OpenAI and SoftBank: Masayoshi Son's Bottomless ATM Is Back in Action

If you thought forty billion dollars was the limit of one investor's generosity, you don't know Masayoshi Son. The head of SoftBank seems to have fully convinced himself that the keys to the future sit in Sam Altman's pocket. While the whole world wonders when generative AI will pay off and whether another bubble is about to burst, Son just keeps writing checks with so many zeros that an ordinary accountant might develop a nervous tic. The new round of thirty billion dollars is not just a helping hand to keep OpenAI afloat; it is a bid for total dominance in the arms race.

Last year was a real stress test for SoftBank. Raising forty billion dollars for OpenAI was no easy task, given the mixed results of the Vision Fund's earlier investments and the overall volatility of the market. The Japanese corporation had to squeeze liquidity out of every asset available to close the deal on time. But Son is an old-school player, and when he sees a target, obstacles such as a lack of free cash barely trouble him. Now additional funding has appeared on the horizon, one that should cement SoftBank's status as the main financial backer of Altman's team.

Why does OpenAI need so much money right now? The answer is prosaic and lies in the architecture of modern neural networks. Training models at the level of GPT-5 and beyond requires not just computing power but a colossal infrastructure of servers and specialized chips. Each new training iteration costs several times more than the last. Sam Altman long ago stopped playing in the startup sandbox and moved into the league of high geopolitics, where budgets are measured in percentages of the GDP of developed countries. Without a steady inflow of capital, ambitious plans to build artificial general intelligence (AGI) risk crashing into electricity bills and invoices from Nvidia.

For SoftBank, this deal is a question not only of profit but of survival as a technology leader. After the loud fiasco of WeWork and the mixed results of other projects, Masayoshi Son needs a win of historic scale. He is putting everything on red, or rather on "smart." If OpenAI really is the first to reach AGI, these seventy billion dollars in total investment will look like the best deal in human history, comparable to buying a stake in Alibaba at the dawn of the internet. If progress slows, SoftBank risks becoming the most expensive monument in the history of tech gambles.

Competitors Anthropic, Google and Meta* watch this financial downpour with barely concealed anxiety. Even with its own cloud resources, Google will struggle to compete with a company backed by such a determined investor. The AI market is finally turning into a war of attrition. Here, victory goes not only to the side with the more talented engineers, but also to the one that can afford to burn billions of dollars a month without blinking. Capital is becoming as important a "layer" of the model as the training data itself.

It is important to understand the context: Son has always loved sweeping maneuvers. His strategy of grabbing a market fast at any cost is now being applied to the most complex technology of our time. This is no longer venture investing in the usual sense; it is an attempt to buy ownership of the intelligence of the future. And by all appearances, OpenAI's leadership has nothing against such patronage, as long as it lets them ignore pressure from public markets and focus exclusively on research and scaling.

The bottom line: SoftBank is methodically turning OpenAI into a capital monopolist. Competitors will have to either find comparable sources of funding or admit defeat in the race for computing superiority. Can Sam Altman convert these billions into a real breakthrough, or are we watching the creation of the biggest financial overheating in Silicon Valley history?

*Meta has been designated an extremist organization and is banned in Russia.