Two decisions by SoftBank Group founder Masayoshi Son have recently sparked considerable discussion in the US capital markets.
On November 11, SoftBank Group released its latest financial report, disclosing that the company had sold all of its Nvidia shares in October. SoftBank cashed out $5.83 billion in shares. SoftBank explained that this move was not due to a pessimistic outlook on Nvidia 's long-term prospects, but rather a strategic asset allocation need. The announcement triggered significant volatility in US tech stocks, with selling pressure spreading rapidly after the market opened, causing a noticeable decline in the share prices of several tech companies, including Nvidia .
However, more than the divestment of its Nvidia shares, many investors are paying closer attention to another SoftBank decision – a large-scale investment in OpenAI. SoftBank Group revealed in its earnings call that it will invest an additional $22.5 billion in OpenAI in December. Some analysts point out that SoftBank's eagerness to cash out its Nvidia shares is precisely to raise funds for this substantial investment.
Many market participants view SoftBank's move to divest its stake in Nvidia and invest in OpenAI as extremely risky. Several industry experts have explicitly expressed their concerns: as a driving force in global artificial intelligence... While OpenAI is a well-known startup in the AI investment boom, its future is uncertain. If the company's performance deviates from market expectations, it could very well become a "time bomb" that detonates in the US capital market and AI industry.
In November 2022, OpenAI released ChatGPT, a large language model that showcased the enormous potential of AI technology through a novel interactive approach, reshaping people's imagination of the AI industry and sparking a global AI investment boom. US technology analyst Scott Galloway stated that over the past two years, approximately 80% of the overall gains in the US stock market have been contributed by the artificial intelligence sector.
As OpenAI continues to release new large-scale models, global tech giants are vying to invest. In just three years, this previously unknown company has seen its valuation soar to over $500 billion, and its market influence rivals that of tech giants.
If we only look at the performance of the large models released by OpenAI, investing in it seems justifiable. However, OpenAI's aggressive development strategy and cooperation model are causing increasing doubts among investors.
In September of this year, NVIDIA and OpenAI announced plans to deploy an AI data center with a capacity of at least 10 gigawatts through a strategic partnership. Under the agreement, Nvidia committed to investing $100 billion in OpenAI in installments and receiving corresponding equity returns. Once the two parties formally sign the procurement agreement, Nvidia will pay the first $10 billion investment, while OpenAI will purchase chips from Nvidia in cash.
Besides Nvidia, OpenAI and Oracle AMD, Broadcom The cooperation models of these companies are also very similar: tech giants provide investment or financing guarantees to OpenAI, and OpenAI then uses the funds raised to repurchase the products or services of these companies.
On the surface, the collaboration appears to be a win-win situation for all parties involved—OpenAI gains funding and computing power, while tech giants receive equity, orders, and growth prospects. However, many investors are wary of this "circular financing" model. After all, the figures on paper ultimately need to translate into tangible results, requiring a rapid increase in productivity and commercial returns from AI to match market expectations. But this chain is extremely fragile and subject to numerous uncontrollable factors, making it fraught with uncertainty.
The lessons of the past are still fresh. Before the bursting of the dot-com bubble in the United States in 2000, major U.S. telecommunications companies, based on optimistic expectations for future traffic, frantically deployed fiber optic cables and base stations, constantly borrowing to expand. To maintain revenue growth, some equipment vendors also provided funds to customers to purchase their own equipment. As a result, market demand growth fell short of expectations, much of the infrastructure became junk assets, and many companies consequently incurred heavy bad debts, ultimately ending in bankruptcy.
In response to the criticism, OpenAI CEO Alter... Mann responded multiple times. He first declared that achieving profitability was "not one of his top ten priorities," and that OpenAI was currently focused on scaling rather than short-term financial returns, with the high investment laying the foundation for the next technological revolution. He then stated that OpenAI's monetization capabilities exceeded market expectations, with revenue exceeding $20 billion this year and projected to grow to hundreds of billions of dollars by 2030.
But once investors examine OpenAI's financial statements, they find it difficult to feel at ease. (According to the Wall Street Journal, based on Microsoft... ) Financial reports and calculations based on OpenAI's shareholding ratio indicate that OpenAI's quarterly loss may exceed $12 billion. Even if OpenAI's annual revenue exceeds $20 billion, this figure wouldn't offset the half-year loss.
Faced with severe financial challenges, OpenAI has already committed to investing a staggering $1 trillion in computing power. To match such a massive investment, OpenAI must achieve commercial returns at an astonishing pace. It must be recognized that while the long-term potential is promising, achieving rapid growth and substantial commercial returns in AI applications is far from straightforward. This is not something that can be determined by the products and technologies of one or a few tech companies; it will also be influenced by multiple factors, including the macroeconomic environment. Therefore, given the many unknowns in the market, Altman's prediction of achieving hundreds of billions of dollars in revenue by 2030 seems more like empty promises than a long-term vision.
Looking further ahead, in the field of AI applications, will OpenAI ultimately be the winner? A temporary lead is by no means a permanent one. Once OpenAI's development encounters setbacks and it loses in market competition, the highly intertwined relationship between it and tech giants will instantly turn into a constraint, causing severe damage to the company and even the entire AI industry.
Harbour Global Securities Analyst Jay Goldberg points out that OpenAI's collaboration with tech giants is essentially mortgaging future demand, and the downside risks will be amplified once the market cycle reverses. Scott Galloway also stated, "If OpenAI's development encounters setbacks, the severity of the market downturn will leave investors with nowhere to hide."
Scott Galloway's statement serves as a significant reminder to investors. There has long been a prevailing view that tech giants, with their deep pockets, wouldn't be severely impacted even if their investment in OpenAI failed. However, the reality is likely far less optimistic. OpenAI's massive funding round has already significantly influenced the performance expectations and production capacity plans of numerous tech giants. For example, if the OpenAI-Nvidia partnership agreement is fully implemented, its scale is roughly equivalent to Nvidia's revenue in the first half of this year. The stock performance of Nvidia and other tech companies has largely reflected this expectation.
It's also important to recognize that this wave of AI investment has put immense pressure on the cash flow of many US tech companies. To alleviate this pressure, tech giants have widely begun issuing bonds on a large scale. Data shows that in September and October alone, US tech companies issued $75 billion in bonds, and this figure is expected to exceed $200 billion for the whole year. Previously financially sound and with ample cash flow, these tech giants are rapidly accumulating debt leverage in a short period. If investment returns fall short of expectations, this high leverage will backfire.
Such concerns are increasingly drawing the attention of professional institutions. Oracle disclosed in September that it was launching an $18 billion investment-grade bond issuance. In November, Barclays... bank Oracle 's debt rating has been downgraded, with a warning that "Oracle's capital expenditures to fulfill its AI contracts have exceeded the limits of its free cash flow, forcing it to rely on external financing." Even if capital expenditures do not increase further, Oracle's cash reserves may be depleted by November 2026.
Another issue to consider is Masayoshi Son's investment acumen, which has been heavily criticized in recent years: his investment in co-working startup WeWork ultimately resulted in losses of tens of billions of dollars and the bankruptcy of the invested company; his sale of Nvidia stock in 2019 caused him to miss out on $150 billion in potential profits, effectively letting a "duck that was within his grasp" fly away. Will Son's recent sell-off of Nvidia and heavy investment in OpenAI be another major misjudgment?
No one can foresee the future, but the risks are already too great to ignore.
(Source: Economic Daily)
