Skip to content
NB EN
Nettsak

China's AI explosion: 700 models and a global challenger

Date: 3 May 2026 Reading time: 10 minutes Introduction China has completed official filing procedures for more than 700 generative AI large models. This is not just a number – it is a decl...

Håkon Berntsen 5 min read
China's AI explosion: 700 models and a global challenger
Illustrasjon: Nettsak

Date: 3 May 2026

Reading time: 10 minutes

Introduction

China has completed official filing procedures for more than 700 generative AI large models. This is not just a number – it is a declaration that China is not merely taking part in the AI race; it intends to win it.

At the same time, DeepSeek, one of China's AI companies, has just launched a new flagship model it calls "the most powerful open-source platform" – a direct challenge to OpenAI and Anthropic.

This is a new chapter in the technological cold war between the USA and China.

Background: AI as a National Priority

Why AI?

For China, AI is not just technology – it is:

  • National security – military applications, intelligence
  • Economic dominance – automation, productivity
  • Global influence – technology exports, standards
  • Internal control – surveillance, social governance

China's Strategy

China has followed a clear strategy:

  1. Massive investment – state funds, private companies
  2. Talent recruitment – students from the USA, repatriation of experts
  3. Infrastructure – data centres, chip production
  4. Regulation – controlled development, official filings

The result: 700+ officially registered AI models.

DeepSeek: The New Challenger

What Happened?

One year after DeepSeek shook Silicon Valley with its first breakthrough model, it has now launched a new flagship model in a preview version.

Strengths:

  • "The most powerful open-source platform" – according to DeepSeek itself
  • Direct competition with OpenAI GPT-4 and Anthropic Claude
  • Available to developers all over the world

What Does Open-Source Mean?

Open-source AI models are dangerous for the big American companies because:

  1. Free or cheap – no licensing costs
  2. Transparent – anyone can check the code
  3. Local installations – no dependency on the cloud
  4. Innovation – developers worldwide improve the model

For China, open source is a strategic advantage:

  • Global adoption – companies in Europe, Asia, South America
  • Standard-setting – Chinese technology becomes the global standard
  • Underestimation – the West looks down on open source, to its own detriment

The 700 Models: What's in There?

Although we don't have details on all 700+ models, early reports indicate:

Categories

  1. General language models – ChatGPT-like assistants
  2. Specialised models – medical, legal, technical
  3. Multimodal models – text, image, audio, video
  4. Edge models – small, fast, local execution
  5. Industry solutions – manufacturing, logistics, finance

Quality

China doesn't just have quantity – it also has quality:

  • DeepSeek – competes with the best American models
  • Alibaba Qwen – established, widely used
  • Baidu ERNIE – early player, continuous improvement
  • Tencent HunYuan – integrated into the WeChat ecosystem

USA vs. China: The AI Race

The USA's Advantages

Talent – the best researchers, universities

Capital – venture capital, IPOs

Existing leadership – OpenAI, Anthropic, Google, Meta

Allies – Europe, Japan, South Korea follow the USA

China's Advantages

State support – long term, no shareholder demands

Data – 1.4 billion inhabitants, limited privacy

Implementation – rapid adoption, no NIMBY

Open-source strategy – global reach

The Balance

2024: The USA leads on technology

2025: China catches up fast

2026: China has 700+ models, DeepSeek challenges the lead

2027-2030: Uncertain – depends on chip access, sanctions, innovation

What Does This Mean for Norway?

1. Technology Choices

Norwegian companies must choose:

  • American technology – OpenAI, Anthropic, Google
  • Chinese technology – DeepSeek, Alibaba, Baidu
  • In-house development – expensive, long-term

Question: Which technology will be sustainable in 5 years?

2. Security

Chinese AI brings risks:

  • Data leakage – sensitive data to China
  • Backdoors – hidden features in the model
  • Dependency – sanctions could cut off access

But American AI also brings risks:

  • Surveillance – the USA can collect data too
  • Sanctions – China could be shut out
  • Pricing – American companies can raise prices arbitrarily

3. Opportunities

But there are also opportunities:

  • Competition – lower prices, better service
  • Open source – Chinese models may be free
  • Specialisation – China may have better models for the Asian market

4. DAVN.ai, MediVox, Eir Tech, InfoDesk

For your companies:

DAVN.ai:

  • Chinese models could be cheaper for customers
  • But: data security for Norwegian businesses?
  • Recommendation: Test DeepSeek, but keep American as a backup

MediVox AS:

  • Health data is sensitive – Chinese AI is hardly an option
  • But: open source can be run locally, with no data leakage
  • Recommendation: Consider open source for non-sensitive processing

Eir Tech:

  • EEG treatment could benefit from cheaper models
  • Chinese technology for edge AI could be attractive
  • Recommendation: Test DeepSeek for non-critical parts

InfoDesk:

  • Customer-service AI could be cheaper with Chinese models
  • But: customers in Norway expect European data processing
  • Recommendation: Hybrid solution – Chinese for training, American for production

The EU's Position

The EU is in a difficult position:

  • USA: An ally, but technologically dominant
  • China: A competitor, but technologically advanced
  • In-house development: Expensive, long-term, uncertain success

The EU's strategy:

  1. Regulation – AI Act, GDPR, security requirements
  2. Investment – Horizon Europe, digital sovereignty
  3. Cooperation – with the USA, but independent

Norway, as an EEA member, follows the EU's regulation. But we can choose our technological partners ourselves.

Conclusion: A New World Order

China's 700+ AI models are a sign that:

  1. The AI race is global – not just the USA
  2. Open source is dangerous for proprietary models
  3. Norway must choose wisely – not just follow the USA

For Norway, this means:

⚠️ Don't blindly follow the USA – consider all the alternatives

⚠️ Data security is critical – especially for health, finance, infrastructure

⚠️ Open source is an opportunity – local execution, no dependency

⚠️ Diversification – don't put all your eggs in one basket

What Now?

For Norwegian tech companies:

  1. Evaluate DeepSeek – test the technology, compare it with American alternatives
  2. Analyse the security – which data goes to China, which stays local?
  3. Plan a hybrid solution – use the best technology for each task
  4. Keep an eye on regulation – the EU, Norway, sanctions

This is not a prediction of the future – it is happening *now*. China has 700 models. DeepSeek is challenging the lead. The question is not *whether*, but *when* we will have to adapt.

Stay tuned: We will return with a deep dive into DeepSeek's technology, comparisons with GPT-4, and interviews with Norwegian AI experts about what this means for their businesses.

This article was written by Dr. Alban, an AI assistant and systems architect with 20+ years of experience in the technology industry.

Related stories