Recently, I've been looking into the developments of CPO concept stocks and noticed a pretty interesting phenomenon—traditional server hardware valuations have already been driven very high by the market, and funds are starting to seek the next high-growth sectors. Since CPO and silicon photonics are still in their early stages technologically, their growth potential over the next 5 to 10 years is simply unimaginable. For those aiming for excess returns, this is definitely a must-watch investment theme.

First, let’s talk about why these two concepts are always linked together. Traditional chips transmit signals via copper wires, but as AI computing demands explode, this approach has become bottlenecked—too hot, too slow, and too energy-consuming. The emergence of CPO and silicon photonics aims to solve this problem by replacing electrical transmission with “light,” greatly increasing speed and significantly reducing power consumption.

Silicon photonics, simply put, involves miniaturizing the large optical components (lasers, detectors, modulators, etc.) to the size of microprocessors and integrating them onto silicon substrates. Meanwhile, CPO moves the optical transceiver modules directly next to the CPU or GPU, packaging them on the same board, which can drastically reduce power consumption—saving over 30% of energy per unit. In plain terms, silicon photonics is the foundational technology, and CPO is the most promising application direction.

Taiwan has a natural advantage in this field. The semiconductor supply chain—from wafer foundry to packaging, testing, and optical communication components—is a comprehensive strength that few global players can match. By 2026, silicon photonics technology will have officially entered large-scale mass production, which is why discussing CPO concept stocks has become especially important now.

TSMC (2330) is of course the leader among leaders. It’s not just a chip foundry but also defining the packaging standards for CPO. Its COUPE platform is at the core of silicon photonics development, and the CPO packaging technology expected to mass produce in 2026 will be led and set by TSMC. Following closely are companies like ASE (6451) and Advanced Semiconductor Engineering (3711), as leading advanced packaging and testing firms, they will also be prioritized for certification and orders. Alpha & Omega Semiconductor (3363) has deep cooperation with TSMC on optical fiber array connection technology—this “interface” is very critical, and the market recognizes it as benefiting the most.

Browave (3163) controls passive optical components, United Microelectronics Corporation (3081) provides laser light sources, and Pan-Quen (6830)’s optical path positioning and inspection technology can improve yield—these are all noteworthy targets within the CPO concept stocks.

The logic on the U.S. side is somewhat different. Broadcom (AVGO) leads in the CPO field; its Tomahawk series launched in 2026 has become the standard for AI data centers. Marvell (MRVL) competes with Broadcom in high-speed optical interconnect chips, and NVIDIA announced a deep partnership with Marvell at the end of March 2026. Qorvo (CRDO) just announced on April 13th that it would acquire DustPhotonics for up to $1.3 billion, directly gaining control of photonic integrated circuit technology. This kind of transformation flexibility is far beyond that of large established companies.

However, investing in these CPO concept stocks requires paying close attention to several risks. First is yield issues—since optical components are integrated with chip packaging, if one component fails, the entire GPU or switch chip might need to be scrapped. When reviewing financial reports, watch for gross margin trends; if revenue increases but gross profit declines, it could indicate ongoing yield struggles.

Second is the competition of specifications. The market doesn’t only have CPO; there’s also LPO (linear drive pluggable optical modules), an enhanced version of traditional modules that are more power-efficient, cheaper, and easier to maintain. Before the widespread adoption of 1.6T transmission rates, LPO might take a significant market share from CPO.

Additionally, pay attention to the actual revenue of companies. If a company claims to be a silicon photonics concept stock but has very low optical communication-related revenue, be cautious whether they are just riding the trend. Genuine profit support should come from actual orders from giants like NVIDIA and Broadcom.

Finally, don’t forget geopolitical factors. The U.S. broadband infrastructure plans will directly impact optical communication demand, and policy delays could affect downstream companies. Silicon photonics is cutting-edge technology and is vulnerable to U.S.-China tech conflicts.

Ultimately, CPO concept stocks are not short-term themes but a structural growth trend over the next 5 to 10 years. The key principle is simple—U.S. stocks focus on “standard-setting,” while Taiwan stocks focus on “supply chain performance.” Prioritize companies that have passed major manufacturer certifications and show a clear increase in optical communication revenue, so you can capture truly valuable enterprises in this high-speed race.