If you want to understand where AI spending is actually going, stop looking at the companies building models and start following the circuit boards.
Victory Giant Technology, a Chinese printed circuit board manufacturer that supplies Nvidia’s AI server ecosystem, surged as much as 60% on its Hong Kong trading debut Tuesday after raising HK$20.1 billion ($2.57 billion) in the city’s largest initial public offering this year. Shares priced at HK$209.88 and promptly rocketed, delivering the kind of first-day performance that signals something deeper than IPO hype: genuine, structural demand for the companies that build the physical infrastructure of artificial intelligence.
From Seventh Place To Market Leader In One Year
Victory Giant’s rise is a case study in how quickly the AI supply chain is reshuffling the competitive order. In the first half of 2024, the company ranked seventh among global PCB makers by revenue in the AI and high-performance computing segment, holding just 1.7% of the market. By the first half of 2025, it was number one, commanding 13.8% market share. That is not incremental growth. That is a company riding a wave so large it lifted it past every established competitor in under twelve months.
The catalyst is straightforward: Nvidia’s AI server demand is so voracious that it is pulling its entire supply chain upward. Victory Giant makes the advanced printed circuit boards that go inside AI servers, the physical substrate on which GPUs, memory, and networking components are mounted. As Nvidia has scaled production of its H100 and successor chips, the companies supplying the surrounding components have seen order volumes that were unimaginable just two years ago.
This is the part of the AI story that rarely makes headlines but determines who actually captures value. Foundation model companies like OpenAI and Anthropic attract the attention. Cloud hyperscalers like Microsoft, Google, and Amazon attract the capital expenditure coverage. But the money flows downhill, and at the bottom of the stack, companies like Victory Giant are converting raw demand into real revenue at margins that the rest of the supply chain envies.
What The IPO Says About AI Infrastructure Demand
The size and performance of Victory Giant’s listing carry implications beyond one company’s stock chart. Hong Kong’s IPO market raised HK$109.9 billion across 40 new listings in the first quarter of 2026, nearly six times the amount raised in the same period last year and triple the number of new listings. Victory Giant’s offering alone accounted for a significant chunk of that total, and investor demand remained strong despite volatility tied to the Iran conflict and broader geopolitical uncertainty.
That tells you two things. First, institutional appetite for AI infrastructure plays is deep enough to absorb multi-billion-dollar offerings without flinching, even in choppy markets. Second, the Hong Kong exchange is emerging as the venue of choice for Chinese AI supply chain companies seeking public capital, a development with strategic implications as U.S. restrictions on Chinese technology investment continue to tighten.
Victory Giant plans to allocate roughly 75% of its IPO proceeds to expanding manufacturing capacity in mainland China. That expansion is a bet that current demand levels are not a peak but a floor, that Nvidia and its competitors will continue scaling AI server production for years, not quarters. Given that every major hyperscaler has publicly committed to increasing AI capital expenditure through at least 2027, the bet looks well founded.
The Supply Chain Geography Problem
Victory Giant’s dominance raises a question that U.S. policymakers and investors should be asking more loudly: how much of the physical AI supply chain is concentrated in China?
The United States has spent the past three years building semiconductor fabrication capacity through the CHIPS Act, luring TSMC and Samsung to build fabs on American soil, and restricting China’s access to advanced chipmaking equipment. But chips are only one component of an AI server. The circuit boards, connectors, thermal management systems, and power delivery infrastructure that surround those chips are manufactured overwhelmingly in Asia, with China and Taiwan accounting for the vast majority of production.
Victory Giant’s IPO prospectus makes the dependency concrete. The company’s manufacturing footprint is entirely in mainland China. Its customers include the largest names in the AI hardware ecosystem. If geopolitical tensions escalate further, whether through trade restrictions, export controls, or a Taiwan contingency, the PCB layer of the supply chain becomes a chokepoint that no amount of domestic chip fabrication can bypass.
This is the unsexy but critical dimension of the AI race: you can design the most advanced chip in the world, but if you cannot mount it on a circuit board manufactured to the required specifications, it does not ship. Victory Giant’s market position is not just a business advantage. It is a strategic one.
Follow The Money Down The Stack
The investor playbook for AI has followed a predictable pattern. First came the foundation model companies. Then the chip designers, led by Nvidia. Then the cloud providers building AI data centers. Each wave pushed further down the technology stack, toward the less glamorous but increasingly essential companies that manufacture physical components.
Victory Giant’s 60% pop on day one suggests the market is now fully pricing in the supply chain layer, at least for the most visible players. The question for investors is whether this extends to other PCB makers, connector manufacturers, and thermal management companies that serve the same ecosystem but have not yet gone public or attracted the same attention.
The data center buildout driving this demand shows no signs of slowing. Microsoft, Google, Amazon, Meta, and Oracle have collectively committed hundreds of billions of dollars to AI infrastructure through 2027 and beyond. Every dollar of that spending requires servers, and every server requires circuit boards. Victory Giant is positioned at the intersection of those capital flows, and Tuesday’s IPO performance confirms that investors are paying attention.
For anyone still wondering whether AI infrastructure spending is sustainable, the answer arrived on a Hong Kong stock exchange terminal Tuesday morning, priced at HK$209.88 and trading 60% higher by lunch. The supply chain is the story now, and the companies building the physical backbone of AI are just getting started.