When SoftBank began accumulating stakes in OpenAI and supporting associated infrastructure projects, few observers imagined that the main obstacle would not be technological, regulatory or competitive. However, the slowdown in discussions aimed at raising at least $6 billion via a loan guaranteed by its stake in OpenAI reminds us of a more fundamental reality: the artificial intelligence revolution is now entering a phase where the question of financing becomes as important as that of innovation.
According to Bloomberg, SoftBank has suspended discussions with several potential creditors to set up this financing. A few weeks earlier, the Japanese group had already lowered its initial target from 10 billion to 6 billion dollars in the face of reluctance from lenders. This caution is surprising even though OpenAI has just submitted its IPO application, a step which should have helped to reassure the markets.
The contrast is all the more striking since SoftBank has today once again become Japan’s largest market capitalization ahead of Toyota. Its stake in ARM benefited fully from the surge in values linked to artificial intelligence. The group also displays significant unrealized capital gains on OpenAI. However, despite this favorable context, transforming these assets into several billion dollars of debt seems more complex than expected.
OpenAI participation remains difficult to promote
The heart of the problem lies in the very nature of the asset used as collateral. A margin loan allows an investor to borrow by pledging a financial asset. The amount granted by creditors generally corresponds to a fraction of the estimated value of the collateral. This mechanism is common for listed stocks. Banks can constantly monitor their market price, measure their liquidity and adjust their risks.
Except OpenAI does not meet any of these criteria, and despite its valuation of several hundred billion dollars, the company remains private. Its price is only known through successive financing rounds or limited secondary transactions. Creditors must therefore rely on estimates rather than on a liquid and transparent market.
This difficulty is reinforced by the particular structure of the company. OpenAI looks neither like a typical listed technology company nor a traditional financial asset. Governance mechanisms, economic rights associated with investors and constraints specific to the organization make the valuation exercise more complex for lenders.
So SoftBank’s OpenAI stake is probably one of the most sought-after assets in the global technology market, but also one of the most difficult to use as bank collateral.
The IPO does not remove uncertainty
The announcement of the filing of the OpenAI IPO file nevertheless modified the discussions. Until now, banks were faced with an uncertain horizon; now creditors can anticipate the appearance of a market price, a float and a capacity for progressive resale of securities.
But this does not seem to resolve all the questions, the banks must still evaluate the effective valuation which will be retained during the introduction, the depth of the market after the listing and the stability of the title in the months which follow. The excesses observed during previous technological waves have left lasting marks on credit teams.
The experience of WeWork or certain overvalued IPOs reminds us that the private valuation of a company is not always synonymous with achievable value on the public markets and has scolded more than one.
Competitive risk becomes more visible
Lenders’ hesitation also reflects a shift in the competitive landscape of artificial intelligence. Two years ago, OpenAI appeared to be the almost undisputed winner of the race for generative models. Since then, the situation has become more complex.
Anthropic has established itself as a credible competitor to large companies. Google continues to invest heavily in Gemini. Meta accelerates its developments. xAI continues its rise in power. Several specialized players are emerging in vertical segments.
For a venture capitalist, this competition can be seen as a driver of innovation, but for a creditor, it represents above all an additional risk factor.
The future value of OpenAI’s stake directly depends on its ability to maintain its technological lead, maintain its market share and transform its growth into sustainable financial flows, all parameters difficult to project on the scale of a decade.
The real issue lies elsewhere
The slowdown in this financing also comes in a particular context for SoftBank, the group must face the expiration of bridge financing of 40 billion dollars used to support its investments in OpenAI. Except this debt matures in March 2027.
The operation currently being studied therefore does not constitute a simple opportunistic refinancing, but is part of a broader strategy aimed at securing the financial resources necessary to pursue Masayoshi Son’s ambitions in artificial intelligence.
SoftBank participates in discussions around Stargate in the United States. The group also plans to invest up to 75 billion euros in data center infrastructure in France. Through ARM, it remains exposed to the entire supercomputing value chain.
Each new initiative mechanically increases financing needs.
Artificial Intelligence Becomes a Profitable Industry
After having been mainly financed by venture capital, the artificial intelligence industry is entering a new phase where its capital needs are closer to those of infrastructure than to those of software. Building a multi-gigawatt data center campus requires tens of billions of dollars of investment. Deploying hundreds of thousands of specialized processors, securing access to energy and developing associated networks now require amounts comparable to those committed to the electricity, telecommunications or transport sectors.
AI players are thus entering a universe historically dominated by banks, insurers, infrastructure funds and pension funds.
The next AI battle will be played out in financial markets
If the slowdown in the loan guaranteed by SoftBank’s OpenAI participation does not call into question the company’s prospects or those of the artificial intelligence market, it nevertheless constitutes an important signal.
The next decade could see the emergence of a different category of winners than that observed so far. Alongside OpenAI, Anthropic or NVIDIA, the real beneficiaries of the AI revolution could also be the institutions capable of organizing and financing the trillions of dollars of investments necessary for its deployment.
