US AI vs China AI Brief — 2026-10-01

Today: The US-China AI contest is increasingly being fought across compute access, model-weight security, chip software, cyber-intelligence and military deployment—not just frontier models.

Top Stories

1. 💻 Tencent reportedly secures access to 100,000 advanced AI chips through Oracle

Financial Times · 2026-10-01

Bottom line: Tencent has reportedly agreed to a five-year, roughly $7 billion Oracle arrangement giving it access to about 100,000 advanced AI chips through Southeast Asian data centres.

The Financial Times reported that the chips are currently unavailable to Tencent inside China, highlighting the importance of overseas compute access as Chinese companies expand AI model development. The arrangement reportedly includes an upfront payment of about 30%, while Oracle and Tencent have not publicly confirmed the details.

Why it matters: The deal illustrates how US-origin cloud infrastructure can remain strategically important to Chinese AI developers even as Washington restricts advanced chip access to China.

🔗 Read the full story

2. 🔒 US lawmakers seek information on Chinese attempts to access AI model weights

Reuters · 2026-10-01

Bottom line: US Congressman Ro Khanna has asked leading American AI companies to disclose attempted or successful Chinese access to their sensitive model weights.

Khanna requested information from OpenAI, Anthropic, Google, Meta and SpaceXAI about efforts to obtain model weights and the cybersecurity measures protecting them. Reuters reported that there are few known cases of actual model-weight theft, although US AI companies have reported Chinese firms attempting to extract capabilities from their models through techniques such as distillation.

Why it matters: Model weights are becoming a strategic asset alongside chips and data, making cybersecurity part of the core US-China technology competition.

🔗 Read the full story

3. 🤖 DeepSeek and Huawei open-source software aimed at reducing dependence on Nvidia’s CUDA

TechNode · 2026-10-01

Bottom line: DeepSeek and Huawei are developing and open-sourcing programming tools optimized for Huawei’s Ascend AI chips as China builds an alternative AI software stack.

The collaboration centers on programming infrastructure designed to improve performance on Huawei hardware, including tools associated with TileLang. The effort is intended to make it easier for developers to use Ascend processors and reduce reliance on Nvidia’s CUDA ecosystem.

Why it matters: China does not only need alternative chips; it needs developers, compilers, libraries and tooling that can make those chips commercially competitive. A stronger software layer could therefore be as strategically important as new silicon.

🔗 Read the full story

4. 🔒 China-aligned hackers target US AI policy experts through impersonation campaigns

Proofpoint · 2026-10-01

Bottom line: Proofpoint says a China-aligned threat actor targeted US AI policy experts with phishing campaigns impersonating prominent US officials and AI-policy figures.

The security company identified the group as TA419 and said its July campaigns targeted experts at US think tanks, universities and legal organizations. The lures focused on subjects including AI policy, export controls and supply chains before directing recipients toward credential-phishing infrastructure.

Why it matters: Intelligence collection around AI policy and export controls shows that strategic competition extends beyond laboratories and chip fabs into the information surrounding national AI strategy.

🔗 Read the full report

5. 🌐 AI spending boom is lifting industrial activity across both US-linked and Chinese supply chains

Reuters · 2026-10-01

Bottom line: Global manufacturing surveys show AI-related capital spending is becoming a significant source of industrial demand across the semiconductor and equipment supply chain.

Reuters reported that September manufacturing activity benefited from strong AI-related demand, with Taiwan, South Korea and parts of Europe seeing particularly strong momentum. China’s factory activity also expanded as production recovered, while demand for AI infrastructure continued to support the wider semiconductor ecosystem.

Why it matters: The AI contest is generating economic effects well beyond model companies, creating strategic demand for chips, servers, networking equipment and data-centre infrastructure across both countries’ technology ecosystems.

🔗 Read the full story

6. 🛡️ Pentagon creates Autonomous Warfare Command to accelerate AI and robotics

Reuters · 2026-10-01

Bottom line: The Pentagon is establishing an Autonomous Warfare Command dedicated to expanding autonomous and robotic capabilities across the US military.

Defense Secretary Pete Hegseth announced the new command as part of a broader push toward AI-enabled autonomous systems and lower-cost military platforms. The initiative reflects the US military’s effort to integrate AI, drones and autonomous systems more rapidly into operational capabilities.

Why it matters: Military AI is becoming a distinct dimension of US-China competition, linking frontier software and compute advances directly to defense capabilities and procurement priorities.

🔗 Read the full story

7. 📊 AI adoption data puts China and the US at the front of enterprise deployment

Reuters · 2026-10-01

Bottom line: A new BearingPoint study found that China and the US reported the highest shares of companies with comprehensive AI implementation among surveyed major economies.

Reuters reported that 20% of Chinese companies and 18% of US companies surveyed described AI implementation as comprehensive, compared with 8% in Germany. The study also found that many companies reported positive financial results from AI even though scaling beyond pilots remains difficult.

Why it matters: The competitive question is increasingly about diffusion and commercial execution—not simply which country produces the strongest individual frontier model.

🔗 Read the full story