arm shared its prospects with them, two people with knowledge of the deal said.

The company also offered many of its customers the opportunity to buy up to $100 million each in the I.P.O., two people involved in the deal said. Arm is expected to raise a total of $735 million from companies including Apple, Samsung, Intel and Nvidia.

SoftBank and Arm have other enduring ties. Arm’s chief executive, Rene Haas, who took the helm in 2022 shortly after the company called off its deal with Nvidia, recently joined SoftBank’s board of directors. He travels to Tokyo regularly to meet with Mr. Son and talks to him multiple times a day, according to two people familiar with those activities.

Mr. Son has rarely left Japan since the start of the pandemic but plans to travel to the United States for the first time since 2020 for the Arm I.P.O., one of the people said. He’ll be watching the start of trading from SoftBank’s offices in California. Mr. Haas will spend the morning at Nasdaq’s New York headquarters.