SoftBank Group has launched more than $11 billion in dollar- and euro-denominated bonds to fund a further investment in OpenAI and general corporate purposes, according to a term sheet reported by Reuters. The financing does not itself build an AI model or data centre. It shows how a major backer is converting its commitment to AI into long-term capital that investors must be willing to buy.
At a glance
- The proposed notes total $10 billion in US dollars and €1 billion, according to the term sheet seen by Reuters.
- Proceeds are intended partly for a $10 billion third tranche of SoftBank’s follow-on OpenAI investment, expected to close October 1.
- The bonds would replace a $10 billion bridge facility used for the investment, shifting the financing toward longer-dated debt.
What SoftBank has launched
Reuters reported on September 21 that SoftBank had launched $10 billion of senior unsecured notes in US dollars and €1 billion of euro-denominated notes. The reported maturities span 3½, 5½ and 7½ years for the dollar notes, and four and six years for the euro notes.
The term sheet said a portion of the proceeds would fund SoftBank’s $10 billion payment for the third tranche of its follow-on investment in OpenAI. Reuters said the new bond issuance is expected to price September 24 and settle September 29. SoftBank could not immediately be reached for comment because of a public holiday in Japan, Reuters reported.
Why replace a bridge loan with bonds?
A bridge loan is usually temporary financing: it lets a borrower move quickly while longer-term funding is arranged. Bonds, by contrast, are sold to investors and commonly carry fixed maturities. Replacing a bridge facility can spread repayment across several years and diversify the funding base.
That does not make the financing risk-free. Interest costs, market conditions and SoftBank’s ability to meet obligations still matter. The transaction also relies on investor demand at the final price, which may differ from the initial terms. In other words, the announcement is a launch, not proof that every planned note has already been sold.
Why link bonds to an OpenAI investment?
Developing and deploying frontier AI requires far more than research spending. It also needs computing capacity, chips, networking, electricity, data centres and skilled staff. Large investments in model makers can therefore be tied to a much wider infrastructure build-out.
SoftBank has made AI a core strategic focus. Its investor-relations materials provide the company’s official disclosures on its financial position and investments, while the term sheet reported by Reuters describes this specific financing. Readers should distinguish between a company’s investment in OpenAI and OpenAI’s own operating revenue, spending or future public-market plans; they are connected, but not the same measure.
Why it matters
The proposed sale is a useful marker of how the AI boom is being financed. Equity investors often receive the most attention, but debt markets are increasingly relevant when companies finance large commitments over multiple years. Bond buyers effectively assess whether an issuer can carry its borrowing alongside the uncertainty of a fast-changing technology market.
It also puts attention on capital structure. A project can be strategically important and still face higher financing costs if interest rates rise or investors demand more compensation for risk. For AI businesses and infrastructure backers, access to durable capital may determine how quickly ambitious plans become operational capacity.
For another view of AI infrastructure demand, read our report on Nscale’s IPO filing. Our explainer on low-carbon aluminium for data centres shows why the build-out also reaches materials and energy supply chains.
What happens next?
The immediate milestones are pricing on September 24, settlement on September 29 and the expected October 1 investment tranche. Investors will also watch the final yields, demand for the notes and any subsequent disclosures from SoftBank.
The broader takeaway is simple: the AI race is becoming a financing story as well as a technology story. Models may capture attention, but the infrastructure supporting them is increasingly funded through ordinary financial instruments such as loans and bonds.
