SoftBank's AI Pivot: Growth Engine Amid Restructuring and Valuation Uncertainty

===========================THE ARTICLE WITH INSERTED TAG===========================

SoftBank Corp. delivered a record fiscal year, posting ¥6.54 trillion in revenue and ¥989 billion in operating income for FY2024, marking 8% and 13% year-over-year growth respectively. This performance exceeded the company's medium-term financial targets a year early, providing a solid base for future ambitions. The strong results signal that SoftBank's core telecom and investment businesses are operating effectively as the group pivots towards high-growth AI initiatives.

Looking ahead, SoftBank has raised its FY2025 operating income target to ¥1 trillion, driven by substantial investments in artificial intelligence infrastructure. This aggressive goal reflects confidence in the strategic shift towards generative AI, including a ¥100 billion commitment to building a large-scale AI data center and developing Japanese large language models like "Sarashina mini". While the AI push offers significant upside potential, achieving this target will depend heavily on the successful deployment and monetization of these new technologies within the coming years.

Within SoftBank's portfolio, PayPay Corporation stands out as a key growth engine. The fintech and payments arm generated over ¥30 billion in operating income during FY2024, demonstrating strong profitability in its core services. The group is accelerating PayPay's development through planned restructuring and an initial public offering, which management expects will further enhance value. This focus on unlocking value from existing assets complements the bold AI bets, though the execution of both strategies simultaneously presents a significant operational challenge.

The enterprise segment also showed robust double-digit growth, highlighting demand for SoftBank's expanding technology services. Combined with the profitability of its financial segment and the rising contribution from PayPay, these results provide the revenue momentum needed to fund ambitious AI investments without external capital raising. This financial self-sufficiency gives SoftBank flexibility, though the high capital intensity of AI data centers and models means sustained profitability throughout FY2025 will be crucial to maintaining investor confidence in the turnaround plan.

AI Infrastructure Expansion

SoftBank Group's recent $40 billion commitment to OpenAI represents a significant escalation in its artificial intelligence ambitions, structured with $30 billion funded directly by SoftBank and $10 billion syndicated to co-investors.

The investment, delivered in two tranches during 2025, intentionally maintains SoftBank's loan-to-value ratio below 25% and prioritizes preserving cash reserves to cover two years of bond redemptions. This deal channels capital into OpenAI's AGI development and aligns with SoftBank's broader "Stargate Project" infrastructure goals. However, the convertible securities component are classified as Fair Value Through Profit and Loss (FVTPL), meaning quarterly earnings will now swing with valuation changes in these assets, creating near-term financial volatility. SoftBank's substantial existing debt burden means this aggressive AI bet carries significant financial risk if returns materialize slower than anticipated.

Parallel to its OpenAI partnership, SoftBank is advancing in-house generative AI capabilities, including developing a 70B-parameter Japanese language model called "Sarashina mini" and building GPU-optimized data centers. These efforts, detailed alongside record fiscal year 2024 financial results showing 6.54 trillion JPY in revenue, reflect a dual strategy of external investment and internal R&D. While the Sarashina mini model demonstrates technical ambition, competing in the rapidly evolving large language model space requires sustained heavy investment and faces intense pressure from global tech giants. SoftBank's ability to translate this technical effort into profitable, scalable products remains unproven, especially given the high costs associated with training and deploying such models at scale.

Valuation Leaps and Market Sentiment

SoftBank's recent share price surge has created a striking market-implied valuation for its OpenAI investment. The Japanese conglomerate's stock currently implies OpenAI is worth $750 billion, representing a significant premium over its last funding round. This figure comes from straightforward math: SoftBank's 11% stake in OpenAI is valued at approximately $82.5 billion based on its current share price, translating to the $750 billion total. This level is 50% higher than the $500 billion valuation reached in OpenAI's most recent funding round, highlighting intense investor optimism around the AI giant. However, this valuation calculation factors in a substantial 30% discount for SoftBank's shares being illiquid, meaning the underlying OpenAI asset value could be even higher if traded freely. Masayoshi Son's personal investment history reinforces this confidence; he has already committed $10.8 billion to OpenAI at prior $300 billion and $500 billion valuations, with plans to add $22.5 billion more, signaling deep conviction in the technology's long-term potential.

This rally is not solely driven by OpenAI. Analysts point to broader AI optimism and SoftBank's significant exposure to chip designer Arm Holdings as key market drivers, indirectly boosting sentiment around all its AI bets. The combination creates a powerful feedback loop where strength in one area (like Arm) lifts the entire AI-focused investment thesis, including SoftBank's OpenAI position. Yet, this lofty implied valuation carries significant risks that warrant caution. OpenAI itself faces challenges, including rising cash burn rates and the absence of a clear timeline for monetization or an initial public offering. The $750 billion figure remains highly theoretical, detached from immediate cash flows or proven business models. The very illiquidity discount factored into the calculation underscores the difficulty in realizing this value soon, if ever. Until OpenAI demonstrates a credible path to significant revenue and profitability, this market-implied valuation stretch, especially given the limited near-term monetization window, remains a major point of uncertainty for SoftBank and its investors.

Execution Risks and Long-Term Viability

SoftBank's $40 billion OpenAI investment faces immediate liquidity pressure. The firm allocated $30 billion directly from its balance sheet, keeping leverage low with a loan-to-value ratio below 25%. This preserves cash for two years of bond redemptions, but the convertible securities tied to the deal will swing quarterly profits with OpenAI's valuation changes. While the Stargate Project infrastructure bets align with SoftBank's AI ambitions, the April and December 2025 funding tranches create sequential execution risks-if either tranche falters, the entire AGI development timeline could stall.

The $750 billion market-implied value for OpenAI appears detached from fundamentals. Analysts note this figure represents a 50% jump from OpenAI's last $500 billion valuation, yet a 30% illiquidity discount applies to SoftBank's stake. The disconnect hinges on unproven revenue models; OpenAI's rising cash burn and undefined IPO schedule mean SoftBank's stock may be overvalued by hundreds of billions. Even SoftBank's Arm Holdings rally can't fully insulate investors from this mispricing risk.

Competitive threats compound these vulnerabilities. SoftBank Corp.'s record revenue growth and AI investments-including a 70-billion-parameter Japanese LLM-show internal innovation capacity. Yet OpenAI's dominance in global LLM development means SoftBank's in-house models face steep adoption hurdles. PayPay's profitability and enterprise segment gains may fund some R&D, but lagging performance against OpenAI's advanced systems could erode SoftBank's strategic relevance in generative AI.

Despite these challenges, SoftBank's diversified bets-including bond redemptions secured and Arm's gains-provide a buffer. However, without concrete monetization pathways for OpenAI assets or accelerated LLM market penetration, the $750 billion valuation premium remains a fragile construct.