---
title: "The AI computing power competition has entered a new stage! The internet firmly stands in the \"C position\" as Morgan Stanley downgrades Corning, but Fabrinet does not change the bullish trend of optical communication stocks"
type: "News"
locale: "en"
url: "https://longbridge.com/en/news/283114497.md"
description: "In the context of AI infrastructure construction, optical communication stocks such as Corning and Fabrinet have been affected by JP Morgan's downgrade of their ratings. Nevertheless, the role of optical modules is becoming increasingly important, and optical communication stocks are still expected to maintain strength in the US AI hardware sector. JP Morgan downgraded the ratings of Corning and Fabrinet from \"Overweight\" to \"Neutral,\" but at the same time raised the target price, indicating a long-term optimism for optical communication stocks"
datetime: "2026-04-17T08:46:08.000Z"
locales:
  - [zh-CN](https://longbridge.com/zh-CN/news/283114497.md)
  - [en](https://longbridge.com/en/news/283114497.md)
  - [zh-HK](https://longbridge.com/zh-HK/news/283114497.md)
---

# The AI computing power competition has entered a new stage! The internet firmly stands in the "C position" as Morgan Stanley downgrades Corning, but Fabrinet does not change the bullish trend of optical communication stocks

According to the Zhitong Finance APP, as U.S. optical communication stocks have recently shown strong performance, Wall Street giant JP Morgan has downgraded the ratings of Corning (GLW.US) and Fabrinet (FN.US), two optical communication stocks. However, investors need not worry about the prospects of optical communication stocks. Against the backdrop of AI infrastructure construction shifting from single-point computing power competition to cluster efficiency games, the role of optical modules is becoming increasingly critical. The optical communication stocks that have continuously reached new highs this year may continue to hold the "pole position" in the U.S. AI hardware sector.

## Morgan Stanley Downgrades Corning and Fabrinet Ratings While Significantly Raising Target Prices

**Morgan Stanley has downgraded the ratings of Corning and Fabrinet from "Overweight" to "Neutral," citing overvaluation and limited visibility of performance as the reasons.** As a result, while other optical communication stocks surged on Thursday—Lumentum (LITE.US) rose over 8%, Applied Optoelectronics (AAOI.US) rose over 10%, and Coherent (COHR.US) rose over 6%—Corning and Fabrinet fell by 1.30% and 1.92%, respectively, in U.S. trading on Thursday.

It is worth mentioning that **Morgan Stanley significantly raised the target prices for these two stocks while downgrading their ratings.** The firm raised its target price for Corning from $115 to $175, which represents about a 5% upside from the stock's closing price of $166.08 on Thursday; it raised the target price for Fabrinet from $530 to $700, which represents about a 4% upside from the stock's closing price of $672.64 on Thursday.

Specifically, the main reason for Morgan Stanley's downgrade of Corning's rating is the stock's overvaluation. The analysis team led by analyst Samik Chatterjee stated in a client report: "We believe that this valuation presents a challenging execution hurdle compared to the earnings expectations embedded by current buy-side investors to justify such a high premium."

The analysts added: "We believe that investors are increasingly shifting their focus to the 2028 outlook and incorporating some degree of idealized scenario forecasting across multiple variables, such as fiber cable/connector pricing and scale expansion opportunities, which leaves almost no room for error regarding capacity risks and the linearity of the optical communication scaling process, not to mention that the company still has about 60% of its business related to non-optical communication markets."

Despite the downgrade, Morgan Stanley has raised its revenue forecast for Corning. The firm increased its full-year revenue forecast for 2026 from $18.6 billion to $19 billion, raised its 2027 revenue forecast from $20.9 billion to $21.7 billion, and provided its first revenue forecast for 2028 at $25.1 billion.

The reason for Morgan Stanley's downgrade of Fabrinet's rating is the "increased volatility" in the short-term customer capacity ramp-up progress, and the limited visibility of the ramp-up pace for new customers in the future. The analysts pointed out: "We expect that the combined effect of these factors will lead to the stock's short-term upside being lower than current buy-side expectations, although we still maintain a positive view on the company's long-term development trajectory, especially considering the manufacturing layout expansion that the company is currently promoting Morgan Stanley has raised its revenue and earnings per share forecasts for Fabrinet. The firm increased its fiscal year 2027 revenue forecast from $5.5 billion to $5.9 billion, raised its fiscal year 2028 forecast from $6.3 billion to $7.1 billion, and provided its first revenue forecast for fiscal year 2029 at $8.5 billion. The firm also raised its fiscal year 2027 earnings per share forecast from $16.65 to $18, increased its fiscal year 2028 earnings per share forecast from $19.40 to $22, and provided its first earnings per share forecast for fiscal year 2029 at $26.50.

Analysts added: "Our updated forecasts are primarily driven by the upside potential in the optical communications business revenue, which includes telecommunications/DCI (data center interconnect) and data communication (Datacom) businesses, expected to achieve compound annual growth rates of approximately 20% and over 30%, respectively, during the forecast period."

Corning and Fabrinet are among the U.S. optical communication stocks that have recently shown strong performance. Morgan Stanley's downgrade of the ratings for these two stocks, while raising the target prices, also indirectly confirms the recent impressive gains in optical communication stocks (which some investors may view as having risen too much) and the general optimistic expectations for the future development of the optical communication industry.

## Continuing to Hit New Highs! U.S. Optical Communication Stocks Show Strong Performance

At the beginning of 2026, global capital markets remain focused on artificial intelligence. While most investors are still debating whether AI computing power has already exhausted growth expectations for the next three years, and media headlines are still dominated by NVIDIA's (NVDA.US) new generation chip launch, the U.S. optical communication sector has quietly emerged with a wave of independent strong performance.

![11.png](https://imageproxy.pbkrs.com/https://img.zhitongcaijing.com/image/20260417/1776414614766770.png?x-oss-process=image/auto-orient,1/interlace,1/resize,w_1440,h_1440/quality,q_95/format,jpg)

Data shows that leading optical communication companies such as Lumentum, Applied Optoelectronics, Coherent, Corning, and Fabrinet have seen their stock prices hit new historical highs this year, significantly outperforming the Nasdaq Composite Index and other AI giants.

The main reason for the recent surge in optical communication stocks is a motivating statement from Lumentum CEO Michael Hurlston. Hurlston stated last Friday: "The capital expenditures of several major cloud providers in the U.S. are extremely large, and there seems to be no signs of slowing down. Our production capacity is increasingly unable to keep up with demand. If the current trend continues, in two more quarters, our capacity for the entire year of 2028 will be completely sold out."

Lumentum had previously disclosed that its production capacity through the end of 2027 has already been fully booked. Therefore, Hurlston's latest statement reinforces market confidence in the optical communication industry's prosperity: even against the backdrop of the Middle East war disrupting the oil market and the global economy, demand for data center equipment remains strong.

Hurlston added: "This situation cannot last forever; it is unrealistic. But for now, this round of the industry cycle can at least maintain its prosperity for about another five years." "When we say that production capacity has been sold out, we mean that irrevocable order agreements have been signed. This is very crucial."

## Paradigm Shift - From "Chips" to "Connectivity"

The strength of optical communication stocks is not merely a rotation of market sentiment, but a long-underestimated industrial logic that is beginning to be realized. Against the backdrop of the exponential expansion of large model scales, the real bottleneck has shifted from "computing power chips" to "computing power connectivity." As the data flow between GPUs grows geometrically, optical modules are no longer supporting roles but have become the "veins and nerves" of AI infrastructure. The construction of AI infrastructure has entered a new deep-water zone—shifting from competition in single-point computing power to a game of cluster efficiency.

In the past two years, the market has generally believed that computing power equals power; whoever has the most GPUs holds the pricing power of the AI era. Computing power suppliers, represented by NVIDIA, have pushed single-card performance to physical limits. From A100 to H100, and then to the next-generation Blackwell platform, the market's focus has always been on the core metric of "trillions of calculations per second." Investors have become accustomed to measuring a tech giant's capability by computing power density, as if sufficient computing power would naturally lead to intelligence.

However, the real industrial reality is that when data centers enter the stages of tens of thousands and hundreds of thousands of cards, computing power no longer depends on single-card performance but on the data exchange efficiency within the cluster. In large-scale distributed training, thousands of GPUs need to work in coordination, and the communication overhead between them becomes the key to determining overall efficiency. In a tens of thousands of card AI training cluster, the internal east-west traffic far exceeds traditional cloud computing loads. Data synchronization, gradient backpropagation, and parameter updates between GPUs are all consuming bandwidth at a frantic pace.

If we compare GPUs to brains, then optical modules are the nerve fibers connecting these brains. When the number of brains increases to a certain extent, if the transmission speed of the nerve fibers cannot keep up, no matter how many brains there are, they cannot form a synergy. This is the famous "communication wall" problem. In the early stages of AI training, communication overhead might only account for 10% of the total time, but in the era of models with hundreds of billions or even trillions of parameters, this proportion could reach 30% or more. Once communication is hindered, expensive GPUs will be left idle, resulting in a huge waste of computing power.

Taking Meta, Microsoft, Google, and other ultra-large-scale data centers as examples, the bandwidth demand within a single AI data center has rapidly jumped from 400G to 800G and has begun large-scale deployment of 1.6T optical modules. Industry estimates show that in AI training scenarios, every $1 of GPU investment often requires nearly $0.5 of network and optical interconnect infrastructure. This means that optical modules are no longer a supplementary cost but have become a rigid prerequisite for computing power expansion.

Against this backdrop, the optical module industry is experiencing a structural turning point, driving the optical communication sector to become "the brightest star" in this year's U.S. stock AI hardware track. Unlike the traditional cyclical expansion of cloud computing, AI data centers exhibit an "exponential upgrade" in demand for high bandwidth, low latency, and low power consumption. Copper cable transmission suffers excessive loss at high rates and is distance-limited. Optical communication technology is precisely the only solution currently capable of scaling to accommodate this surge in traffic When the computing power density increases tenfold, the demand for optical interconnection does not rise linearly, but rather non-linearly. This is because as the scale of clusters increases, the complexity of connections between nodes grows quadratically. Therefore, the growth rate of the optical communication industry is expected to surpass that of computing power chips in the coming years.

## From Cyclical Products to Strategic Assets: Re-evaluation of the Optical Module Business Model

Historically, optical modules have long been viewed as "communication cyclical stocks." Demand fluctuates with the capital expenditures of telecom operators, profit margins are constrained by fierce price competition, and the valuation center has lingered in the manufacturing sector, typically between 15 to 20 times the price-to-earnings ratio.

However, the advent of the AI era has completely changed the pricing logic of the industry. First, there has been a leap in product structure. The technical threshold for 800G and 1.6T optical modules is far higher than that of earlier 100G/200G products. High-speed rates impose extremely high requirements on silicon photonics technology, packaging processes, and thermal management capabilities. For example, silicon photonics solutions require higher integration, and CPO (Co-Packaged Optics) technology necessitates deep collaboration with chip manufacturers. As a result, industry concentration has rapidly increased, and manufacturers with vertical integration and R&D capabilities have begun to command premiums.

Lumentum, with its deep accumulation in the fields of lasers and optical components, has become a core supplier of high-speed modules for several cloud companies. Coherent has been deeply engaged in the fields of photonic chips and advanced packaging for many years, with significant technical barriers that are difficult to replace easily. Corning has formed a monopolistic advantage in optical fiber materials and data center cabling, mastering the discourse power of underlying materials at the physical layer.

More critically, the customer structure has undergone fundamental changes. In the past, optical modules were primarily aimed at telecom operators, but now the core demand comes from large-scale cloud service providers and AI companies. These customers have larger order volumes and faster product iterations. Although they have strong bargaining power, they require extremely high supply stability. Once integrated into the core supply chain, long-term cooperative relationships often form, even leading to joint development of the next generation of products. This binding relationship significantly reduces customer churn rates and enhances revenue visibility.

As a result, the market has begun to reprice—optical modules are no longer low-margin "OEM products," but rather irreplaceable key components in AI infrastructure, with their business model shifting from "manufacturing-driven" to "technology-driven." Compared to the leading GPU companies, which often have price-to-earnings ratios exceeding 30 times, optical communication companies have long had their price-to-earnings ratios suppressed below 20 times. When fundamentals and industry positions are re-evaluated, valuation expansion and profit growth create a double impact.

Of course, the market is not without skepticism. Opponents argue that AI investments ultimately experience cyclical fluctuations. If the demand for large model training slows down, or if algorithm efficiency improvements lead to a decrease in demand for computing power, capital expenditures in data centers may decline, and the demand for optical modules could repeat the past cycle of "oversupply—price reduction—profit compression."

However, the fundamental difference in this round of investment compared to the past is that AI computing power construction has strategic attributes, rather than being purely commercial expansion. Whether it is the competition among American tech giants or the global investment in sovereign computing power, AI infrastructure is becoming a long-term national and enterprise-level strategic investment. Such investments are rigid and will not easily stop due to short-term economic fluctuations More importantly, the technological generational upgrade has not yet ended. From 800G to 1.6T, and then to the future 3.2T, each bandwidth upgrade signifies a replacement demand for existing equipment. The optical module industry has for the first time a structural growth curve similar to the "continuous iteration" of semiconductors. In the traditional telecommunications era, an iteration cycle might last five years. However, in the AI era, the iteration cycle for optical modules has shortened to about 18 months. This means that even if the total volume does not grow, structural upgrades can still bring continuous revenue momentum

### Related Stocks

- [GLW.US](https://longbridge.com/en/quote/GLW.US.md)
- [FN.US](https://longbridge.com/en/quote/FN.US.md)

## Related News & Research

- [Corning stock is falling. Why JP Morgan disagrees with the 'blue-sky scenario.'](https://longbridge.com/en/news/283041060.md)
- [What You Need to Know Ahead of Corning’s Earnings Release](https://longbridge.com/en/news/282715024.md)
- [What Allbirds needs to do to make its Hail Mary AI pivot succeed](https://longbridge.com/en/news/283043758.md)
- [These Allbirds AI jokes are as fire as the company's stock price](https://longbridge.com/en/news/282954479.md)
- [2 Undervalued AI Stocks That Could Skyrocket Soon](https://longbridge.com/en/news/282991032.md)