AI tận cùng là ánh sáng: Bản đồ chuỗi ngành cổ phiếu tăng gấp 10 lần mà đa số bỏ qua

Compiled & Translated by: Odaily TechFlow

Host: Brian, formerly at Target & Amazon

Podcast Source: BWB - Business With Brian

Original Title: Millionaires are Hitting These 10X Stocks HARD!

Air Date: June 21, 2026

Key Takeaways

The real bottleneck for AI data centers isn't a single chip winner, but the entire photonics supply chain that converts electrical signals to optical signals and moves vast amounts of data out. Brian's core thesis is: when the industry upgrades from 800G to 1.6T, and eventually to 3.2T, the ones capturing the most excess returns are often not the hottest front-line stars, but suppliers like Corning, Amphenol, and Ciena that all giants must use, as well as the upstream materials and testing stages. This podcast provides a comprehensive breakdown of the complete photonics supply chain and the golden high-growth stocks you need to watch closely before the Wall Street herd rushes in.

Highlights & Key Insights

Why AI Data Centers Must Transition to Optical Communication

• "Copper is hitting a physical limit. Every data center will eventually have to switch to optics. China has already shown the industry that this path can go even further."
• "Once data needs to travel beyond about 3 feet, copper rapidly loses its advantage, generating more heat and consuming more power; optics can solve both problems simultaneously."
• "Transitioning from electricity to optics – that's the core meaning of photonics technology."

Why Investment Opportunities Often Lie in the Supply Chain, Not Just Star Companies

• "Once a new technology is proven viable, the biggest wealth often flows first to the companies that all participants must rely on, not just the single name grabbing headlines."
• "Glass, lasers, connectors, materials, and testing equipment – none are optional. That's where the most valuable part of the photonics chain lies."

Key Points on Corning

• "Corning's optical communications revenue grew 36% last quarter, but corresponding profits grew 93%. This shows pricing power and economies of scale are materializing simultaneously."
• "Corning has become a core supplier relied upon by Meta, Amazon, Google, Microsoft, OpenAI, and Nvidia. No other competitor in the world has this exact client list. These relationships have translated into multi-year locked-in revenue."
• "Meta committed up to $6 billion, Amazon signed multi-billion dollar contracts, and two other hyperscalers signed similar-sized agreements. And this revenue is locked in years before the fiber is even drawn."

What Amphenol and Credo Represent Respectively

• "If you want a broader, less volatile optical interconnect target with reasonable valuation, Amphenol is a name worth watching long-term."
• "Credo acts as a bridge between the old and new world, maximizing the lifespan of intra-rack copper cabling while also extending into the optical domain."
• "Credo's risk is clear: extremely high customer concentration. If one large customer pauses procurement, the stock price could take a significant hit."

Opportunities in System-Level, Upstream Materials, and Testing

• "Ciena's value lies in enabling existing fiber to carry more data without needing to dig and lay new cables."
• "AXT sits further upstream, being a scarce supplier of critical wafer materials for optical lasers. However, it faces significant China export license risks."
• "VEO Solutions is more like the 'picks and shovels' provider in the optical world. Fiber links, transceivers, and system equipment all need testing before deployment and monitoring after launch. VEO sells the tools these devices must pass through."

Thematic Allocation and ETF Choices

• "If you don't want to pick individual companies, there are now pure-play photonics ETFs that cover this theme in one click."
• "However, these funds are newly established, with small AUM and higher expense ratios. They are suitable for a watchlist first, not for blindly chasing highs."

China Has Raised the Fiber Optic Transmission Ceiling Significantly

Brian:

Engineers recently lit up a single strand of glass fiber in China, with data-carrying capacity five times higher than current levels elsewhere in the world. This isn't a newly laid line or massive excavation. It's a fiber thinner than a hair activated directly within existing buried cables. Previously, moving all the data from the US Library of Congress might take 30 minutes; now it takes about 5.

This technology hasn't really been deployed in the US yet, illustrating just how fast this field is evolving. AI is generating a flood of data that surpasses current transmission capabilities, and US data centers are increasingly lacking this capacity. Eventually, all hyperscale cloud providers will need it. And they won't build the entire system from scratch; they will buy it.

What they buy is glass, lasers, and the chips that convert electricity to light. The suppliers that can truly provide these are actually just a few. Having spent years on the procurement side of the supply chain, my experience tells me the pattern for these opportunities hasn't changed much. Once a new technology is proven viable, the biggest wealth often flows first to the companies that all participants must rely on, not just the single name grabbing headlines.

Why Photonics Is Becoming the Core Variable Now

Brian: Today, I want to break down the entire photonics supply chain for you and explain the public companies at each layer. Why photonics? Why now?

The answer boils down to a hard constraint. Inside a data center, all chips need to communicate. Over short distances, copper remains the winner. But once you want to transmit data beyond about 3 feet, copper's problems quickly surface. The longer the distance, the worse the heat and the higher the power consumption.

Optics solves most of these problems in one go. It transmits further, generates less heat, and consumes only a fraction of the power of copper solutions. Transitioning from electricity to optics – that's the core meaning of photonics technology.

More importantly, this inflection point is now very clear. Every data center is upgrading from 800G connectivity to 1.6T, and even 3.2T is on the table. Simultaneously, the Chinese fiber mentioned earlier has pushed the industry's ceiling significantly higher overall.

Breaking it down, the most critical layers include: the glass, fiber, and cables that actually carry the optical signal; the connectors that tie everything together; the system equipment that lights up the fiber and transmits data between buildings and countries; and the underlying materials and testing equipment. I've covered chips, lasers, and silicon photonics separately in previous videos, so today I'll focus on these often-overlooked but equally profitable segments.

Within these groups, I'll provide names worth adding to a watchlist. But let me be clear upfront: many of these have already risen significantly. Truly favorable entry points will likely require pullbacks.

At the Glass and Fiber Layer, Why Corning Is the One to Watch Most

Brian:

Let's start with the most basic layer: glass. The first name is Corning. This 175-year-old materials company is the one actually drawing the fiber – the "glass thread" mentioned at the start of the video. It holds roughly a 20% share of the global fiber market, making it a very core player.

Where Corning truly differentiates itself is technology. Its latest generation fiber can pack roughly twice as many cores into the same physical space as standard cables – exactly what overcrowded AI data centers desperately need. Its bend-insensitive glass is also difficult for competitors to replicate. Add to that the world's largest fiber manufacturing plants and a US-based supply chain compliant with "Buy America" rules, and these factors combined form a formidable moat.

This is why Corning has become a core supplier relied upon by Meta, Amazon, Google, Microsoft, OpenAI, and Nvidia. No other competitor in the world has this exact client list. And these relationships aren't just "great collaborations"; they've translated into multi-year locked-in revenue.

Industry demand for fiber is growing at about 22% to 25% annually, but the total new supply capacity is only about half that pace. Lead times have stretched to over 60 weeks. Consequently, hyperscale customers are booking capacity years in advance, sometimes paying upfront to secure it. Meta committed up to $6 billion, Amazon signed multi-billion dollar contracts, and two other hyperscalers signed similar-sized agreements. And this revenue is locked in years before the fiber is even drawn.

What impresses me most is the profit leverage. Corning's optical communications revenue grew 36% last quarter, but segment profit grew 93% – more than double the revenue growth rate. That's what pricing power and economies of scale look like. Company-wide, its operating margin has already improved from around 8% two years ago to over 16% now, with management targeting 20% by the end of this year.

Of course, we must acknowledge the reality: this story isn't a secret anymore. Corning's current PEG ratio is near 3, and its price-to-sales ratio is about 9x. For a materials company, this isn't cheap. So, if you want the most stable, least dramatic way to play the fiber layer, Corning is suitable, but a more sensible approach is to wait for the next decent pullback.

Key Companies in the Interconnect Layer: Amphenol and Credo

Brian:

Next, we move to the interconnect layer – the one that physically connects all the components. The first name is Amphenol. It's a relatively low-profile giant that manufactures high-speed connectors and cables, both copper and optical. You can find its products in almost every new AI server rack.

The key to understanding Amphenol is that it's essentially a highly efficient M&A machine. In January this year, it spent $10.5 billion to acquire CommScope's entire fiber connectivity business, transforming overnight from a connector company into a significant optical player. Now, the AI data center business is the core engine for the entire company, its largest single segment, with organic growth exceeding 80% last quarter.

Its order backlog is also at a record $9.4 billion, and new orders are still outpacing shipments. As quarterly revenue jumped from around $4 billion to just over $7 billion, its operating margin wasn't dragged down; instead, it expanded from 22% to nearly 28%.

This is noteworthy. Normally, a $10 billion+ acquisition would cause at least a year or two of integration pressure, with margins often declining first. But Amphenol's margins actually improved. This shows its strong ability to quickly integrate acquired companies into its own high-standards operating system. Consequently, such a large deal didn't become a burden but acted as a profit enhancer.

What's more remarkable is that its valuation isn't unreasonable. Amphenol's PEG ratio is only about 0.7, and its price-to-sales ratio is around 7x. This is uncommon for a company growing this fast. So, if you want a broader, less volatile optical interconnect target with reasonable valuation, Amphenol is a name worth watching long-term.

Also in the interconnect layer is Credo Technology. Its role is more like a bridge between the old world and the new. On one hand, it uses low-power chip technology to squeeze the maximum transmission capability out of intra-rack copper cabling. On the other hand, it's developing 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 up to 1.6T, and is already shipping to all top five US hyperscale cloud providers. Its growth has been explosive: quarterly revenue surged from $135 million to $437 million in just six quarters, more than tripling.

Another telling indicator is its gross margin of roughly 68%. This number looks more like a software company than a hardware one. Meanwhile, its operating margin has nearly doubled to 37% as scale increased. Management's revenue guidance for the next fiscal year still points to over 80% growth.

However, the risk here is very specific and must be heeded. Although it supplies all five major clients, just three of them account for 88% of the company's revenue. If one hyperscale customer slows its procurement, this stock could get hammered by the market quickly. Add to that continuous insider selling during the rally, and the company's current price-to-sales ratio of roughly 35x – the market is essentially pricing in "everything going right." A PEG near 1 indicates strong growth, but this kind of stock is more of a high-conviction idea best considered only after a deep pullback.

The Truly Key Player in the System Layer: Ciena

Brian:

Moving up to the system layer – which lights up the fiber and moves data between buildings and countries – the most critical name here is Ciena. It's the leader in Western coherent optics. Its proprietary WaveLogic technology is the world's first solution to pack 1.6T of data into a single wavelength optical signal. You can think of it as a "capacity upgrade without digging" cheat code, because it enables existing fiber to carry more data without needing to lay new cables.

This isn't just a lab demonstration. In just two quarters, this single product has already won 49 customers. Simultaneously, its position with key accounts is very strong. Its relevant solutions have been adopted by three of the top four hyperscale cloud and cloud service providers, and cloud clients now contribute nearly half of the company's revenue.

For me, the most critical data point is still the order backlog. Last quarter, Ciena's backlog increased by about $2 billion in 90 days, reaching nearly $7 billion. Almost all of it is scheduled for delivery next year, meaning it has already locked in over a year's worth of revenue.

With revenue reaching record highs and growing 40% year-over-year, its operating margin has doubled from under 8% to over 15%. The problem is the market knows this good news too, so the valuation is already very aggressive. Ciena's current forward P/E ratio is around 120x – this is essentially pricing in "perfect execution." So, the fundamentals are great, but one must be wary of the price.

The Real Upstream Bottlenecks: AXT and Testing