The entire optical communication industry's production capacity is sold out—where is the current point of contention?

The ongoing wave of AI data center construction continues to drive demand for optical communications, resulting in significant recent increases in related stocks, with valuations surpassing historical ranges across the board.

According to Windy Trading Desk, Morgan Stanley released a research report on April 20 stating that a new wave of investors has continued to flock to the optical communications sector after OFC (Optical Fiber Communication Conference), with buying power still expanding, providing sustained support for stock prices.

Many optical communications companies have sold out their production capacity for this year, significantly reducing revenue uncertainty. Profit margin has become the most critical metric for the coming quarters.

The progress of laser production capacity expansion remains opaque, making it more difficult to accurately gauge supply-demand gaps. Attention should also be given to the "Narrow and Fast" versus "Wide and Slow" architecture debate. The "Wide and Slow" solution introduces the possibility of using MicroLED to replace InP lasers, which could reshape the long-term supply chain landscape.

Morgan Stanley has raised target prices for four major optical communications stocks: Corning from $127 to $140; Lumentum from $595 to $710; Coherent from $250 to $290; and Ciena from $286 to $405.

Sold-out production capacity reshapes the competitive logic, profit margin becomes the core variable

Against the backdrop of widespread production tightness, the investment logic in the optical communications sector has quietly shifted. Since many companies have locked in capacity through most of this year or 2026, revenue projections have become relatively clear. As a result, the market is focusing on whether companies can effectively capture profit margin upside in a supply-constrained scenario.

On this topic, LITE currently holds the most advantage. The company has aggressively raised prices, which is a key reason for buy-side EPS expectations being much higher than consensus sell-side forecasts. However, this pricing power is not universally replicable across the entire supply chain.

GLW’s situation is more complex. Asian optical fiber prices have risen by 75%, reaching a seven-year high, and investors hope GLW will benefit as well. However, GLW management remains cautious, stating that "there is enough optical fiber globally to meet demand," and that price improvement will be more driven by product innovation than direct spot price hikes.

Two marginal changes warrant attention: First, over 30% of BEAD funds (US broadband infrastructure subsidy) are shifting toward fixed wireless and satellite, likely reducing optical fiber demand;

Second, after CommScope's optical fiber business was sold to Amphenol, Corning canceled its original raw material fiber orders from CommScope, leading to a chain of canceled orders and a slight tightening in the fiber market.

Laser capacity expansion progress remains opaque; supply pacing becomes a key uncertainty

Indium Phosphide (InP) lasers are currently the core devices for AI optical interconnects; their capacity dynamics directly impact overall supply and demand across the supply chain and are among the most watched topics for investors.

From disclosed information:

LITE plans to expand EML capacity by more than 50% in fiscal year 2026;

COHR plans to double capacity in fiscal year 2026, and increase by over 100% again in fiscal year 2027;

Sumitomo Electric aims to double between 2024 and 2026, and increase another 40% between 2026 and 2028;

Broadcom stated it will expand capacity by four to six times within the next year.

However, several key variables remain unclear.

Firstly, the allocation ratio between EML and CW lasers in 1.6T demand is disputed—LITE expects CW to account for about 30% to 40%, while Sumitomo believes it will exceed 50%.

Secondly, many manufacturers are switching to 6-inch wafers, but yield rates remain highly uncertain. In addition, Broadcom’s expansion details are still unclear.

Due to the lack of reliable market share data and only LITE having disclosed more detailed expansion information, the accuracy of overall supply-demand modeling is limited. Whether the pace of supply expansion can match the growth momentum of AI demand will be an important variable affecting sector sentiment.

The "Narrow and Fast" vs "Wide and Slow" architecture debate may reshape the long-term supply chain landscape

In CPO (Co-Packaged Optics), the debate between "Narrow and Fast" and "Wide and Slow" architectures is becoming an increasingly focused long-term topic for investors.

The "Narrow and Fast" solution achieves higher single-channel rates with fewer channels. A typical configuration is 1.6T = 8 channels × 200G, optimizing bandwidth density and power efficiency, and is the main trend in current AI cluster construction.

The "Wide and Slow" solution uses more channels and lower single-channel rates to achieve the same total bandwidth. A typical setup is 1.6T = 16 channels × 100G. It relies on a more mature ecosystem but consumes more space and fiber cores.

The strategic significance of the "Wide and Slow" solution lies in its introduction of MicroLED as a possible replacement for InP lasers.

Microsoft’s MOSAIC architecture exemplifies this direction—by using imaging fiber (containing thousands of fiber cores) to support numerous low-speed parallel optical channels, it may reduce reliance on InP lasers for certain scenarios. COHR advocates that VCSEL technology is also competitive in the "Wide and Slow" architecture.

This debate is expected to take years to resolve, but it is already impacting the long-term positioning of COHR and LITE—if the "Wide and Slow" approach accelerates penetration, vendors that currently rely on InP lasers as their core competitive edge may face larger long-term challenges, while companies involved with MicroLED stand to benefit.

Valuations break through historical ranges, no catalysts yet to disprove the "bull market hypothesis"

Valuations for optical communications stocks have now fully departed from historical averages, becoming another key source of controversy.

Major targets including CIEN, LITE, COHR, and GLW are currently trading at 20 to 25 times projected 2028 EPS, while their 10-year average P/E ratios hover around 17 to 19 times. CIEN and LITE are valued at about three times their respective five-year historical averages.

By contrast, storage and hard drive companies (such as MU) still sport single-digit to low double-digit P/E multiples, this sharp divergence prompting frequent questions about the validity of valuations for optical communications stocks.

The rationale for the premium boils down to the following:

First, the application scenarios of optical communications continue to expand (Co-Packaged Optics CPO, Optical Circuit Switching OCS, coherent transmission, etc.), making growth prospects more explicit;

Second, there is a sizable gap between buy-side expected EPS and consensus sell-side forecasts, with most catalysts (such as new product ramp-ups and pricing effects) not expected to materialize until 2027–2028;

Third, unlike storage, optical communications currently lack recent data points that could disprove the bull market hypothesis, while AI capital expenditure remains robust, lending further support to valuations.

However, current valuations are historically high, and if profit margins fall short of expectations, Broadcom’s laser expansion triggers a supply-demand rebalance, or the competitive landscape for OCS deteriorates, there is a risk of valuations reverting to more moderate levels.

 

~~~~~~~~~~~~~~~~~~~~~~~~

The above highlights are from Windy Trading Desk.