SOFTBANK GROUP CORP
9984
BUFFETT / MUNGER / KLARMAN SUMMARY
Price¥4089
Market CapJPY 23.3T
EVJPY 45.8T
Net DebtJPY 14.3T
Shares5,699M
SoftBank Group is a technology investment holding company whose value is driven by its ~90% stake in Arm Holdings (semiconductor IP, ~55% of NAV), its ~11% stake in OpenAI (~28% of NAV), a ~40% stake in SoftBank Corp (Japan telecom), and the Vision Fund portfolios (300+ companies, 129 unicorns). Under Masayoshi Son, SBG has pivoted to become an "AI-era industrial holding company" committing $200B+ to AI infrastructure (Project Stargate), AI chips (Project Izanagi), and Physical AI (ABB Robotics).
Revenue: JPY 7,244B
Organic Growth: 8.2%
Arm (Wide moat): 99%+ smartphone processor market share, massive switching costs, network effects from 280B+ chips shipped. Expanding into automotive and data center. Parent (Narrow): Scale advantage writing $10B+ checks, Son's network for proprietary deal flow, unique 90% Arm stake. No structural moat at holding company level - Vision Fund investing is highly competitive with many capable rivals.
CEO: Masayoshi Son (since 1981, founder)
Aggressive, visionary, high-conviction style. Brilliant early bets (Yahoo Japan, Alibaba, Arm) offset by catastrophic losses (WeWork, Wirecard, SVF1 losses). Currently in "total offense mode" committing $200B+ to AI. No succession plan. Key-man risk is extreme. Son IS SoftBank.
3.8%
Op Margin
3.7%
ROIC
JPY -651B
FCF
4.5x
Debt/EBITDA
FCF/ShareJPY -114
FCF YieldNegative
DCF RangeJPY 4,400 - 7,000
NAV-based valuation. Base case JPY 4,400-4,900/share assumes Arm at 50-60x earnings, OpenAI maintains current valuation, net debt of ~JPY 5-7T at holding company level. Bull case JPY 6,000-7,000 assumes AI revolution delivers fully. Bear case JPY 2,100-2,600 assumes AI winter with 40-50% decline in key asset values.
| Kill Event |
Severity |
P() |
E[Loss] |
| AI Winter / Arm valuation collapse (P/E from 80x to 30x) |
-50% |
20% |
-10.0% |
| OpenAI competitive disruption / open-source wins |
-30% |
25% |
-7.5% |
| Credit crisis / forced asset sales at distressed prices |
-60% |
10% |
-6.0% |
| Masayoshi Son incapacitation / departure (key-man risk) |
-30% |
10% |
-3.0% |
| Project Stargate/Izanagi execution failure (>$100B committed) |
-25% |
30% |
-7.5% |
Tail Risk: The true tail risk is correlated AI collapse: Arm, OpenAI, and Vision Fund portfolio all declining simultaneously while JPY 18.9T in debt remains fixed. During 2022, SoftBank lost JPY 1.7T in a single year. An AI winter could produce losses of JPY 5-10T, potentially threatening solvency if LTV exceeds 35% and triggers margin calls. Sub-investment-grade credit rating (BB+/Ba2) means limited refinancing flexibility in a crisis.
Downside Case
In an AI winter scenario, Arm's market cap drops from $130B to $50-60B (still a premium business, but at 30x earnings instead of 80-100x). OpenAI's private valuation gets marked down to $200B from $750B+ as competition intensifies and unit economics disappoint. Vision Fund portfolios take another $5-10B in impairments. SoftBank's NAV falls to JPY 12-15T while debt remains at JPY 18.9T. Stock falls to JPY 2,000-2,600 range. LTV exceeds 30%.
Why Market Wrong
The market may be underpricing the probability that AI is truly the next internet-scale revolution and that SoftBank's 90% Arm ownership is an irreplaceable asset. If Arm's data center and automotive businesses scale as Son envisions, Arm alone could be worth $300B+ by 2030. The NAV discount at current prices may represent a rare opportunity to buy the AI revolution at a discount through a holding company structure.
Why Market Right
The market may correctly price SoftBank at roughly NAV because: (1) the holding company adds no value - just leverage and risk, (2) Son's track record includes as many disasters as successes, (3) the $200B+ in new commitments require funding that will dilute or leverage existing shareholders, (4) OpenAI's $750B valuation may prove unsustainable, and (5) credit ratings are sub-investment-grade for good reason.
Catalysts
Positive: ARM earnings acceleration, OpenAI IPO or value crystallization, Project Stargate progress, credit upgrade to investment grade, buyback execution ($30-40B over 3-4 years). Negative: AI winter, Arm earnings miss, OpenAI competition from open-source, rising rates increasing debt costs, Son health or succession issues.
C
Rejected
Strong Buy¥2800
Buy¥3500
Sell¥6500
At JPY 4,089, SoftBank trades at roughly its estimated base-case NAV with essentially no margin of safety for a sub-investment-grade holding company with 80%+ concentration in two AI assets, negative FCF, $200B+ in new commitments, 65% annualized volatility, and extreme key-man risk. The stock would become interesting below JPY 3,500 (meaningful NAV discount) and compelling below JPY 2,800 (pricing in AI bear case). This is a speculation, not an investment. For AI exposure, buying Arm directly eliminates holding company risks.
SoftBank Group (9984) - Ultrathink
The Real Question
The real question with SoftBank Group is not whether AI is transformative -- it almost certainly is. The real question is whether Masayoshi Son's approach to capturing AI value will reward shareholders or destroy them.
Son's track record is a paradox that defies simple categorization. The same man who turned a $20M investment in Alibaba into $60B+ also poured billions into WeWork at a $47B valuation, watched Sprint hemorrhage cash for years, and presided over JPY 3T in cumulative Vision Fund losses from 2022-2024. He is simultaneously one of the greatest venture investors in history and one of its most reckless allocators of capital. The question is not which Son will show up -- both Sons always show up. The question is whether the wins will outweigh the losses on a probability-weighted basis.
Buffett has said he avoids businesses that require a genius to run them. SoftBank does not merely benefit from Son's genius; it is constitutionally incapable of existing without it. There is no SoftBank investment thesis that does not begin and end with faith in one man's judgment. That is not investing. That is hero worship with a ticker symbol.
Hidden Assumptions
When you buy SoftBank at current prices, you are implicitly assuming:
Arm's valuation is sustainable at 80-100x earnings. This requires Arm to execute flawlessly on its expansion into data center, automotive, and IoT while maintaining 99%+ mobile market share. Any competitive threat from RISC-V, any pricing pressure, any execution stumble, and a ~$130B market cap could halve.
OpenAI's $750B private valuation is real. OpenAI generates perhaps $10-15B in annual revenue. You are paying approximately 50-75x revenue for a company that faces existential competition from Google, Meta, Anthropic, and the open-source community. If open-source models reach 90% of GPT-5's capability at 1% of the cost -- a scenario many AI researchers consider plausible -- OpenAI's moat evaporates.
SoftBank can fund $200B+ in new commitments without catastrophic dilution. Son has committed to Project Stargate ($100B), Project Izanagi ($100B), the ABB Robotics acquisition ($5.4B), and an additional $30B in OpenAI. The math does not work without massive new debt, monetizing Arm shares, or attracting partner capital. Each path carries significant risk for existing shareholders.
The next credit crisis will not happen during SoftBank's heavy investment phase. With sub-investment-grade debt (BB+/Ba2), JPY 18.9T in total obligations, and declining cash reserves (JPY 3.7T, down from JPY 6.9T), SoftBank is vulnerable to a liquidity squeeze. The company has historically navigated crises by selling assets -- but its portfolio is now concentrated in illiquid (OpenAI) or market-sensitive (Arm) positions.
Masayoshi Son's judgment is more right than wrong over the next decade. At 68, with no succession plan, the company's future is inseparable from his health and decision-making.
The Contrarian View
The contrarian bull case is more interesting than it first appears. Consider: SoftBank owns 90% of the company that designs the brains of essentially every smartphone, an increasing share of data center chips, and the future of automotive computing. Arm is not just a company; it is infrastructure. It is to the chip industry what Windows was to PCs in the 1990s -- except Arm faces even less competition, charges even lower royalties relative to value delivered, and has an even larger addressable market.
If you believe that AI will indeed transform the world -- not just chatbots, but autonomous vehicles, robotics, edge computing, and scientific discovery -- then Arm's architecture will be embedded in trillions of devices running AI inference at the edge. The royalty per chip is rising from pennies to dollars. The addressable market is expanding from phones to everything with a processor. At $3.5B in revenue, Arm may be capturing less than 5% of the value it ultimately creates.
Son's grand vision -- controlling the semiconductor architecture (Arm), the AI models (OpenAI), the physical infrastructure (Stargate), the custom chips (Izanagi), and the robotics (ABB) -- is either megalomaniacal or genuinely visionary. If even a fraction of this vertical integration works, SoftBank becomes the Berkshire Hathaway of the AI age: a holding company whose disparate parts create more value together than separately.
The NAV discount, if it widens to 30-40% in a market downturn, could represent a genuine opportunity to buy the AI revolution at fire-sale prices through a structure that, for all its flaws, gives you exposure to assets no other public equity can replicate.
Simplest Thesis (One Sentence)
SoftBank Group is a leveraged bet on Masayoshi Son's ability to position the company at the center of the AI revolution through Arm and OpenAI, but at current prices offers no margin of safety for the extreme risks involved.
Why This Opportunity Exists
The NAV discount exists because the market correctly recognizes that:
- The holding company structure adds complexity, leverage, and key-man risk without adding value
- Son's capital allocation history includes catastrophic losses alongside brilliant wins
- Net income is meaningless as a metric -- it swings from -JPY 1.7T to +JPY 3.2T based on paper gains
- The stock is too volatile (65% annualized) for most institutional investors
- Sub-investment-grade credit ratings exclude it from many fixed-income and equity mandates
- Three consecutive years of net losses (FY2022-2024) damaged confidence
The market is not mispricing SoftBank out of ignorance. It is pricing in a rational assessment of the risks. The question is whether the market is too pessimistic about the probability and magnitude of the upside scenario.
What Would Change My Mind
I would seriously consider buying SoftBank if:
- Price falls below JPY 2,800 (40% discount to base NAV), providing genuine margin of safety for the risks
- Son announces a credible succession plan, reducing key-man risk
- SoftBank achieves investment-grade credit rating (BBB-/Baa3 or higher), indicating improved financial discipline
- OpenAI IPOs or provides transparent financials, allowing proper valuation of the second-largest NAV component
- Free cash flow turns positive at the consolidated level, indicating the heavy investment phase is yielding returns
- The NAV discount widens to 35-40% during a market dislocation, creating a genuine Klarman-style margin of safety
Any three of these conditions would shift my assessment from WAIT to ACCUMULATE.
The Soul of This Business
SoftBank Group has no soul in the way that a great operating business does. It is not Costco's relentless focus on member value, or Danaher's continuous improvement machine, or Arm's quiet excellence in chip design. SoftBank Group is a vessel for one man's vision of the future.
That vision is currently: artificial superintelligence will arrive within a decade, and the company that controls the compute architecture (Arm), the intelligence layer (OpenAI), the physical infrastructure (Stargate), and the embodiment (robotics) will become the most valuable entity in human history.
This may be correct. It may also be the same type of grandiose thinking that led Son to value WeWork at $47B.
The difference between investing and speculating, as Graham taught us, is the margin of safety. At current prices, SoftBank offers the speculator an intriguing vehicle to bet on AI's future. It does not offer the investor the margin of safety required to sleep well at night. And as Munger would say, the first rule of compounding is to never interrupt it unnecessarily. Owning a leveraged, sub-investment-grade holding company concentrated in two correlated AI assets and run by a 68-year-old visionary with no succession plan is a fine way to interrupt compounding.
Wait for the price. Wait for the margin of safety. If AI is truly as transformative as Son believes, there will be other ways to profit from it without taking on SoftBank-level risk. And if the market ever offers SoftBank at a deep enough discount, the risk-reward might finally tip in the patient investor's favor.
SoftBank Group Corp (9984.TSE) - Investment Analysis
Date: 2026-02-27
Currency: JPY unless otherwise stated
Price at analysis: JPY 4,089
Executive Summary
Thesis (3 sentences): SoftBank Group is a JPY 23.3T technology investment conglomerate whose value is overwhelmingly determined by its 90% stake in Arm Holdings (JPY 16T at current prices) and its massive $34.6B bet on OpenAI. Under Masayoshi Son's visionary-but-volatile leadership, the company has pivoted from a diversified telecom/VC holding company to an "AI-era industrial holding company," committing to $500B+ in AI infrastructure (Project Stargate) and $100B in a proprietary AI chip venture (Project Izanagi). While trading at a significant discount to its sum-of-parts NAV, the company's extreme concentration risk, leverage, and history of spectacular losses make it unsuitable for conservative value investors.
| Key Metric |
Value |
| Market Cap |
JPY 23.3T |
| Enterprise Value |
JPY 45.8T |
| P/E (Trailing) |
7.6x |
| P/E (Forward) |
13.8x |
| P/B |
1.51x |
| EV/EBITDA |
31.2x |
| ROE (TTM) |
26.0% |
| ROE (FY2025 Reported) |
~10.0% |
| Operating Margin |
3.8% |
| D/E Ratio |
1.64x |
| Dividend Yield |
~0.27% |
| LTV Ratio |
~18-20% |
| 52-Week Range |
JPY 1,433 - 6,924 |
Business Description
What SoftBank Is
SoftBank Group Corp is fundamentally a technology investment holding company headquartered in Tokyo. Despite being classified under "Telecom Services," the parent company's value and earnings are almost entirely driven by its investment portfolio, not operational businesses.
How It Makes Money
SoftBank Group operates through five reporting segments:
- Investment Business of Holding Companies - Direct investments by the parent, most notably the ~90% stake in Arm Holdings (semiconductor IP design) and the growing stake in OpenAI
- SoftBank Vision Funds - SVF1 (
$100B fund, largely wound down) and SVF2 ($56B, SBG's own capital), investing in late-stage tech/AI companies. 300+ portfolio companies, 129 unicorns
- SoftBank Corp (Telecom) - ~40% owned listed subsidiary (TSE: 9434) providing mobile, broadband, and enterprise services in Japan. The stable cash cow generating ~JPY 500B+ annual operating income
- Arm - ~90% owned, listed on NASDAQ. Designs semiconductor IP (microprocessor architecture) licensed to virtually all smartphone makers. Revenue ~$3.5B, growing 25%+ YoY
- Other - PayPay (fintech), Yahoo Japan, media, baseball team
Revenue Decomposition
Consolidated revenue of JPY 7.2T is misleading because it primarily reflects the telecom subsidiary. The real economic engine is unrealized/realized investment gains, which flow through net income but not revenue. This explains why operating margin appears low (3.8%) while net margin can swing wildly (-27.5% to +48.2%).
Net Asset Value (NAV) - The True Measure
As an investment holding company, SoftBank should be valued on NAV, not operating metrics:
| Asset |
Estimated Value (JPY T) |
% of NAV |
| Arm Holdings (~90% stake) |
~16.0T |
~55% |
| OpenAI stake (~11%) |
~8-10T |
~28% |
| SoftBank Corp (~40%) |
~3.5T |
~12% |
| Vision Fund Portfolios |
~2-3T |
~8% |
| Other (PayPay, etc.) |
~1T |
~3% |
| Gross Asset Value |
~31-34T |
100% |
| Less: Net Debt (holding co) |
~(5-7T) |
|
| Estimated NAV |
~25-27T |
|
| NAV per share |
~JPY 4,400-4,700 |
|
| Current Price |
JPY 4,089 |
|
| Discount to NAV |
~5-13% |
|
Phase 1: Risk Analysis
1. Concentration Risk (CRITICAL)
SoftBank's NAV is ~80%+ concentrated in just two assets: Arm and OpenAI. Both are in the AI/semiconductor space, making them highly correlated. A broad AI sentiment reversal would simultaneously crush both positions.
- Arm alone represents ~55% of gross NAV
- OpenAI (private, illiquid) represents ~28%
- Combined: ~83% in two AI bets
- This is extreme single-sector concentration for a company of this size
2. Leverage and Financial Risk (HIGH)
| Metric |
Value |
Assessment |
| Total Debt |
JPY 18.9T |
Very high |
| Net Debt |
JPY 14.3T |
Substantial |
| D/E Ratio |
1.64x |
Elevated |
| LTV Ratio |
~18-20% |
Within policy (<25%) |
| S&P Rating |
BB+ |
Below investment grade |
| Moody's Rating |
Ba2 |
Below investment grade |
The LTV metric is misleading because it is based on market values of assets that can decline rapidly. During the 2022 tech downturn, LTV spiked dramatically. The sub-investment-grade credit ratings tell the real story: this is a highly leveraged structure.
3. Vision Fund History (CAUTIONARY)
The Vision Fund track record is checkered:
- SVF1 ($100B): Invested in WeWork (massive loss), Wirecard (fraud), Greensill (collapsed), among others
- FY2022: Net loss JPY 1.7T driven by Vision Fund impairments
- FY2023: Net loss JPY 970B
- FY2024: Net loss JPY 228B
- Three consecutive years of losses before FY2025 recovery
4. Masayoshi Son Key-Man Risk (HIGH)
Son is 68 years old and is the visionary-architect of every major SoftBank bet. His investment style is:
- High conviction, concentrated bets
- "300-year vision" time horizon
- Willingness to leverage up for opportunity
- Track record: Brilliant (early Yahoo, Alibaba, Arm) AND catastrophic (WeWork, Sprint's early years, Vision Fund 1 losses)
There is no clear succession plan. Son IS SoftBank.
5. Execution Risk on New Ventures (HIGH)
Current commitments are breathtaking in scale:
- Project Stargate: $500B AI infrastructure (SBG pledging $100B)
- Project Izanagi: $100B AI chip venture
- ABB Robotics: $5.4B acquisition
- OpenAI follow-on: Additional $30B investment
- Total new commitments: $200B+
Funding this requires either: (a) massive new debt, (b) monetizing Arm/other assets, or (c) partner capital. Each carries significant risk.
6. AI Bubble Risk (MODERATE-HIGH)
SoftBank's entire thesis depends on AI being a transformative revolution comparable to the internet. While AI is certainly significant, valuations in the space are extreme:
- Arm trades at ~80-100x earnings
- OpenAI valued at $750B+ with limited revenue
- If AI hype deflates (as dot-com did in 2000), SoftBank's NAV could decline 50%+
7. Currency and Macro Risk (MODERATE)
- JPY weakness vs USD inflates the value of USD-denominated assets (Arm, OpenAI)
- Rising Japanese interest rates could increase debt servicing costs
- Global recession would hit tech valuations and Vision Fund portfolio
Risk Score: 7/10 (High Risk)
Phase 2: Financial Analysis
Income Statement Analysis
| FY End |
Revenue (B) |
Op Income (B) |
Net Income (B) |
Op Margin |
Net Margin |
| 2022-03 |
6,222 |
714 |
-1,708 |
11.5% |
-27.5% |
| 2023-03 |
6,570 |
633 |
-970 |
9.6% |
-14.8% |
| 2024-03 |
6,756 |
560 |
-228 |
8.3% |
-3.4% |
| 2025-03 |
7,244 |
730 |
1,153 |
10.1% |
15.9% |
Key observations:
- Revenue grew modestly (8.2% latest year) - driven by telecom + Arm revenue
- Operating income is relatively stable (JPY 560-730B range) - telecom is the anchor
- Net income is wildly volatile due to investment gains/losses. This is NOT an operating company.
- Q3 FY2025 (9 months to Dec 2025): JPY 3.17T net income, nearly 5x prior year - driven by OpenAI fair value gains
Balance Sheet Analysis
| FY End |
Assets (T) |
Equity (T) |
Debt (T) |
Net Debt (T) |
D/E |
| 2022-03 |
47.5 |
10.0 |
22.3 |
16.3 |
2.24 |
| 2023-03 |
43.9 |
9.0 |
20.3 |
12.6 |
2.25 |
| 2024-03 |
46.7 |
11.2 |
21.4 |
14.4 |
1.91 |
| 2025-03 |
45.0 |
11.6 |
18.9 |
14.3 |
1.64 |
Key observations:
- Deleveraging trend: D/E improved from 2.25x to 1.64x
- But absolute debt remains massive at JPY 18.9T
- Cash declined sharply from JPY 6.9T (2023) to JPY 3.7T (2025) - funding investments
- Net debt relatively stable at ~JPY 14.3T as debt was repaid
Cash Flow Analysis
| FY End |
Operating CF (B) |
CapEx (B) |
FCF (B) |
Dividends (B) |
| 2022-03 |
2,725 |
-835 |
1,890 |
-76 |
| 2023-03 |
741 |
-634 |
108 |
-70 |
| 2024-03 |
251 |
-623 |
-372 |
-64 |
| 2025-03 |
204 |
-854 |
-651 |
-64 |
Key observations:
- Operating cash flow declining sharply: JPY 2.7T -> JPY 204B
- FCF turned deeply negative (-JPY 651B) as CapEx grew (Arm + data centers)
- Dividends are minimal (~JPY 64B, ~0.27% yield)
- The company is in heavy investment mode, not cash-generating mode
Profitability Metrics
| Metric |
Value |
Buffett Standard |
Pass? |
| ROE (TTM reported) |
26.0% |
>15% |
Misleading - includes unrealized gains |
| ROE (FY2025 actual) |
10.0% |
>15% |
Fail |
| ROIC |
3.7% |
>10% |
Fail |
| Operating Margin |
3.8% |
>15% |
Fail |
| FCF Margin |
Negative |
>5% |
Fail |
These metrics are fundamentally misleading for an investment holding company. SoftBank should be evaluated on NAV growth and LTV management, not traditional operating metrics.
Phase 3: Moat Analysis
Moat Rating: NARROW (at holding company level) / WIDE (at Arm level)
Arm Holdings (the crown jewel) has a WIDE moat:
- Network Effects: The more devices using Arm architecture, the more developers write for Arm, attracting more device makers. 99%+ of smartphone processors use Arm designs.
- Switching Costs: Transitioning from Arm to a competitor would require years of re-engineering and billions in R&D. The entire mobile ecosystem is built on Arm.
- Installed Base: 280+ billion chips shipped on Arm architecture since inception
- Expansion: Moving into automotive, data center, IoT - each creating new switching cost moats
SoftBank Group (parent) has a NARROW moat:
- Scale Advantage: Can write $10B+ checks that few others can, accessing unique deals (OpenAI, Arm)
- Masayoshi Son's Network: Personal relationships with tech founders give access to proprietary deal flow
- Arm Ownership: The 90% stake in Arm is itself a competitive advantage - no one else has this card to play
No moat in:
- Vision Fund investing (many competitors: Tiger Global, Sequoia, sovereign wealth funds)
- Telecom (SoftBank Corp faces intense competition from KDDI, NTT Docomo)
- There is no structural barrier preventing NAV erosion in a downturn
Competitive Position
SoftBank has positioned itself as the only publicly traded company offering pure-play exposure to AI through both infrastructure (Arm chips) and applications (OpenAI, Vision Fund portfolio). This is a unique positioning but also extreme concentration.
Phase 4: Decision Synthesis
The Core Investment Question
Is SoftBank Group worth buying as a discounted way to own Arm, OpenAI, and the AI revolution, given the risks of leverage, concentration, and Son's aggressive capital deployment?
Valuation Scenarios
Bull Case (NAV JPY 35-40T, ~JPY 6,000-7,000/share):
- AI revolution delivers on its promise
- Arm grows into $10B+ revenue at 50%+ margins
- OpenAI reaches $100B valuation as leading AI platform
- Vision Fund portfolio produces several big winners
- Project Stargate/Izanagi create additional value
Base Case (NAV JPY 25-28T, ~JPY 4,400-4,900/share):
- AI growth continues but moderates
- Arm maintains premium valuation but P/E compresses from 100x to 50-60x
- OpenAI faces growing competition from open-source models
- Net debt stays manageable with LTV <25%
Bear Case (NAV JPY 12-15T, ~JPY 2,100-2,600/share):
- AI winter: Arm and OpenAI valuations decline 40-50%
- Vision Fund losses resume
- Credit downgrade forces asset sales at distressed prices
- LTV exceeds 35%, triggering margin calls
Entry Prices
| Level |
Price |
Rationale |
| Strong Buy |
JPY 2,800 |
~40% discount to base NAV, pricing in significant AI downside |
| Accumulate |
JPY 3,500 |
~20% discount to base NAV |
| Fair Value |
JPY 4,500 |
Approximate base case NAV per share |
| Sell |
JPY 6,500 |
Approaching bull case; NAV premium territory |
Position Sizing
Given the extreme volatility (65% annualized) and concentration risk, position sizing should be 1-2% maximum for aggressive investors, 0% for conservative/Buffett-style investors.
Verdict: WAIT (at JPY 4,089)
At JPY 4,089, SoftBank Group trades roughly at its estimated base-case NAV per share (~5-13% discount). This is not enough margin of safety for a company with:
- Sub-investment-grade credit rating
- 80%+ NAV concentration in two AI assets
- Negative free cash flow
- $200B+ in committed new investments requiring funding
- Extreme stock price volatility (65% annualized)
- Key-man risk with no succession plan
The stock would become interesting below JPY 3,500 (a meaningful NAV discount) and compelling below JPY 2,800 (pricing in significant AI bear case). At current prices, you are paying roughly fair value for an exceptionally risky holding company with no margin of safety.
For investors wanting AI exposure, buying Arm directly (NASDAQ: ARM) eliminates the holding company risks, leverage, and Vision Fund baggage. For value investors, SoftBank fails nearly every Buffett test: no durable competitive advantage at the parent level, no consistent earnings, no FCF generation, extreme leverage, and dependence on one man's vision.
Recommendation: WAIT for JPY 3,500 or below. This is a speculation, not an investment.
Sources