Hangzhou Pivots to Hardware: The Rise of the 'New Eight Steeds'

Hangzhou is fundamentally restructuring its AI ecosystem. The city known for Alibaba and digital commerce is moving away from pure software applications toward hard infrastructure.

This strategic shift aims to reduce dependency on foreign silicon. It focuses on building a self-sufficient supply chain for artificial intelligence.

The centerpiece of this transformation is the emergence of eight specific companies. These firms are collectively dubbed the "New Eight Steeds."

They represent a stark contrast to the previous generation of startups. Unlike the "Six Little Dragons," these new entities focus on deep tech and industrial foundations.

Key Facts: The New Eight Steeds

• Xi Wang (Xiwang) targets a 90% reduction in inference costs, aiming for pricing as low as $0.01 per million tokens.
• The eight selected sectors include inference chips, processing-in-memory, dexterous hands, and smart glasses.
• Additional critical tracks involve space-based computing, AI data infrastructure, intelligent chassis, and synthetic biology.
• This represents a decisive move from application-layer innovation to hardware-level sovereignty.
• The shift addresses vulnerabilities exposed by reliance on imported GPUs and Western toolchains.
• These companies are positioned to support the next wave of industrial AI integration.

From Digital Commerce to Industrial Foundation

For over two decades, Hangzhou defined the narrative of China’s digital economy. Its story began in the apartments of Lakeside Garden, evolving into the world’s largest e-commerce ecosystem.

The city revolutionized payment habits through Alipay and built Asia’s leading cloud platform with Alibaba Cloud. Success was measured in code, algorithms, and commercial efficiency.

However, this software-centric prosperity relied on an invisible foundation. Servers ran on chips designed elsewhere. Models trained on imported graphics processing units (GPUs).

Development tools originated across the ocean. As applications grew more complex, the lack of control over underlying hardware became a strategic risk.

The higher the application layer rose, the more precarious the unseen foundation appeared. This vulnerability became acute as AI moved from labs into real-world industries.

Hangzhou recognized that sustainable growth requires owning the stack. The city is now digging deeper into the roots of technology.

The 'New Eight Steeds' vs. The 'Six Little Dragons'

The distinction between the old and new guard is visible in their public perception. Two years ago, the "Six Little Dragons" captured global attention.

These earlier startups focused on tangible products. Consumers could easily visualize large language model chats or four-legged robots.

The "New Eight Steeds" operate differently. Their work is less visible to the average user but far more critical to industry.

Consider the specific technologies involved. Inference chips optimize how models process data after training. Processing-in-memory architectures reduce data movement latency.

Dexterous hands enable robots to perform delicate tasks, while smart glasses bridge physical and digital worlds.

These sectors are deeply technical. They require significant capital and long development cycles compared to software apps.

Yet, they form the backbone of future automation. Without these components, advanced AI remains theoretical rather than practical.

Strategic Imperatives: Reducing Inference Costs

A primary driver for this hardware pivot is cost efficiency. Current AI deployment is prohibitively expensive for many enterprises.

Xi Wang, one of the key players, has set an ambitious target. The company aims to slash inference costs by 90%.

The ultimate goal is a price point of one cent per million tokens. This would democratize access to powerful AI models.

Such reductions require specialized hardware. General-purpose GPUs are often inefficient for specific inference tasks.

Custom silicon can offer better performance per watt. This is crucial for large-scale deployments where energy costs dominate.

Hangzhou’s focus on AI data infrastructure supports this efficiency. Better data pipelines reduce waste and improve model accuracy.

By controlling both hardware and data flow, the city aims to create a competitive advantage. This advantage is not just in speed, but in economic viability.

Industry Context: Global Supply Chain Tensions

This local strategy reflects broader global trends. Geopolitical tensions have disrupted semiconductor supply chains.

Western export controls limit access to advanced chips. This forces Chinese tech hubs to innovate domestically.

Hangzhou’s approach mirrors Silicon Valley’s early days. It combines academic research with venture capital and government support.

The focus on space-based computing indicates long-term thinking. Orbital data centers could offer unique advantages in latency and security.

Synthetic biology adds another dimension. It merges biological systems with computational power, creating new material possibilities.

These diverse fields suggest a holistic view of technology. Hangzhou is not just building chips; it is building an ecosystem.

What This Means for Developers and Businesses

For global developers, this shift signals increased competition in the hardware space. Expect more optimized tools for specific AI workloads.

Businesses operating in Asia may benefit from lower infrastructure costs. Local solutions could offer better compliance and data sovereignty.

However, fragmentation is a risk. Diverging standards between East and West may complicate global deployments.

Companies should monitor these developments closely. Partnerships with local firms may become essential for market access.

Investors should watch the intelligent chassis sector. Autonomous driving relies heavily on integrated hardware-software solutions.

The rise of smart glasses also presents opportunities. Enterprise augmented reality could see accelerated adoption.

Looking Ahead: The Road to 2026

The timeline for this transformation is clear. By 2026, these eight sectors are expected to mature significantly.

Success will depend on execution. Turning research prototypes into mass-produced goods is challenging.

Talent acquisition will be critical. Engineers skilled in chip design and robotics are in high demand globally.

Regulatory frameworks must evolve. Safety standards for autonomous systems need international alignment.

If Hangzhou succeeds, it will redefine the AI landscape. The city could become a hub for industrial AI innovation.

Failure to adapt might leave it dependent on foreign tech. The stakes are high for the entire region.

Gogo's Take

• 🔥 Why This Matters: This is not just about local pride; it is a survival strategy. By mastering inference hardware, Hangzhou aims to break the bottleneck of GPU scarcity. If Xi Wang achieves its $0.01/million token target, it could disrupt global cloud pricing models, forcing US giants to compete on cost rather than just capability.
• ⚠️ Limitations & Risks: Hardware is hard. Unlike software, you cannot patch a flawed chip overnight. The capital expenditure required for fabrication and R&D is immense. There is a risk of fragmentation if these domestic solutions do not achieve global compatibility, potentially isolating Chinese AI from the broader ecosystem.
• 💡 Actionable Advice: Watch the processing-in-memory sector closely. This technology offers the most immediate ROI for reducing energy costs in data centers. For investors, look for partnerships between these "New Eight Steeds" and traditional manufacturing giants, as this integration will drive real revenue.