Analysis of the Impact of Nscale's Large-Scale Financing on AI Infrastructure's Competitive Landscape and Industry Chain Investment

Based on the latest data and analysis I have collected, I now present to you a

.

---

Nscale is a data center company focused on AI infrastructure, spun off from Australian cryptocurrency mining firm Akrom Energy in 2023[1]. According to the latest news, the company is working with Goldman Sachs and JPMorgan to seek nearly $2 billion in financing[2][3]. Prior to this, the company completed two rounds of financing in September and October 2024, raising over $1.5 billion in total[2]. Notably, NVIDIA participated in Nscale’s $433 million SAFE financing in October 2025[1][4].

Nscale is currently building large-scale data centers in the UK and Norway for OpenAI’s “Stargate” project, making it an indispensable infrastructure partner in NVIDIA’s ecosystem[1]. Looking at its financing history, Nscale has demonstrated remarkable financing capabilities: it completed an $1.1 billion Series B financing in September 2024 (the largest AI infrastructure financing in European history), followed by a $433 million Pre-Series C financing in October[5].

Nscale’s large-scale financing reflects the continued rise of capital enthusiasm in the AI data center sector. The global AI data center market is undergoing profound changes, and is expected to grow from $12.8 billion in 2023 to $125.8 billion in 2033[6]. Against this backdrop, Nscale’s rise is reshaping the competitive landscape:

Although NVIDIA still holds over 90% market share in the GPU segment[7], it is facing challenges from multiple players. Hyperscale cloud providers have carried out vertical integration, with Amazon (Trainium 3), Google (TPU v6/Ironwood), and Microsoft (Maia series) collectively ramping up efforts in 2025[7]. According to a report by Nomura, the ASIC chip shipments of Google and Amazon in 2025 are expected to reach 40% to 60% of NVIDIA’s GPU shipments[7].

Nscale has a unique positioning—it is both an investment target of NVIDIA and the core contractor for OpenAI’s data center construction, forming a dual-binding model of “capital + orders”. This role has given it a differentiated advantage in AI infrastructure competition, avoiding direct conflicts with traditional cloud giants.

Competition in AI infrastructure has shifted from pure technical competition to system-level competition. NVIDIA continues to consolidate its dominant position with its full-stack solutions (GPU architecture, CUDA ecosystem, and high-speed interconnection technology)[7], but the focus of competition is shifting from competing for single-card peak computing power to pursuing system-level efficiency and total cost of ownership optimization in large-scale clusters. As a specialized data center operator, Nscale’s value lies in providing one-stop “computing power + energy + land” solutions for AI companies, which forms a complementary rather than substitutive relationship with traditional cloud service providers.

The Nscale financing event highlights NVIDIA’s strategic investment layout in the AI industry chain. According to statistics, NVIDIA has participated in nearly 90 investment transactions in 2025, covering multiple fields such as infrastructure, model layer, application layer, embodied intelligence, and energy infrastructure[8][9]. NVIDIA’s investments demonstrate strong precision—with itself as the core, it is building an “ecosystem moat” that covers the entire AI industry chain.

Jensen Huang divides the AI industry chain into five layers—energy, chips and systems, infrastructure and software, AI models, and applications[8]. Nscale belongs to the infrastructure layer, which is a key support point of NVIDIA’s “ecosystem moat”.

NVIDIA’s investment strategy has an obvious locking effect. The commitment to invest $100 billion in OpenAI means that OpenAI will purchase 4 to 5 million NVIDIA GPUs[10]. Similarly, in xAI’s $2 billion financing, NVIDIA plans to invest $200 million to help xAI purchase more NVIDIA equipment[10]. This strategy of “I invest in my customers’ equity, and customers place orders with me” has enabled NVIDIA to achieve synergies between investment and sales.

The Nscale financing event has the following impacts on industry chain investment:

Industry Chain Segment	Change in Investment Enthusiasm	Driving Factors
Upstream - Chips	★★★★★	Mass production of NVIDIA Blackwell, tape-out of Rubin; rise of ASIC route
Midstream - Data Centers	★★★★★	Active financing of Nscale, Crusoe, etc.; OpenAI’s Stargate project
Energy Infrastructure	★★★★☆	Highlighted power gap; focus on nuclear fusion and clean energy
Model R&D	★★★★☆	Continuous financing of OpenAI, xAI, Anthropic
Application Layer	★★★☆☆	AI Agent and edge AI become new hotspots

Nscale’s financing is a microcosm of the “arms race” in AI infrastructure. Global tech giants’ capital expenditures continue to rise:

• Microsoft:
  Capital expenditure in fiscal year 2025 reached $64.6 billion, a year-on-year increase of 45%; it is expected to reach $100-140 billion in fiscal year 2026[11]
• Amazon:
  Capital expenditure in 2025 is expected to reach $125 billion, nearly doubling compared to 2022[11]
• Google:
  2025 capital expenditure guidance range is $91-93 billion; it is expected to reach $122 billion in 2026[11]
• Meta:
  The upper limit of capital expenditure in 2025 has been adjusted to $72 billion[12]

A report from Industrial Securities pointed out that the total AI infrastructure capital expenditure of leading enterprises will reach as high as $1.4 trillion from 2025 to 2027[12].

Although NVIDIA is still the absolute leader in the AI chip sector (with a market value exceeding $5 trillion[13]), it is facing competitive pressure from multiple sides:

The ASIC route represented by Google TPU is on the rise, and the focus of competition has shifted to system-level efficiency[7]. According to a report by Semi Analysis, the total cost of ownership of TPUv7 servers is about 44% lower than that of NVIDIA GB200[14].

Cloud giants such as Microsoft, Google, and Amazon develop their own chips, and OpenAI plans to cooperate with Broadcom to develop its own AI chips[13] to reduce dependence on NVIDIA.

High-end chips are fading out of the Chinese market, and domestic substitution is accelerating[7]. According to Frost & Sullivan’s forecast, the compound annual growth rate of China’s AI chip market will reach 53.7% from 2025 to 2029, with the market size surging from 142.5 billion yuan in 2024 to 1.34 trillion yuan in 2029[7].

New entrants such as Nscale adopt differentiated strategies to avoid direct competition with traditional cloud giants:

• Specialized Positioning:
  Focus on AI data center construction, providing customized infrastructure for AI companies
• Energy Efficiency Advantage:
  Use renewable energy and clean computing to reduce operating costs
• Global Layout:
  Enter various countries’ markets through the “AI Sovereignty Initiative” to obtain policy support

The rise of companies such as Crusoe and Nscale has transformed the AI infrastructure market from “oligopoly” to “diversified competition”.

Based on the current pattern evolution, the AI infrastructure market will show the following trends in 2026:

1. Rise of Inference Computing:
   AI workloads are shifting from training to inference, and demand for high-energy-efficiency, low-cost ASIC chips will experience explosive growth, with ASIC shipments possibly surpassing GPU shipments for the first time[7]
2. Ecosystem Competition Escalates:
   Competition between GPU and ASIC has escalated into ecosystem competition, and Google is cooperating with Meta to promote the “Torch TPU” initiative[7]
3. Intensified Supply Chain Game:
   TSMC’s CoWoS advanced packaging capacity has become a key contested resource
4. Ecosystem Construction in the Chinese Market:
   2026 will be a key year for the construction of China’s AI chip ecosystem[7]

• NVIDIA industry chain (TSMC CoWoS packaging, SK Hynix HBM)
• ASIC chip design companies (Broadcom, Marvell)
• Domestic substitution leaders (Huawei, Cambricon)

• Data center operators (Nscale, Crusoe, CoreWeave)
• Energy infrastructure (nuclear fusion, clean energy)
• Liquid cooling technology, power grid management

• AI Agent platforms
• Edge AI chips
• Vertical industry applications

Nearly 80% of enterprises deploying AI have not yet achieved net profit growth[12], leading to uncertainty in the investment return cycle

The outcome of competition between GPU and ASIC routes is still unclear, and betting on the wrong technical route may cause significant losses

Intensified Sino-US technological competition may lead to supply chain disruptions or market fragmentation

The International Energy Agency estimates that the total power consumption of global data centers will exceed 800 TWh in 2026[8], and the power gap may restrict expansion

The “capital expenditure/operating net cash flow” ratio of tech giants continues to rise[12], increasing external financing pressure

Nscale’s nearly $2 billion financing event is a microcosm of the continued rise of capital enthusiasm in the AI infrastructure sector. This event has far-reaching impacts on the competitive landscape and industry chain investment:

1. Competitive Landscape:
   The AI infrastructure market is evolving from “NVIDIA’s sole dominance” to “diversified competition”, with traditional cloud giants developing their own chips, new entrants (such as Nscale) adopting differentiated layouts, and geopolitical factors jointly reshaping the market pattern
2. Industry Chain Investment:
   NVIDIA is building an industry-wide moat through its “ecosystem moat” strategy, and the “capital-order” internal cycle model locks in future demand; investment enthusiasm in all links of the industry chain continues to rise, but vigilance against valuation bubbles is needed
3. Development Trends:
   2026 will be a key turning point—rise of inference computing, escalation of ecosystem competition, intensified supply chain game, and ecosystem construction in the Chinese market

AI infrastructure has entered a fierce “arms race” stage, and the Nscale financing event marks the upgrade of market competition from single-point technical competition to systematic ecosystem competition. Investors need to pay close attention to the evolution of technical routes, changes in supply chain patterns, and geopolitical impacts, and prudently manage risks while seizing opportunities.

---

[1] Zhiyuan Community - What has Jensen Huang Invested in 2025? 50 Deals, 32 Companies Covering the Industry Chain Closed Loop (https://hub.baai.ac.cn/view/49607)
[2] Eastmoney - NVIDIA-backed Data Center Company Nscale Seeks Nearly $2 Billion in Financing (https://wap.eastmoney.com/a/202601093613299370.html)
[3] Phoenix Net Finance - AI Data Center Company Nscale Considers Raising Approximately $2 Billion (https://finance.ifeng.com/c/8plM6AAM7yi)
[4] VC Memo - Inventory of NVIDIA’s AI Investment Empire: 67 Deals in 2025 (https://vcsmemo.com/newsflashdetail/a0bf4781-9012-4543-b084-9d1c5330a295)
[5] Nscale Official Website - Nscale 2025: A landmark year for AI infrastructure (https://www.nscale.com/blog/nscale-2025)
[6] Spherical Insights - Global AI Data Center Market (https://www.sphericalinsights.com/images/rd/global-ai-data-center-market.png)
[7] PE Daily - AI Chips 2025: Giants Battle for Power Shift (https://news.pedaily.cn/202601/559632.shtml)
[8] The Paper - Crazy NVIDIA, Crazy Spending (https://m.thepaper.cn/newsDetail_forward_32186519)
[9] Jiemian News - Crazy NVIDIA, Crazy Spending (https://www.jiemian.com/article/13768802.html)
[10] QbitAI - What has Jensen Huang Invested in 2025? 50 Deals, 32 Companies Covering the Industry Chain Closed Loop (https://hub.baai.ac.cn/view/49607)
[11] Sina Finance - Increasing Investment Scale: Deconstructing the AI Scale Competition of the Seven U.S. Stock Giants (https://finance.sina.com.cn/jjxw/2025-12-14/doc-inhatxvm5227798.shtml)
[12] People’s Daily - Hidden Worries of Bubbles Amid Overseas AI Carnival (http://paper.people.com.cn/zgjjzk/pc/content/202512/15/content_30125822.html)
[13] The Paper - 2025 AI Giants: All “Villains”: Grudges, Love-Hate, and Calculations (https://m.thepaper.cn/newsDetail_forward_32334285)
[14] TMTPost - AI Chips 2025: Giants Battle for Power Shift (https://m.thepaper.cn/newsDetail_forward_32307594)
[15] Frost & Sullivan - 2025 China AI Infrastructure Market Report (https://pdf.dfcfw.com/pdf/H3_AP202509011737555566_1.pdf)