SoftBank raised a large amount of US$11.1 billion for AI investment and issued junk bonds, setting a record for high-yield global corporate bonds

The Zhitong Finance App learned that SoftBank Group founder Sun Zhengyi has long clearly regarded artificial intelligence as the group's core bet on the future. To finance this ambitious strategy, SoftBank is leveraging the global credit market with unprecedented strength. The company completed issuing high-yield bonds worth US$11.1 billion last week, setting a record for the global corporate junk bond market.

Much of this financing was used to replace SoftBank's previous $40 billion bridge loan to invest in OpenAI. Faced with huge financing needs, SoftBank finally decided to return to the US institutional bond market and issue securities to US institutional investors with strong financial strength. This is also the first time in over a decade that the company has used this method of financing in the US market.

In order to enter the approximately $23 trillion US bond and other debt securities market, SoftBank's current offering needs to meet the relevant requirements of US Securities and Exchange Commission (SEC) Rule 144A. In the past ten years, SoftBank's smaller overseas bond financing mainly used the Regulation S framework and was issued only to investors outside the US.

To this end, SoftBank and its team of lawyers have been preparing for several months. One of the key issues is determining whether SoftBank will be recognized as an “investment company” under US law. This identification may directly affect its ability to enter the US capital market.

According to the US Investment Companies Act of 1940, if a company holds more than 40% of the specified assets, it may be recognized as an investment company unless it meets the relevant exemptions. Once a foreign company falls into this category, it may be necessary to obtain special approval from the SEC before issuing securities to US investors.

SoftBank stated in this bond issuance memorandum that the company believes that its investment securities account for less than 40% of unconsolidated assets, and positions itself as a “strategic holding company focused on the technology sector.”

People familiar with the matter revealed that the final judgment was based on a comprehensive test of SoftBank's overall business and assets. One important factor is that SoftBank's holding about 86% of the chip design company Arm is classified as an operating subsidiary rather than passive investment securities, which helps SoftBank avoid being included in the category of investment companies defined by the Investment Companies Act.

SoftBank declined to comment.

Arm (ARM.US) stock price rebound creates a window for huge debt issuance

Just a few months ago, there was still great uncertainty about whether SoftBank could successfully complete such huge bond financing. Affected by factors such as the overall correction of AI concept stocks, heightened valuation concerns, and the unease of some investors about SoftBank's continuous increase in OpenAI, SoftBank's stock price plummeted by nearly 50% from early June to late July.

Since then, SoftBank's stock price has clearly rebounded, with a cumulative increase of about 42% so far this year. People familiar with the matter said that the stronger Arm stock price in particular provided important support for SoftBank, and also gave the company an opportunity to test demand in the bond market.

SoftBank executives then launched an intensive global investor roadshow. In early September, they traveled to London to meet with major European investors and attended a leveraged finance conference hosted by Goldman Sachs and J.P. Morgan. Afterwards, SoftBank Chief Financial Officer Yoshimitsu Goto and other executives flew to New York to hold talks with investors at the Citibank office to understand the market's interest in issuing potential junk bonds.

Investors were quick to respond to these meetings. People familiar with the matter said that many institutions later submitted letters of intent to the underwriting bank, indicating that if the transaction starts, they may participate in the subscription.

However, US investors still raised numerous questions about SoftBank's complex corporate structure, future investment plans, and data center project progress.

Contrary financing under the AI “bubble” controversy

The market environment facing SoftBank's current financing is also not ideal. US Treasury yields have risen sharply, and market discussions on the “AI bubble” driven by debt financing have intensified.

At the same time, a number of AI industry leaders have begun warning of possible security risks posed by advanced artificial intelligence models and calling for a slowdown in the development of cutting-edge models. OpenAI CEO Sam Altman also said that the company will not go public this year, which means investors who have been waiting for OpenAI to exit the channel for a long time still need to be patient. For SoftBank, which has already bet huge sums of capital on OpenAI, this has also increased uncertainty about the valuation and monetization prospects of its portfolio.

Hiroki Takei, strategist at Resona Holdings, pointed out that for AI-related bond investors, the risk is not just an extreme situation where the AI boom completely breaks down.

He said that if the current model of “pre-investment and lagging commercial monetization” continues, huge AI capital expenses may continue to erode corporate free cash flow, making enterprises more dependent on external financing and refinancing. Should AI demand then slow down, credit spreads are likely to widen significantly.

However, at this stage, many investors still view these scenarios more as risks to consider in due diligence rather than an impending industry reversal, and the global AI debt financing boom continues as a result.

In order to advance SoftBank's huge deal, banking teams in Asia and Europe even worked continuously the weekend before the launch. People familiar with the matter said that the relevant conference spanned three time zones, and staff from London and Tokyo stayed up all night communicating with American investors.

Nearly 10% yield finally attracts investors

There were also new twists and turns in the release process. The news that SoftBank-backed data center service provider SB Energy has postponed its initial public offering (IPO) has raised further concerns among investors. The underwriting bank then received numerous questions about when OpenAI went public, SB Energy's delayed IPO, and what these changes meant for SoftBank's holdings. Ultimately, the high yield offered by SoftBank became the key to attracting investors. The longest tenure bond issued this time was 7.5 years, and the final pricing yield reached 9.75%, which is close to double-digit levels.

Both S&P Global Ratings and Fitch gave SoftBank a “BB+” rating, the highest of the two institutions' non-investment grade ratings.

SoftBank, on the other hand, mainly measures its debt level through the loan value ratio (LTV), that is, it uses the ratio of net debt to the value of the equity assets it holds to calculate leverage. The company promises to generally keep LTV below 25%, and only allows it to rise to 35% in exceptional circumstances. As of the end of June this year, the indicator was only 13%.

Satoru Aoyama, senior director of Fitch Asia Pacific Corporate Ratings, said that investors are aware of SoftBank's risks as a BB+ issuer, and a bond yield of close to 10% is already close to the level of return on stock investment. However, at the same time, SoftBank did not issue bonds for up to 30 years, and the company maintained financial discipline such as lower LTV, which also provided some indirect protection for creditors.

SoftBank's record bond issuance is the latest example of how the AI investment boom is reshaping the global credit market. As data centers, chips, computing power infrastructure, and AI model development require more and more capital, technology companies' reliance on debt financing is deepening.

Aoyama said that although the market is still debating the risks that may be caused by excessive borrowing in the AI industry, demand for financing is clearly growing rapidly and has covered the entire AI industry chain.