Key Points Summary
The real bottleneck of AI data centers is not the winner of a single chip, but the entire photonics industry chain that converts electrical signals into optical signals and then transports massive data. Brian's core judgment is: as the industry upgrades from 800G to 1.6T, and even continues to move towards 3.2T, the first to gain excess returns are often not the hottest front-line companies, but suppliers like Corning, Amphenol, Ciena that all giants cannot bypass, as well as the more upstream materials and testing links. This podcast fully outlines for you the complete photonics supply chain, and before the Wall Street mainstream rush in, the high-growth gold stocks that you need to focus on tightly.
Highlights Summary
Why AI Data Centers Must Switch to Optical Communication
· "Copper cables are hitting a physical limit, and every data center must eventually transition to optics. China has proven to the entire industry that this path can continue even further."
· "Once the data transmission distance exceeds about 3 feet, copper cables will quickly lose their advantage, with higher heat generation and power consumption; while optics can solve these problems simultaneously."
· "Moving from electricity to light, this is the core significance of photonics technology."
Why Investment Opportunities Often Lie in the Supply Chain Rather Than Star Companies
· "Once a new technology is proven viable, the greatest wealth often flows first to those companies that all participants must rely on, rather than the sole company in the limelight."
· "Glass, lasers, connectors, materials, and testing equipment, none can be omitted; this is the most valuable part of the photonics industry chain."
Key Points on Corning
· "Corning's optical communications revenue grew by 36% last quarter, but the corresponding profit increased by 93%, indicating that pricing power and economies of scale are being realized simultaneously."
· "Corning has become a critical supplier designated by Meta, Amazon, Google, Microsoft, OpenAI, and Nvidia. There is no second competitor in the world that can boast such a client list. These relationships have translated into pre-committed revenue for many years."
· "Meta has committed up to $6 billion, Amazon has also signed multi-billion-dollar contracts, two other hyperscale customers have also signed similar-sized agreements, and this revenue has been locked in many years before the fiber is actually pulled."
What Amphenol and Credo Represent
· "If you are looking for a broader coverage, relatively stable, and reasonably valued optical communication interconnect target, Amphenol is a name worth keeping an eye on for the long term."
· "Credo serves as a bridge between the old world and the new world, squeezing the most out of the lifespan of copper cables within the rack and extending towards the optical communication side."
· "The risks associated with Credo are also clear, with extremely high customer concentration. A halt in purchases by a single large customer could significantly impact the stock price."
Opportunities in System Level, Upstream Material Layer, and Testing Layer
· "Ciena's value lies in enabling existing optical fibers to carry more data without the need for re-excavation and rewiring."
· "AXT operates in a more upstream position, serving as a near-monopoly supplier of critical wafer materials for optoelectronic devices. However, it also faces the high risk of China export license restrictions."
· "VEO Solutions is more like a 'scoop seller' in the optical communication world because whether it's an optical fiber link, transceiver, or system equipment, they all need to be tested before going online and monitored after operation. And VEO sells precisely the testing tools required for all these devices."
Thematic Allocation and ETF Selection
· "If you don't want to handpick individual companies, there are now pure-play photonics thematic ETFs that can cover this theme in one go."
· "However, these funds have only recently been established, have a small scale, and relatively high fees. It is suitable to put them on the watchlist first rather than blindly chase after them."
China Raises the Fiber Optic Transmission Ceiling Once Again
Brian: Recently, an engineer in China lit up a single-mode glass fiber, which has a data-carrying capacity five times higher than existing levels elsewhere in the world. This was not a newly laid line or a large-scale excavation, but the direct activation of a fiber optic cable that was already buried underground, thinner than a hair. In the past, transferring all the data from the entire U.S. congressional library might have taken half an hour, but now it takes about 5 minutes.
This technology has not yet been fully implemented in the United States, indicating just how fast-paced advancements in this field truly are. AI is generating a data deluge far beyond existing transmission capacities, and U.S. data centers are increasingly lacking in this capability. All hyperscale cloud providers will eventually need it, and they will not build the entire system from scratch; they will buy it.
What they are buying is glass, lasers, and chips that convert electricity into light, and in reality, there are only a few suppliers that can provide these things. Over the years, I have always observed the industry from the supply chain procurement side, and experience tells me that the dynamics of such opportunities have hardly changed. As soon as a new technology proves itself, the greatest wealth often flows to those companies that all participants must rely on, rather than the one name in the headlines.
Why Photonics Technology Is Now a Core Variable
Brian: Today, I want to break down the entire photonics value chain for you and explain clearly the corresponding publicly traded companies at each level. Why photonics technology, and why now?
The answer ultimately boils down to a hard constraint. Inside data centers, all chips need to communicate with each other. At short distances, copper cables still prevail; but once you want to transmit data about 3 feet away, the issues with copper cables quickly become apparent— the longer the distance, the more heat is generated and the higher the power consumption.
Light, on the other hand, almost instantaneously solves all of these issues. It can travel farther, generates less heat, and consumes only a fraction of the power compared to copper solutions. The core significance of photonics technology is the transition from electricity to light.
Moreover, the turning point is now very clear. Every data center is upgrading from 800G connections to 1.6T, and even 3.2T is already on the table. At the same time, the fiber optic cable mentioned in China has once again raised the industry's ceiling as a whole.
Breaking it down further, the key layers include: the glass, fiber optics, and cables that truly carry the light signal; the connectors that link the entire system together; the system equipment responsible for lighting up the fiber optics and transmitting data between buildings and nations; and the underlying materials and testing equipment. I have previously covered chips, lasers, and silicon photonics in separate videos, so today, I am focusing on these often overlooked but equally profitable segments.
In these groups, I will provide names worthy of being added to the watchlist. However, I must clarify that many assets have already experienced significant price appreciation, and a favorable entry point may require waiting for a pullback.
Glass and Fiber Optics Layer: Why Corning Stands Out
Brian: Starting with the fundamental material of glass, the first name is Corning. This is a 175-year-old material company that is responsible for manufacturing optical fiber, the type of "glass thread" mentioned at the beginning of the video. It holds approximately a 20% share of the global optical fiber market and is a key player in the industry.
Corning's true differentiator lies in its technology. Its latest generation of optical fiber can accommodate approximately twice the standard cable's fiber count in the same physical space, a capability that is highly sought after by the already crowded AI data centers. Its bend-resistant glass is also difficult for competitors to replicate. Coupled with the world's largest fiber optics factory and onshore U.S. supply chains compliant with the "Buy America" policy, these factors together constitute its real moat.
For this reason, Corning has become a core supplier relied upon by Meta, Amazon, Google, Microsoft, OpenAI, and Nvidia. There is no second competitor globally that can boast such a customer list simultaneously. These relationships go beyond mere cooperation and have already translated into locked-in multi-year revenue.
The industry's demand for optical fiber is growing at approximately 22% to 25% annually, yet the entire industry's new supply capacity is only about half of this growth rate, leading to delivery lead times of over 60 weeks. As a result, large-scale customers reserve capacity several years in advance, prepaying to secure production capacity. Meta has committed to up to $60 billion, Amazon has also signed contracts worth billions of dollars, and two other large-scale customers have signed agreements of similar amounts, with these revenues locked in many years ahead of the actual deployment of the fiber.
What caught my attention the most is the profit elasticity. Corning's last quarter's optical communications revenue grew by 36%, but the segment's profit surged by 93%, more than double the revenue growth rate. This exemplifies the simultaneous impact of pricing power and economies of scale. At the company-wide level, its operating profit margin has increased from approximately 8% two years ago to over 16% now, with management aiming for 20% by the end of this year.
Of course, the reality side also needs to be acknowledged that this story is no longer a secret. Corning's current PEG is close to 3, with a P/E ratio of about 9 times, which is not cheap for a materials company. So if you want to allocate fiber optics in the most stable and least dramatic way, Corning is suitable, but the more reasonable pace would be to wait for the next more decent pullback.
Core Companies in the Interconnect Layer: Amphenol and Credo
Brian: Next, moving into the interconnect layer, which is the layer that truly connects all components, the first name is Amphenol. It is a rather low-key giant that deals with high-speed connectors and cables, including both copper and fiber optics. Now, almost every new AI server rack prominently features it.
Understanding Amphenol is key; it is fundamentally an extremely efficient M&A machine. In January this year, it spent $10.5 billion to acquire CommScope's entire fiber optic connectivity business, overnight transforming from a connector company into a substantial player in the fiber optics space. Now, the AI data center business has become the core engine of the entire company, its largest single segment, with organic growth exceeding 80% last quarter.
Its order backlog has also reached a record $9.4 billion, with new orders still outpacing shipments. As quarterly revenue jumped from around $4 billion to just over $7 billion, its operating margin not only was not dragged down but expanded to nearly 28% from 22%.
This is very worth noting because under normal circumstances, a $10 billion-plus acquisition would at least make a company endure integration pressures for one to two years, often resulting in an initial decline in profit margins. But Amphenol's profit margin has actually increased, indicating its strong ability to quickly align the acquired target with its high standard operating system. It's also because of this that such a large transaction has not become a burden but rather a profit enhancer.
What's even rarer is that its valuation is not unreasonable. Amphenol's PEG is approximately only 0.7, with a P/E ratio of around 7 times. For a company growing so rapidly, this valuation is not common. So if you want an interconnect target in the optical communication sector that has a broader coverage, relatively stable fluctuations, and a reasonable valuation, Amphenol is a name worth long-term observation.
Similarly, the second name still in the interconnect layer is Credo Technology. Its role is more like a bridge between the old world and the new world. On one hand, it maximizes the transmission capacity of copper cables inside the rack using low-power chip technology; on the other hand, it is also involved in optical communication chips and cables, seamlessly taking over when signals need to travel further.
It recently acquired a silicon photonics company, completing its end-to-end product stack to 1.6T, and has already shipped to all of the top five hyperscale cloud providers in the United States. Its growth has been quite impressive, with quarterly revenue skyrocketing from $135 million to $437 million in just six quarters, more than tripling.
Another key indicator is its approximately 68% gross margin. This number is more typical of a software company rather than a hardware company. Meanwhile, its operating margin has also nearly doubled to 37% as it scales, and the management's revenue guidance for the next fiscal year still indicates over 80% growth.
However, there is a very specific risk here that must be taken seriously. While it supplies to all five major customers, just three of them account for 88% of the company's revenue. If one of these hyperscale customers slows down their procurement, this stock is very likely to be swiftly punished by the market. Additionally, insider selling has continued during the rise, and the company's current P/S ratio is around 35 times, essentially pricing in a scenario where "almost everything continues to go smoothly." With a PEG ratio close to 1, indicating strong growth, companies like this are more of a target for those with high conviction and are willing to wait for a significant pullback.
The Key Player at the System Level: Ciena
Brian: Moving up the chain, we reach the system layer that lights up fiber optics and transports data between buildings and across nations. The most critical name in this link is Ciena. It is a leader in Western coherent optics, and its proprietary WaveLogic technology is the world's first solution to squeeze 1.6T of data into a single-wavelength light signal. You can think of this as a "capacity expansion without digging" cheat code because it allows existing fiber optics to carry more data without the need for rewiring.
And this is not just a technology showcase in a lab. In just two quarters, this single product has already won over 49 customers. At the same time, its position in major customer relationships is also very strong. Its solutions have been adopted by three out of the four largest hyperscale cloud and cloud service providers, with cloud customers now contributing to nearly half of the company's revenue.
For me, the most crucial data is still the order backlog. Last quarter, Ciena's backlog increased by approximately $2 billion in a span of 90 days, reaching close to $7 billion. Almost all of this is planned for delivery next year, meaning it has already locked in a revenue scale equivalent to over a year in advance.
With revenue hitting a historic high and growing 40% year-over-year, its operating profit margin has also doubled from under 8% to over 15%. The issue, however, is that the market is also aware of this good news, so the valuation has become very aggressive. Ciena's current expected price-to-earnings ratio is around 120 times, which is nearly pricing in "perfect execution." Therefore, while the fundamentals are strong, caution must be exercised regarding the price.
The True Upstream Bottleneck: AXT and Testing Equipment Company VEO Solutions
Brian: Next, we move into my favorite layer, which is the "supplier behind the supplier." The top-tier name in the entire video is AXT. There are few companies globally that can truly produce indium phosphide wafers, and this particular crystal is the key material upon which all optical lasers must be built. From this perspective, it is almost a natural moat because the crystal's growth capability is not something you can achieve overnight.
Now, as all laser manufacturers are scrambling for supply, AXT's orders for this type of material have backlog exceeded $100 million, setting a record. But its risk is also at the forefront of the entire list, and it is very specific. Almost all of the company's manufacturing is in China, and China now requires government approval for every export order. This has directly impacted its revenue recognition in the last quarter, and even with a record backlog sitting there, it does not guarantee smooth delivery.
In addition, the company has raised approximately $550 million in financing, leading to shareholder dilution; insiders have also sold over $70 million worth of stock during the uptrend. More trouble, it is still not profitable at the moment, although the losses are narrowing, and the gross margin has increased to about 30%.
So, AXT's story is real, the bottleneck is real, but if you look at its price-to-sales ratio of nearly 66 times, you will understand that this is more like a high-volatility, high-risk small position "lottery ticket" rather than a core asset suitable for a heavy position.
The last name in this layer is VEO Solutions. It is more like the "seller of shovels" in the optical communication world because whether it's fiber optic links, transceivers, or system equipment, they all need to be tested before deployment and monitored after operation, and VEO sells the testing tools that all these devices must go through.
The beauty of this position is that it is not sensitive to specific winners. You don't need to bet on which laser company will eventually prevail, or speculate on which transceiver will become mainstream because no matter who wins, they all eventually need to test their equipment, which is a business position I personally really like.
Over the years, VEO's revenue has remained relatively flat, stuck around approximately $2.85 billion per quarter. It wasn't until the acceleration of AI infrastructure development that its network testing business started to take off, experiencing a surge of over 54% as data centers rushed to validate a large number of new optical communication devices. Now, its quarterly revenue has surpassed $4 billion, and the operating profit margin has returned to double digits.
In addition, the company has a quieter but highly profitable second business of producing anti-counterfeiting coatings printed on global banknotes, providing another high-margin support for the entire company. It is worth noting that insiders, including the CEO, have been consistently selling off their shares. Its current PEG ratio is around 1.4, not too exaggerated, but like other companies on the list, considering its significant previous gains, the more sensible approach is still to let it cool off first.
If you don't want to pick individual stocks one by one, there is a newly launched pure-play photon ETF
Brian: Many people may be wondering why there is no focus today on Coherent, Lumentum, Marvell, Broadcom, and those semiconductor foundries that actually manufacture chips?
Of course, these companies are all worth watching as they are at the core of this industry chain, in the laser and silicon photonics segment, and are a very important part of the entire theme. It's just that I have already done a separate in-depth analysis of them before. So today, I want to focus on the parts that others often overlook, such as glass, connectors, system equipment, and their suppliers.
If you don't want to individually select companies, there is now a one-click method to configure the entire theme. A pure-play photonics theme ETF has just appeared in the market, with the code FOTO, known as the Tuttle Capital Pure Play Photonics Fund.
One thing I like about it is that it does indeed carry out "pure-play" filtering. It excludes large conglomerates where the photonics business is only a small part and focuses its positions on true optical companies. Its largest holding is exactly the batch of laser and transceiver companies I mentioned earlier, such as Lumentum, Coherent, Fabrinet, as well as IPG Photonics and Inphi. The entire fund only has 15 names, with the top 10 holdings accounting for nearly 90%, so the style is very concentrated and clearly bets on "pure photonics."
Of course, full disclosure is necessary. This fund was just launched a few weeks ago and does not yet have a real long-term performance record. The AUM is approximately $140 million, and the expense ratio is 0.75%. So, you should definitely do your own research first. But if you just want to initially include the "copper-to-light" theme in your watchlist, FOTO is indeed worth paying attention to.
Summary: Copper has reached its limit, and the real beneficiaries will diffuse along the entire photonics chain
Brian: Let's wrap it up. Copper has started hitting its physical limit, and every data center on Earth must transition to light. China has already proven to the whole industry that this path can go even further.
Funds are now pouring into this chain, and the real question is no longer whether this will happen, but rather, in this accelerated transition, which group of companies will capture the most incremental value.
If you take the laser and silicon photonics companies I talked about earlier, and add in today's discussion of glass, connectors, systems, and upstream suppliers, you actually have a fairly complete map of the photonics industry chain from end to end.
Welcome to join the official BlockBeats community:
Telegram Subscription Group: https://t.me/theblockbeats
Telegram Discussion Group: https://t.me/BlockBeats_App
Official Twitter Account: https://twitter.com/BlockBeatsAsia
