SoftBank's OpenAI & DigitalBridge Investments: AI Industry & Tech Investment Trends

Based on the latest information, I first need to clarify an important fact: According to web search data, SoftBank’s investment in OpenAI is

(not the $4 billion mentioned in your question)[1]. Additionally, SoftBank recently announced a $4 billion acquisition of data center investment firm DigitalBridge; these two investments together form the core of SoftBank’s AI strategic layout[2]. Below is a comprehensive analysis of the impact on the AI industry landscape and tech investment trends:

---

SoftBank’s $4 billion investment in OpenAI marks the AI industry entering the

. Currently, OpenAI has become a that major tech companies are competing to partner with[3]:

• Amazon
  : Considering an investment of approximately $10 billion[3]
• Disney
  : Has invested $1 billion and authorized Star Wars, Marvel, and Pixar characters for use by Sora[3]
• Oracle
  : Signed a cloud computing power contract worth up to $300 billion over approximately 5 years[3]
• Microsoft
  : Committed to investing $250 billion[4]
• CoreWeave
  : $22.4 billion contract[3]
• NVIDIA
  : Plans to invest up to $100 billion and continue supplying data center chips[3]

: OpenAI has shifted from a cooperation model relying solely on Microsoft to an ecosystem of , significantly reducing single-partner risk and enhancing bargaining power and strategic flexibility.

Global tech companies are accelerating investments of billions of dollars into AI infrastructure[3], with

becoming the core battlegrounds:

Investment Direction	Key Players	Investment Scale
Data Center Construction	SoftBank, OpenAI, Oracle	“Stargate” project up to $500 billion[3]
AI Chip Supply	NVIDIA, AMD, Broadcom	NVIDIA supplies 10 GW of GPUs to OpenAI (worth approximately $300 billion)[4]
Cloud Computing Power Cooperation	Oracle, Microsoft, Amazon	Oracle provides a $300 billion contract, Microsoft $250 billion[4]

: Computing power demand is growing exponentially: OpenAI’s computing power increased from approximately 200 MW at the end of 2023 to 2 GW at the end of 2025 (a 10x increase), while revenue grew from $2 billion to nearly $20 billion over the same period[4]. This has redefined the investment return logic of the AI industry.

By investing in both OpenAI (model layer) and DigitalBridge (infrastructure layer), SoftBank demonstrates its

:

• Model Layer
  : OpenAI continues to lead with products like ChatGPT and Sora
• Chip Layer
  : Collaborating with Broadcom to develop custom AI processors and signed a multi-year supply agreement with AMD[3]
• Infrastructure Layer
  : Layout of data center assets via DigitalBridge (managing approximately $108 billion in assets)[2]
• Application Layer
  : Partnering with companies like Disney to explore industry application scenarios[3]

: AI competition has shifted from a single model capability race to a . Enterprises with a complete “model-chip-computing power-application” chain will gain significant competitive advantages.

---

SoftBank CEO Masayoshi Son recently sold all his NVIDIA shares, cashing out $5.83 billion, and reallocated funds to AI investments in a “painful” way[2]; this

:

• Traditional Tech Stock Positions
  : SoftBank liquidated NVIDIA (despite NVIDIA’s 39% rise in 2025[5])
• Heavy Bet on AI Infrastructure
  : $4 billion acquisition of DigitalBridge + $4 billion investment in OpenAI
• Future Vision
  : Son stated that he will push SoftBank to become a “leading super AI platform provider”[2]

: Tech investment is shifting from ; investors prefer to make large-scale layouts in core positions across the entire AI industry chain rather than分散投资 multiple niche tracks.

OpenAI has committed to future expenditures totaling up to

to multiple enterprises[4]; this long-term commitment spanning 8 years has changed the traditional time frame of tech investment:

OpenAI's Future 8-Year Commitments (by Category):
- Oracle: $300 billion (cloud computing power)
- Microsoft: $250 billion (cloud services)
- Amazon: $38 billion (infrastructure)
- CoreWeave: $22.4 billion (computing power)
- NVIDIA Chips: ~$300 billion
- Custom Chips (Broadcom): ~$150 billion
- "Stargate" Data Centers: Up to $500 billion

: AI investment has transcended traditional quarterly/annual ROI assessments and entered a . Investors need to accept the new paradigm of “invest first, return later”, similar to the investment cycles of 19th-century railway construction or 20th-century telecom network deployment.

The market’s valuation logic for AI enterprises is undergoing profound changes:

• OpenAI Current Status
  : 2025 revenue ~$13 billion, weekly active users reach 800 million, enterprise customers 1 million[4]
• Valuation Target
  : IPO valuation target up to $1 trillion[4]
• Internal Forecast
  : 2030 revenue could reach $200 billion[4]

: Even if revenue grows as planned, HSBC estimates that OpenAI will still have a by 2030[4]

: AI enterprise valuation is no longer based solely on PE or PS multiples, but on:

1. Ecosystem Position
   : Whether it is at a core node of the AI value chain (e.g., OpenAI)
2. Computing Power Control
   : Whether it can obtain stable, low-cost large-scale computing power
3. Data Moat
   : Whether it has unique, hard-to-replicate training data
4. Application Scenario Breadth
   : Whether it can quickly penetrate multiple vertical industries

The market’s attitude toward AI investment has shown obvious differentiation:

(most analysts):

• Even if there is a slowdown, it won’t happen in 2026[4]
• AI is like the dial-up era of the Internet; even after the dot-com bubble, today’s usage methods are still far beyond what the most enthusiastic friends could imagine 30 years ago[4]
• From a valuation perspective, AI stocks may have bubbles, but long-term prospects are still optimistic[5]

:

• Investor Brad Gerstner questioned: “How can a company with $13 billion in revenue bear $140 billion in expenditure commitments?”[4]
• DA Davidson’s Gil Luria pointed out: “OpenAI is starting to go off track by building its own data centers, chips, and hardware, which is causing those huge commitments”[4]
• Market concerns about the AI bubble heated up in November 2025: Oracle fell 23% in a single month, AMD 15%, and CoreWeave 45%[4]

: The market has entered a ; short-term volatility has increased, but long-term funds are still continuously flowing into the infrastructure sector.

---

According to the latest market research, the global AI industry scale is expected to reach

by 2025[6], forming a :

• Algorithm Innovation
  : OpenAI (GPT series), Google (Gemini 3 won multiple industry benchmark tests[4])
• Chip Monopoly
  : NVIDIA dominates the AI computing market; AMD’s server market share increased from less than 0.5% in 2014 to 41%[4]
• Data Center Ecosystem
  : Three super cloud service providers—Microsoft Azure, Amazon AWS, and Google Cloud

• Application Penetration
  : AI leads in penetration in sectors like the Internet (89%) and finance (over 60%)[6]
• Data Advantage
  : Internet giants like Alibaba and Tencent hold massive amounts of data
• Chip Substitution
  : Enterprises like Cambricon are accelerating domestic substitution, but there is still a gap in deep learning computing power[6]

• Technological Sovereignty
  : Countries are accelerating the construction of independent and controllable AI industry chains
• Standard Setting
  : China and the US are competing in AI ethics, regulation, and technical standards
• Talent Competition
  : Global AI talent has become a strategic resource, with salaries continuing to rise

---

According to industry forecasts,

[7]:

• NVIDIA’s next-generation “Vera Rubin” architecture and “Kyber” system are expected to launch in 2027
• The 800V DC power supply (800VDC) architecture will reshape data center capital expenditure structures
• Single cabinet power density will reach an ultra-high computing environment of 1MW

:

• Focus on AI infrastructure suppliers (data centers, liquid cooling technology, power semiconductors)
• Be wary of overvalued pure concept AI enterprises
• Focus on laying out AI application enterprises with real cash flow and customer bases

:

• The 800VDC architecture transformation will spawn a new round of investment opportunities[7]
• AI Agent moves from experimentation to commercial deployment
• Large-scale implementation of enterprise AI, accelerating cloud-edge-end integration

:

• Cumulative investment in AI infrastructure may reach $500 billion[7]
• Global AI industry scale may exceed $100 billion
• The Sino-US bipolar pattern will further solidify, with technological blockades and cooperation coexisting

1. Technical Route Risk
   : If the marginal收益递减 of large models continues, it may lead to huge investments failing to generate expected returns[6]
2. Regulatory Policy Risk
   : Regulatory frameworks like the EU’s AI Act may limit the speed of AI commercialization[6]
3. Geopolitical Risk
   : Sino-US tech decoupling may split the global AI market
4. Bubble Burst Risk
   : If enterprise AI demand falls short of expectations, current infrastructure investments may face overcapacity

---

SoftBank’s $4 billion investment in OpenAI and $4 billion acquisition of DigitalBridge are

. This investment portfolio clearly demonstrates Son’s judgment on the future of AI:

For investors, key cognitive shifts include:

1. From focusing on single model capability to full-stack ecosystem layout
2. From pursuing short-term ROI to accepting long-term infrastructure investment logic
3. From technology worship to focusing on commercialization implementation capability

In the next 5 years, AI investment will present a pattern of

; giants with complete “computing power-data-application” closed loops like OpenAI, Microsoft, Google, and Amazon will continue to expand their advantages, while small and medium AI enterprises lacking core competitiveness will face the risk of being integrated or eliminated.

---

[1] Yahoo Finance - “SoftBank has completed its $40 billion investment in OpenAI” (2025-12-30) - https://finance.yahoo.com/news/softbank-fully-funded-40-billion-141537239.html

[2] Yahoo Hong Kong Finance - “SoftBank spends $4 billion to acquire DigitalBridge to seize AI infrastructure opportunities” (2025-12-29) - https://hk.finance.yahoo.com/news/軟銀斥資40億美元收購digitalbridge-搶攻ai基礎設施商機-143235559.html

[3] Yahoo Hong Kong Finance - “An AI computing power arms race is underway globally! OpenAI becomes the biggest investment magnet” (2025-12-26) - https://hk.finance.yahoo.com/news/一場ai算力軍備競賽全球上演-openai成投資最大磁鐵-121006858.html

[4] Yahoo Hong Kong Finance - “OpenAI named Yahoo Finance 2025 Company of the Year” (2025-12) - https://hk.finance.yahoo.com/news/openai當選yahoo財經2025年度企業-051905048.html

[5] Yahoo Hong Kong Finance - “Alphabet’s brilliant transformation up 65% in 2025 leading the Magnificent Seven” (2025-12-29) - https://hk.finance.yahoo.com/news/alphabet華麗轉身-2025年大漲65-領跑科技七雄-001003848.html

[6] Yahoo Hong Kong Finance - “2025 Global AI Industry Chain: Technology Competition Under Sino-US Bipolar Pattern” - https://hk.finance.yahoo.com/news/2025全球ai產業鏈-中美雙極格局下的技術競逐-092003249.html

[7] Yahoo Hong Kong Finance - “2027 Critical Point! NVIDIA’s 800V Voltage Revolution: Global Data Centers Face the Largest Infrastructure Transformation in History” - https://hk.finance.yahoo.com/news/2027年臨界點-輝達800伏電壓革命-全球資料中心面臨史上最大規模基礎設施改造-031005070.html

[8] Jinling API Data - Sector Performance Data (2025-12-30)

---