The global fiber optic gap soars to 40%! "Price changes daily," after a 10x increase by Longi, who will be the next to multiply tenfold?

February 8+ and March 25 I wrote two in-depth research reports, saying:[Taoguba]
** 《Global Fiber Optic Shortage 30%-40%! “Price per day”—Is the next CPO cycle coming?》**
** 《Global Fiber Optic in Emergency! “Price per day”—the gap hits 30%! China’s 4 major giants monopolize 60% of the market!》**
But back then, a bunch of people blasted us, saying: talking about this again at high levels!
But for any trend and industry, only when it erupts can people really see it!
And what we’re thinking about isn’t just value—we also need to see future growth and explosive momentum!
And in the end, the leader: ZTT even completed the 10x myth in one year—though chasing it now definitely isn’t right. So the real question is: who might become the next “follow-on rally” player?

Just like last year: innovative drugs—Sintai 10x! Yipin 4x, Guangzhou Biosth? 8x!—and there were also follow-on rallies in the back rows!
So don’t just look at the K-line. If you think deeply: industry development variables + company value + trend!

I. Why have fiber optics and cables been continuously blasting higher against the trend recently? ZTT Fiber Optic even completed a 10x move in one year!

The fiber optics and cables industry is going through a super-cycle driven by the “AI compute revolution,” with a “rising together in both volume and price” surge. Its breakout is the result of multiple factors—demand, supply, technology, and policy—resonating at the same time. ZTT Fiber Optic, with its unique moat, has become the most extreme beneficiary in this round of the market.

1、Demand side: AI compute exploding exponentially (the main engine)
North American CSPs (Microsoft/Amazon/Google) are shifting capex toward AI clusters in 2026-2028. For a single 10k-card (10k) GPU cluster, fiber demand is 5-10x that of traditional IDC
Interconnection inside data centers (DCI) + short-reach multimode (OM4/OM5) demand surges, and the share of AI fiber rises from <5% in 2024 to 35% in 2027
Overseas orders: high certainty, low price sensitivity, long-cycle lock-in—directly pulling up both quantity and pricing across the industry

2、Supply side: capacity is rigidly constrained + price skyrockets (the strongest catalyst)
The fiber preform expansion cycle is 2 years. In the past few years the industry was sluggish and almost no expansion happened. Globally, there are about 2.5 hundred thousand tons of fiber preforms, so in the short term volume can’t be ramped up
2025Q4-2026.2: G.652D stray fibers jump from 25 yuan to 50 yuan per fiber-kilometer (doubling in 2 months), and some quotes reach 85-120 yuan
Domestic operators’ centralized procurement has had multiple failed bids (price caps lower than market prices); the March quarterly reports confirm the pricing increase’s sustainability

3、Technology side: commercial breakthroughs in hollow-core fiber (valuation re-creation)
ZTT hollow-core fiber: 0.04dB/km (lowest globally), latency -31%, speed +47%, far beyond the Microsoft/Lumenisity line (0.091dB/km)
For supercomputers/high-frequency trading/backbone networks, hardcore demand starts the replacement cycle from solid-core to hollow-core
4、Key events (accelerating fuse in Jan-Mar 2026)

Hollow-core fiber procurement for overseas CSP AI clusters is realized—ZTT/HenTong win large orders
Domestic operators’ centralized procurement prices rise by 100%+ (e.g., Heilongjiang Telecom, etc.)
OFC conference (March): 1.6T/3.2T optical modules, hollow-core fiber, liquid cooling—far beyond expectations
ZTT’s 2026Q1 AI orders +187.3%, confirming a profit-margin inflection point

II. Can fiber optics and cables keep exploding in the future?

The answer is yes. The industry’s high level of business conditions has staying power, and the growth structure will evolve toward higher-end offerings.

1、Pricing logic: the capacity expansion cycle > the demand surge cycle time mismatch:
The fiber preform expansion cycle lasts as long as 1.5-2 years, while AI compute and overseas demand explosions are immediate. In 2026, there will be no effective new capacity additions in the industry, but demand growth is expected to exceed 25%, leading to a quantified supply-demand gap of 138 million fiber-kilometers (gap ratio 16.7%).

Cost rigidity: The fiber preform cost is relatively fixed (about 35 yuan per fiber-kilometer). Almost all of the price increase converts into gross margin.

2、Far-reaching impact: earnings leverage far exceeds revenue leverage—gross margin multiplies:
Calculations show that per-fiber-kilometer gross profit can jump from the bottom level of 6 yuan to the expected 110 yuan, with an increase as high as 1733%.
Pattern optimization: Smaller manufacturers are constrained by capital and approval, and their willingness to expand capacity is low; leading manufacturers, leveraging fiber preform self-sufficiency and overseas layout, will capture the vast majority of the pricing upside, further increasing industry concentration.
Valuation rework: The industry’s nature shifts from “cyclical manufacturing” to “core assets of compute infrastructure,” and the valuation framework has the potential to align with the AI compute hardware segment.

3、Global supply-demand gap—full panorama estimates (unit: 25k fiber-kilometers/year)

Detailed estimates:

4、Why will the gap keep widening?

Demand side: the threefold surge is irreversible
AI data center demand explodes: a single 100k-card (10,000-card) level AI cluster needs 100k fiber-kilometers; globally, in 2026 there will be 50+ new AI clusters → 5 million+ fiber-kilometers in demand (doubling versus 2025)
Key validation: The fiber preform/cable capacity of Corning, Fujikura, and Preussman was already running at full production capacity in 2025Q2, and they start buying from China—proving AI demand has spilled over globally
Drone private network demand is above expectations: Cru data shows demand of 50 million fiber-kilometers in 2025 → 80 million+ in 2026 ( +60% ), and actual could be higher (requests from leading companies like DJI and EHang may not be fully included)
US Bead project accelerates: Federal government allocates 60 billion dollars and requires 100% fiber coverage—demand in 2026 will reach 40 million fiber-kilometers (5% of global)

Supply side: capacity expansion is hard, and gets harder
Preform production bottleneck: fiber manufacturing’s core is the “preform” (pre-made preforms), accounting for 60% of costs—requires big investment (1.6T+ yuan for 10,000 tons of preforms) and has a long cycle (2-3 years)

Current extreme capacity:
Domestic four major preform producers’ capacity: 4.8 hundred thousand tons/year → 500 million fiber-kilometers/year
Industry-wide maximum capacity: with “squeezing processes,” at most another 10-15% → 550-575 million fiber-kilometers
The dilemma of adding capacity:
Since 2025Q3, there has been no new preform capacity plan (long payback cycle)
International giants like Corning and Fujikura in Japan hit full production in 2025Q2 and turn to purchasing from China
Key conclusion: a gap of typically 10%+ usually leads to price hikes of 50%+. This gap is 16.4%, which explains why the 657 A2 price breaks through 100 yuan/fiber-kilometer (up 150% from the 2024 low of 40 yuan). The gap will keep going until end of 2027, because the fastest fiber preform capacity release is at end of 2027.

Core contradiction: capacity buildout lags behind the demand surge, and in 2026 the supply-demand gap will reach its peak.
5、Capacity expansion cycle:

6、Research on current capacity status of leading manufacturers

III. Why is it ZTT that completed a 10x move in one year? Not someone else?

1、Preform process: the only global method with “three-in-one” (the core moat)
Three-process matrix vs. peers comparison

2、Comparison of hollow-core fiber performance gap

Hollow-core fiber: performance gap leading by 2 years
ZTT: 0.04dB/km, 91.2km single preform, commercial deployment achieved
Overseas: Lumenisity (Microsoft) 0.091dB/km, laboratory stage
Domestic: HenTong/CAM? developing in-house; maybe mass production only after 2027
3、Global capacity + customers: overseas AI orders dominate
Overseas revenue 42.7%; North America CSP core supplier
Delivery capability: 8 global bases, with no geopolitical supply cutoff risk

4、 Costs and yield: nonlinear earnings leverage
The industry’s #1 in drawing length → unit cost -15%-30%
High yield and low scrap rate → gross margin elasticity exceeds peers by 2-3 percentage points

IV. Who could become the next ZTT fiber optics—or the next “follow-on rally” leader?

1、In-depth comparison

2、Follow-on rally path and catalyst time axis

3、And one more surprise!

Summary: fiber optics and cables are not just a simple “price-hike cycle”—it’s a fundamental reconfiguration of the demand structure—
It’s also the AI compute-driven 2026-2028 super-cycle: supply-demand gaps + price surges + technological iterations, with ongoing explosive momentum for more than 2 years.
ZTT 10x: three-process monopoly + absolute lead in hollow-core + global capacity + volume-price resonance—unreplicable. But thinking about a follow-on rally might be even more worth deep thought! Like in CPO, after Cambridge, and after Yi Zhongtian! So that’s why I said: fiber optics is a CPO-style way of thinking!
After reading, can you understand who the follow-on rally stock will be? Will the price increase keep going? Who will be the most valuable in the future? But it’s not recommended to chase the rally—maybe a pullback is a better pattern! But it still needs reasonable risk control!
If you don’t know, please hit like + share + comment: **The global fiber optic gap skyrockets to 40%! “Price per day”—after ZTT’s 10x, who takes over the next 10x? **

The above is only my personal trading review and summary of my thinking. Investing involves risk; trading requires caution! Plans never change faster than reality—everything must be based on the market action. The article content is my personal thoughts and records; it serves only as a personal sharing record of how I understand the market, and does not constitute any investment advice. For reference only. Any buying/selling based on it is your responsibility for profits and losses!
(It’s not easy to dig up and整理资料; your: likes + shares + comments are the motivation for our effort—thank you!）

Overall market situation: On Monday this week, I reminded everyone—after the turning point at No. 1, we should be bearish until around the 7th. The adjustments over the last two days were to prove our judgment is correct. So I suggest reducing exposure to within 50%. The recent question is correct. And today, the number of stocks falling is 4,459, while only 698 are up. Trading volume has also fallen to 1.65 trillion, and it’s getting harder and harder with that volume. However, after risk keeps declining, we overall see support at 3,850, or possibly 3,794. As long as these levels aren’t broken, it’s still okay—so just wait patiently!

Sentiment: sentiment is at an extreme low. There are 36 stocks hitting the daily limit, 24 hitting the floor limit, with a board-closure rate of 75%, and the total number of stocks in consecutive boards is 4, with as high as 6 boards.
On sectors: the current core main line is innovative drugs, while the strongest sector is: compute power engineering + innovative medicine + robotics! Overall compute power engineering keeps strengthening, and innovative drugs are now starting to diverge—whether to strengthen is key. At the moment, the overall center is this, and other defensive themes have no appeal anymore! Supporting plays: chemical industry + power + chips!
Pharma: the leader is Jinyao? (Jin Ya? Pharmaceutical) with 6 boards, while the runner-up Zhongyao? with 2 boards has become the leader, and in the back row it’s not necessarily sustainable—so we need further observation.
Compute power engineering: the leader is Huiyuan with 2 boards; and right now, the core is that ZTT is the strongest core,
Robotics: the leader is Rongtai with 1 board, while back-row plays like Falan? have爆发, but it’s not clear if it’s sustainable—so we can only observe further!
There’s no shortcut to success—only self-discipline and persistence. Wishing you—those who keep striving— the more you work hard, the luckier you get!

Today’s highlights:
1、Cyberspace Administration solicits public comments: any organizations and individuals that provide or use digital virtual human services should respect intellectual property and business ethics, and must not infringe on the intellectual property that others legally enjoy
2、A French-flagged container ship sends a signal indicating it has exited the Strait of Hormuz.
3、At the 2026 Space Compute Industry Conference, the “Space Compute Professional Committee” for the compute industry development faction will be established.
4、Nine departments including the Ministry of Commerce jointly issued the 《2026 Work Plan for Actions to Enhance Services Consumption and Benefit the People》

Tracking commodity themes

1、生猪 9.98（-0.49%，pig cost 12, Muyuan 11.6 yuan, feed over 50%, labor 10-15%.)
2、Battery-grade lithium carbonate 151k（-1.23%，salt lake cost 30-40k, mica 60-90k, spodumene? 60-80k.)
3、Praseodymium neodymium oxide 722.5k（0.21%）, NdFeB N35: 1.7M（0%）（, light cost 350k-400k. MP790k, Australia 450-500k. Mining: 8,000–10,000 yuan/ton, medium-heavy rare earth 1.30–1.50 million yuan/ton）
Like first then watch—monthly income in the millions! Thank you all for your support!