Jun 6, 2026 · 8:14 AM
Subscribe
Home Ai

ADTEK is testing Hong Kong's appetite for AI infrastructure listings

Shenzhen ADTEK Technology is reportedly preparing a Hong Kong IPO that could raise at least $500 million and value the optical connectivity supplier at up to $4 billion. The deal would test whether investors see fiber systems and data-center connectivity as a real AI infrastructure trade or just another IPO-cycle story.

Judith Murphy
· 5 min read · 315 views
ADTEK is testing Hong Kong's appetite for AI infrastructure listings

ADTEK's reported Hong Kong IPO plan shows how the AI trade is spreading from chips into the optical plumbing that keeps data centers connected.

Shenzhen ADTEK Technology is not a household AI name, and that is exactly why its reported listing plan matters. The company makes connectors, fiber systems and data-center connectivity products that sit behind the headline race for GPUs, models and cloud capacity. Now that quieter part of the stack is being pulled toward the public markets.

According to Bloomberg, ADTEK is considering a Hong Kong initial public offering that could raise at least $500 million, with a target valuation in the range of $3 billion to $4 billion. Local market reports said the company has been working with Jefferies and CITIC Securities on the potential deal, with a listing application possible as soon as Friday, May 15, 2026. The terms are not final, but the timing is the signal. Investors are still looking for ways to own the AI buildout, and Hong Kong has become one of the places where Chinese suppliers are trying to turn that demand into liquidity.

ADTEK was founded in Shenzhen in 2007. Its own company materials describe it as a provider of optical connectivity solutions for telecom operators, cloud service providers, equipment manufacturers and system integrators. Its product list is not flashy: optical connectors, fiber patch cables, fiber management systems, WDM components, data-center cabling products, optical distribution frames and fiber tracing solutions. But in the AI data-center economy, boring equipment can become strategic very quickly.

The first wave of AI infrastructure investing was easy to understand. More models meant more accelerators. More accelerators meant more Nvidia, more domestic Chinese GPU challengers and more high-end memory. That part of the story is still true, but it is no longer enough. Large clusters also need dense optical links, structured cabling, low-loss connectors and systems that help operators see, manage and reconfigure thousands of fiber connections without turning the data hall into a manual troubleshooting exercise.

That is where companies like ADTEK come in. A modern AI data center is not simply a warehouse full of servers. It is a network machine. Training and inference workloads move data across racks, halls and campuses, and the cost of poor connectivity shows up as latency, downtime and wasted compute. A high-density fiber product may never get the attention of a new AI chip, but it can still matter when operators are trying to scale capacity quickly and keep expensive hardware working at full utilization.

ADTEK has also positioned itself around newer optical needs. Its public product categories include CPO-related offerings, optical cabling systems, wavelength division multiplexers and fiber optic tracing solutions. It says it developed high-density cabling products in 2018, added DAC and AOC connectivity products in 2019, established a wave division production line in 2020 and expanded manufacturing in Vietnam. Those details will matter if the IPO moves ahead, because they help investors separate real AI-linked demand from a company simply dressing a telecom supply story in AI language.

Hong Kong is becoming the release valve

The proposed deal also fits a broader pattern. Reuters reported that Chinese AI startup MiniMax spearheaded six Hong Kong listings worth HK$16.7 billion, or about $2.15 billion, around the turn of the year, as AI and chipmakers helped rebuild the city's IPO market. Biren Technology, Zhipu AI and Iluvatar CoreX were part of the same public-market wave, giving investors a clearer menu of AI exposure beyond the private funding rounds that dominate the United States.

For Chinese companies, Hong Kong offers something useful. It is close to mainland capital, open to international investors and better suited than New York for companies caught in the politics of advanced technology. That matters for AI infrastructure suppliers because their growth plans often require heavy spending before public investors see the full payoff. Factories, new product lines, overseas capacity and research partnerships all require capital. A Hong Kong listing can help turn private-market AI enthusiasm into money that actually builds capacity.

The risk is that the label starts doing too much work. ADTEK sells into telecom and cloud markets, not only AI data centers. Fiber demand has been supported by 5G, broadband expansion and general cloud growth for years. Investors need to ask how much of the company's revenue is directly tied to AI clusters, how much comes from ordinary network upgrades, and whether higher-density products carry better margins or simply higher production demands. Those answers will matter more than any neat AI infrastructure narrative.

Still, the direction is hard to miss. The AI boom is widening. It is moving from model labs to chip designers, from chip designers to power equipment, and now from power equipment to optical interconnect suppliers. ADTEK's reported IPO plan is a reminder that the market is starting to price the whole data-center supply chain, including the companies that make sure the machines can talk to one another.

If ADTEK files and finds strong demand, it will give other Chinese optical and connectivity suppliers a useful benchmark. If the deal struggles, it will show that investors are becoming more selective about what counts as an AI infrastructure company. Either way, the next phase of the AI capital cycle will be less about whether demand exists and more about who can prove they sit in the critical path.

Also read: Kioxia is chasing a U.S. listing as AI makes storage scarcePOET turns a $50 million Lumilens order into a bigger AI optics testllama.cpp gives RDNA3 users a sharper local AI path

TOPICS
Judith Murphy is a financial journalist and market analyst covering AI, technology stocks, and emerging market trends. She has contributed to multiple financial publications and brings a data-driven approach to her coverage of the technology sector and its impact on global markets.
Related Articles
More posts →
Loading next article…
You're all caught up