Top of main content

CIO Academy - Open models, lower prices: Will China compress America’s AI moat?

Thought Leadership
Tech
CIO Academy
SaaS
AI

CIO Academy - Open models, lower prices: Will China compress America’s AI moat?

Oct 9, 2026

Highlights: A wave of low-cost, open-weight Chinese models is making AI cheaper and more accessible worldwide. As models’ performance converges and inference prices fall, pressure is building on frontier US model-layer economics. Does this mean that China’s AI models will compress America’s AI moat? The answer, in our view, is multi-layered and nuanced. We think the competitive battleground is moving beyond just the cost of AI models towards reliability, safety, compute, distribution and enterprise integration. Model pricing alone will not determine who ultimately captures the value of AI, other factors will:

- China is winning usage, not the AI economics yet. Lower prices have helped Chinese models gain a larger share of the global token volumes. They have not, however, translated into gaining a corresponding share of model spend. Moreover, today’s pricing may prove difficult to sustain

- Enterprise buyers will pay for trust. Models’ safety, reliability and governance increasingly matter alongside performance. AI guardrails, auditability, traceability, telemetry and constrained generation are more important for enterprises than the lowest token price

- It’s the AI ecosystem that’s the real moat, not the model. Models work within a broader stack of compute, platforms, distribution, security, governance and integrations. US AI model companies increasingly own the enterprise relationship. Switching providers can mean retraining employees, rebuilding integrations & repeating compliance reviews — all big switching costs

- Compute remains America’s structural advantage. Access to advanced chips, manufacturing capability and leading-edge AI infrastructure gives US labs greater capacity to train and improve frontier models. Supply of memory and compute remain constrained in China. US export controls and China’s limited access to advanced chipmaking tools reinforce that advantage for the US

- Agentic AI could widen the gap again. As ‘enterprise AI’ moves from answering questions to executing multi-step workflows with AI agents, demand for more capable models, greater compute, and stronger reliability and AI safety of models will rise. That favours the US frontier labs—provided they can convert technical leadership into dependable enterprise products

- Geopolitics will shape commercial adoption. Chinese models may face restrictions in Western enterprise markets, just as advanced US chip exports are restricted for China. A future ban on the use of Chinese models by US administration remains a non-zero probability in our view, and hence it’s a risk that US enterprises need to factor in when choosing AI vendors

Eventually, we envisage the co-existence of two AI ecosystems: a US-led, high-performance and enterprise-oriented stack; and a China-led, lower-cost and more open stack. Their relationship may eventually resemble iOS and Android—distinct ecosystems, different strengths and different price points.

For investors, convergence in model performance suggests that value may migrate across the AI stack. Standalone model providers could face margin compression, while owners of scarce assets—compute, chips, power and distribution—may capture a greater share of AI economics. But the picture is not one-sided. US power availability and permitting could constrain domestic expansion, while China has built substantial power capacity. But Chinese data centres face a critical limitation - access to advanced chips.

What do we like? Chips, compute, AI-enabled applications and the companies adopting AI at scale.

 

This is a marketing communication from HSBC Private Bank, which is the main private bank business within the HSBC Group. Private banking services are delivered by various HSBC companies around the world, depending on local laws and regulations. The services described in this document may be provided by different HSBC entities, and members of the HSBC Group may also trade in the products mentioned here.

 

This document is not independent investment research under the European Markets in Financial Instruments Directive (‘MiFID’) or other relevant regulations and is not subject to restrictions on dealing ahead of its distribution. This means HSBC and its staff may have an interest in the products or services mentioned before this document is shared with you.

 

The information in this document is for general information only and is intended for HSBC Private Bank clients. It does not constitute, and should not be construed as, legal, tax or investment advice, or a solicitation, offer, or recommendation to buy or sell any financial products or services.

 

Some HSBC offices may act only as representatives of HSBC Private Bank and are not permitted to sell products, provide services, or offer advice to customers. Not all products or services are available in all jurisdictions. For a complete list of HSBC Private Bank entities and their regulatory status, please visit our HSBC Private Bank website.

 

Before proceeding, please refer to the full long macro disclaimer and the Terms and Conditions available at HSBC Private Bank website which provide further important information about the use of this material.

 

© Copyright HSBC. All rights reserved.

Listening to what you have to say about services matters to us. It's easy to share your ideas, stay informed and join the conversation.