According to Zhitong Finance APP, SB, the data center developer under SoftBank Group, led by Japanese investment legend Masayoshi Son
Energy is considering delaying its initial public offering until as late as next year. The company is building a large artificial intelligence data center in Ohio, USA, for OpenAI.
Tech media The Information reported on the 22nd that SB
Energy is considering delaying its IPO until December this year or next. The media previously reported in August that the company aimed to go public as early as September, indicating a significant delay in the listing timeline.
This shift reflects a rapid cooling in investor interest in IPOs for data center and AI cloud computing companies. Recent volatility in cloud computing stocks, coupled with soaring yields on certain CoreWeave corporate bonds, has prompted other data center developers to reconsider their IPO plans. Australian cloud service provider Firmus has already withdrawn its IPO plans earlier this month.
The report states that SB
Energy filed its prospectus on the 2nd of last month and submitted a revised version on the 21st of the same month. Companies typically begin roadshows for IPO investors about two weeks after publicly filing their prospectus with the U.S. Securities and Exchange Commission. However, when market conditions become volatile, the preparation period between filing and actual issuance may be extended.
SB
Energy is seeking to finance its Ohio data center through an IPO. The company plans to raise $5 billion to $7 billion (approximately KRW 6.61 trillion to KRW 9.25 trillion) to build the facility, which will be leased by OpenAI. The data center will utilize NVIDIA chips, with a planned capacity of 8.5 gigawatts, and will be supported by a $105 billion credit guarantee provided by NVIDIA.
The investment bank's valuation of SB Energy is approximately $60 billion, equivalent to about 79.35 trillion Korean won. However, the
Information says that IPO investors have expressed difficulty accepting this valuation, as it is entirely based on projected future revenues, despite the company having no data centers yet operational.
Ohio data center accounts for SB
Energy has secured the majority of the $430 billion in contracted data center revenue. The company plans to phase in operations of the facility starting in 2028 and will generate income through lease agreements lasting 15 to 20 years following operational commencement. In contrast, SB
Energy's revenue for the first half of this year was $139 million, which annualizes to $278 million. At a $6 billion valuation, this corresponds to a price-to-sales ratio exceeding 200x. A higher price-to-sales ratio indicates that the company is being valued more highly relative to its current revenue.
The Information, citing a person familiar with the matter, said SB
Energy executives are closely monitoring Anthropic’s anticipated IPO next month. Anthropic is viewed as the first AI startup to go public; if its stock performs well after listing, SB
Energy may begin issuance as early as December.
$60 billion valuation meets cold reception: AI funding pressure is spreading from the bond market to the IPO market.
SB
Energy's consideration to delay its IPO reveals a financing dilemma: massive long-term orders but insufficient near-term cash flow in AI infrastructure.
This data center developer, a subsidiary of SoftBank, originally sought to go public as early as September but is now considering delaying the IPO until December or even 2027. The planned $5 billion to $7 billion in funding will support OpenAI’s project in Ohio. Investor resistance to the company’s approximately $60 billion valuation centers on the timing of revenue realization: while the company holds around $430 billion in data center contracts, the facilities have not yet been operational, with major projects scheduled to roll out in phases starting in 2028. Capital markets are reassessing: how high a current valuation can be supported by these future rental commitments, and how much additional capital will be required before construction is complete.
This shift occurred after global assets had just experienced a prolonged interest rate shock. The yield on the 10-year U.S. Treasury had previously risen to 5.34%, a level not seen since 2002, significantly raising the discount threshold for future cash flows. However, recent market conditions have changed: on October 22, a sharp sell-off in tech stocks triggered safe-haven buying, causing U.S. Treasury yields to drop by more than 19 basis points in a single day, reaching approximately 4.97% during Asian trading on the 23rd. Thus, the current key market narrative is that while benchmark yields have retreated, credit risk premiums for companies linked to AI infrastructure and equity financing pressures remain pronounced.
The stock market has fully reflected this concern. On October 22, the Philadelphia Semiconductor Index fell 6.7%, declining more than 26% from its June closing high; the Nasdaq Composite Index dropped 2.60%. SoftBank Group also fell approximately 6.9% in Tokyo on the same day, retreating about half from its June high, with its credit default swap (CDS
Default
Swap and CDS prices had previously risen to their highest levels in nearly four years. Although SoftBank recovered most of its intraday losses on the 23rd, closing down just 0.02%, the short-term price recovery has not altered market scrutiny of its financing burden.
The credit market has provided more direct pressure indicators for SB Energy, a data center developer under SoftBank Group. On October 22, reports showed that the yield on CoreWeave’s 2031-maturity bond reached 15.76%, while its five-year CDS spread widened to 1,115 basis points; prior media reports indicated that its CDS pricing had once implied an approximate 63% cumulative five-year default probability—this probability incorporates credit risk premiums and model assumptions and should not be interpreted directly as the company’s actual likelihood of bankruptcy. SoftBank itself has also faced significantly higher financing costs: in September, it announced the issuance of approximately $11.1 billion worth of foreign currency bonds, including a 7.5-year USD bond issued at par with a coupon rate of 9.75%, with proceeds primarily used to fund its investment in OpenAI and for general corporate purposes. The secondary market yields on CoreWeave’s bonds and SoftBank’s new bond coupon rates together indicate that investors are demanding higher returns to continue financing AI expansion.
The order locks in the future; financing determines whether you can sustain until delivery.
SB
The key constraint for Energy is the timing mismatch between capital expenditures and rental income. Land, power connection, substation facilities, server rooms, and cooling systems require upfront investment; only after construction is completed and the facilities are ready for delivery can long-term leases gradually generate revenue. The total contract value of $430 billion represents the scale of business spread over multiple years, not cash reserves, nor the present value of these revenues.
Reuters Breakingviews analysis points out that only about one-tenth of the related data center contract revenue is expected to be realized within six years; this means that even if demand is genuine, the project must still secure sufficient funding before commencement and bear the accumulating financing costs during the interim period.
A $60 billion valuation divided by $278 million in annualized revenue equals approximately a 216x price-to-sales ratio.
Ratio, P/S); however, current revenue primarily comes from existing energy operations, while future valuation depends largely on data centers that are not yet operational, creating a disparity in business stages. What truly determines valuation resilience is how much cash remains after deducting construction, operational, and financing costs from future rents, and when that cash will be received. When benchmark interest rates and issuer credit spreads rise, increasing the cost of capital, investors demand lower entry valuations; delaying the IPO also postpones equity replenishment, further increasing the developer’s reliance on alternative financing channels.
The $105 billion disclosed by Huyang is a conditional guarantee cap related to designated lease and power payment obligations corresponding to approximately 4.25 gigawatts of IT load at the Ohio site; it typically becomes effective upon the commencement of each lease term and is triggered in the event of a specific tenant default; it has not been delivered to SB.
The cash from Energy does not cover the full cost of the project. The planned scale of 8.5 gigawatts mentioned in the report does not align with the officially disclosed guaranteed scope; gigawatts measure electrical capacity and cannot directly represent model computational performance.
SB
The adjustment in Energy's listing plan highlights a growing distinction within the AI supply chain between "having orders" and "having the capital to fulfill orders." The more secure the financing arrangement and the closer the project is to power delivery, the higher the certainty that procurement demands will translate into revenue from chips, power equipment, and data centers. If Anthropic successfully lists, it will test market acceptance of AI companies’ growth and valuation, and improve the fundraising environment for subsequent issuances; but SB
Energy must ultimately convince investors to pay today’s price for future rent payments based on the project’s economic viability and financing structure.