SoftBank Beat on Intel, Not OpenAI: 347 Billion Yen Profit Masks a Volatility Problem

Intel drove the beat

SoftBank's earnings beat was real, but it was also very specific. The company reported 347.3 billion yen of net profit versus 120.23 billion expected. At the same time, profit was still nearly 18% lower than a year earlier, which suggests a quarter helped by portfolio marks rather than a full operating reset.

The main catalyst was Intel. SoftBank recorded a 1.3 trillion yen gain on its Intel stake, and its investment division posted segment profit of 1.05 trillion yen. That single position did most of the heavy lifting.

ByteDance also helped, while OpenAI did not. A rise in the value of SoftBank's ByteDance stake helped offset weaker results elsewhere in the Vision Funds, and the company said it recorded no investment gain or loss related to OpenAI. In other words, the current quarter leaned on Intel, while the prior quarter had been more exposed to AI-related revaluations.

That is the core tension. When portfolio marks can shift from one mega-position to another, the earnings headline can look stronger and less durable than the underlying business.

SoftBank's earnings are uneven, and that is the risk

Repeatability is the real question

SoftBank has already shown how quickly this pattern can change. In the January-March quarter, the company reported 1.9 trillion yen of net profit, up from 517 billion yen a year earlier. That was a much larger beat, but it also illustrates the broader point: SoftBank can produce blockbuster quarters, yet that does not make such results predictable.

This quarter's beat matters because it arrived on top of a still-lower baseline, with gains concentrated in a few positions. A rise in ByteDance's value helped offset gains elsewhere in the Vision Funds, while no investment gain or loss related to OpenAI showed up. That is better described as episodic portfolio revaluation than a steady earnings engine.

Vision Funds showed the drop-off more clearly

The Vision Funds segment posted just 5.4 billion yen of profit, versus 451.4 billion a year earlier. The segment also had no OpenAI-related gain or loss this quarter. That makes the mix of winners more important, because it shows how much headline profit can depend on where private and public marks move.

Management's comment that it had not seen any new information to justify a change in the valuation of OpenAI is the watchpoint. If OpenAI's mark is not shifting, the next large gain has to come from another position instead.

That leaves public stakes with outsized influence. Intel shares have surged nearly 400% over the last 12 months, and that move helped create the 1.3 trillion yen stake gain. When that happens, results look strong. When it does not, the earnings profile can cool quickly.

What matters for the next few prints

The next test is not whether SoftBank can beat again. It is whether investors treat another surprise as repeatable or as another one-off. The recent 1.9 trillion yen net profit shows how large these prints can get. The latest Intel-driven result shows how fast expectations can still be moved.

For investors, the practical read is simple:

• Bull case: new public winners or fresh private marks keep feeding gains across the portfolio.
• Bear case: gains remain dependent on one or two positions, while core segment profit stays uneven.
• Watch items: Vision Funds profit trend, OpenAI revaluation activity, and whether Intel continues to dominate the gain mix.

Stay interested, but do not pay up on faith. The opportunity is still real if SoftBank keeps finding fresh upside across the portfolio. The clearest risk is the same as the opportunity: one-stake gains can lift the headline fast, but they can also make the earnings story less repeatable.