Key Takeaways
- SoftBank’s stock has pulled back after enthusiasm for Masayoshi Son’s expansive AI strategy drove a sharp rally.
- Continued growth in AI infrastructure supports the investment thesis, but high capital requirements and concentrated valuations increase the downside risk.
- Arm, Nvidia and OpenAI connect SoftBank to several layers of the AI market, leaving the group exposed to both infrastructure demand and valuation shifts.
SoftBank founder Masayoshi Son’s sweeping bet on artificial intelligence sent the Japanese conglomerate’s shares soaring. That momentum has now weakened, bringing the stock back down to earth and focusing attention on how much future AI growth is already reflected in market valuations.
The retreat does not necessarily signal that enterprise demand for AI infrastructure is fading. Quite the opposite. Global spending on AI infrastructure reached $318 billion in 2025, more than twice the level recorded in 2024, according to IDC. IDC forecasts spending of $487 billion in 2026 and more than $1 trillion by 2029.
Those projections help explain why SoftBank has positioned AI as the foundation of its next investment cycle. The group’s exposure includes Arm, whose chip architecture is used across mobile devices, cloud systems and an expanding range of data-center processors. Son has also pursued ambitions connected to OpenAI, while Nvidia remains central to the wider computing ecosystem in which SoftBank is investing.
A large and growing market can still produce painful investment losses. Entry price, financing structure and the timing of exits matter, particularly when a strategy depends on companies whose valuations assume years of rapid expansion.
Market concentration makes that tension harder to ignore. Bloomberg found that more than half of the S&P 500’s gains since ChatGPT’s launch came from 10 AI-linked companies. When a relatively small collection of stocks accounts for such a large share of market performance, a change in expectations can travel quickly through portfolios, funding markets and private-company valuations.
For SoftBank, that could affect more than its listed holdings. A broad technology sell-off may reduce the value of future exits, complicate fundraising and raise the cost of capital for infrastructure projects. Private investments can appear insulated from daily market swings, but comparable-company valuations eventually influence new funding rounds and asset sales.
The spending mix creates another layer of exposure. Servers represented about 98% of the $86 billion global AI infrastructure market in Q3 2025, while cloud and shared deployments accounted for more than 86%, according to IDC. This concentration suggests that much of the current boom is tied to expensive computing hardware and large-scale cloud environments rather than a broadly distributed set of assets.
That said, capacity constraints could keep investment elevated even during periods of weaker technology valuations. Service-provider procurement for AI infrastructure is forecast to climb from $401 billion in 2025 to $1.45 trillion annually by 2030. Data-center facilities capital expenditure is projected to rise from $276 billion to $655 billion over the same broad expansion cycle.
Can demand grow fast enough to justify both the infrastructure buildout and the valuations attached to it? That is the central question for SoftBank investors. The answer may vary sharply across chips, model developers, cloud providers and applications.
Capital remains heavily concentrated geographically as well. The Stanford HAI AI Index reported that U.S. private AI investment reached $109.1 billion in 2024, compared with $9.3 billion in China and $4.5 billion in the U.K. Generative AI attracted $33.9 billion globally. That gap gives U.S.-centered companies greater access to funding, but it also leaves investors sensitive to any shift in American capital markets.
Governance will increasingly shape commercial adoption, too. Frameworks such as the NIST AI Risk Management Framework and ISO/IEC 42001 give enterprises structured approaches to model oversight, accountability and operational controls. Stronger governance can support deployment, although it can also lengthen purchasing cycles and add implementation costs.
SoftBank’s pullback therefore reflects more than a change in market mood. Son’s strategy connects the group to a genuine surge in computing demand, but also to a capital-intensive market with elevated expectations and crowded positioning. If infrastructure spending follows current forecasts, SoftBank retains substantial upside. If valuations reset faster than demand converts into durable revenue, the same concentration that powered the rally could amplify the decline.
⬇️