First batch of 17 integrated circuit projects under Optics Valley's "Plane Tree Plan" signed, with total contracted investment of over 6 billion yuan

On August 22, at the Wuhan East Lake Hotel, the first batch of 17 projects under the Optics Valley "Phoenix Tree Plan" were signed and landed in a concentrated ceremony.

This was no ordinary signing. Centered on the integrated circuit industry chain, all of the leading domestic enterprises were present: AMEC, Skyverse, Leadmicro, Zhenbao Technology... From lithography, etching, and deposition to polishing, from equipment, components, and metrology and inspection to design, the entire chain was covered, with a total signed value of more than 6 billion yuan.

What is truly worth savoring is the pace after the signing.

In just over two months, among the 17 projects, 14 companies including AMEC have completed registration of their Wuhan subsidiaries, 10 have confirmed their sites and are about to begin construction, and they are expected to be put into use in about a year. At the same time, the second batch of enterprises under the "Phoenix Tree Plan" is accelerating its rush toward Optics Valley.

A single "phoenix tree" has drawn numerous "phoenixes" competing to perch, and is quietly rewriting the ecology of the entire forest.

Overlooking the brilliantly lit Optics Valley Plaza, from here all the way east, an increasing number of technological innovation enterprises have gathered. (Photo by Wei Lai, Hubei Daily all-media reporter)

One tree changes a forest

First, understand what the "Phoenix Tree Plan" is.

It takes the chain-leading storage enterprise as the center and a radius of 4.5 kilometers, concentrating the introduction of core equipment, materials, and chip design enterprises within a 60-square-kilometer area to create a 5-minute industrial collaboration circle; then, relying on the depth of the hundred-billion-level storage industry block, it supports a 30-minute supply circle.

In one sentence: it aims to write a grand article on industrial collaboration in spatial terms.

Why make an article in spatial terms?

To produce a single chip, a wafer fab must go through hundreds of processes including lithography, etching, thin-film deposition, polishing, and metrology and inspection. If any single piece of equipment or any chemical falls short, the entire production line has to stop and wait. In the past, these supporting enterprises were scattered across the country, and a single emergency allocation of spare parts often required traveling thousands of kilometers.

Five minutes means engineers from upstream and downstream enterprises can meet on foot. When a process parameter needs joint debugging, there is no need to book flights or arrange schedules; they can sit down together just by pushing open the door.

Thirty minutes means that from silicon wafers, gases, and polishing slurries to precision components, there is a local supply radius. Wafer fabs no longer have to "wait for express delivery from a thousand kilometers away."

"We are seizing the opportunity to accelerate the improvement and upgrading of the integrated circuit industry chain ecosystem," said the head of the Investment Promotion Bureau of the East Lake High-tech Zone. Supporting this spatial logic is a set of continuously strengthened mechanisms: government-enterprise joint weekly meetings and monthly scheduling working mechanisms; landed enterprises enjoy a dedicated task force for each enterprise; state-owned platforms join with industrial funds to drive projects through investment promotion linkages... Entering a brand-new "uncharted territory"

The chip industry has never been merely a product of the market.

Since the invention of the transistor, no region in the world has grown its integrated circuit industry through market forces alone. The U.S. CHIPS Act, Japan and South Korea's industrial mobilization, and Taiwan, China's park model are all, without exception, cases of the government making the first move.

An effective market determines the efficiency of resource allocation, while a capable government determines the starting point of industrial competition. The chip contest appears on the surface to be a competition among enterprises, but at bottom it is a competition in government organizational capacity.

Wuhan Xinxin building

Hubei made this move early. Twenty years ago, Hubei Province, Wuhan City, and the East Lake High-tech Zone jointly invested 10 billion yuan to found Wuhan Xinxin, built the first 12-inch integrated circuit production line in central China, and planted the first seed of Hubei's chip industry; afterward, the national memory base was located here, leading chip enterprises took root, and the Jiufengshan Laboratory and Jiangcheng Laboratory were successively established, as Optics Valley grew into one of the country's four major integrated circuit industrial bases.

After scale was achieved, shortcomings increasingly became a constraint. Storage manufacturing strength is solid, but high-end chip design, core equipment, and key materials support are relatively weak. The etching machines, polishing slurries, and precision components required by wafer fabs were mostly procured from thousands of kilometers away in the past. The further the industry chain extends toward the high end, the more critical it becomes whether ecosystem partners can live next door.

And a deeper change is taking place: Optics Valley's processes are beginning to reach the world's forefront, in turn "using processes to lead equipment research and development." This means there is no longer ready-made equipment to buy by imitation or mature recipes to copy—China's integrated circuit R&D and production have already entered "uncharted territory."

Wuhan Xinguang Industrial Park, located in the core area of the Optics Valley memory industry innovation block

This is by no means alarmist, nor is it a test unique to Optics Valley. The hardest stretch of "uncharted territory" in semiconductor history was traversed by Intel, Samsung, and TSMC alike.

The most representative case is Intel. When it first settled in Silicon Valley, the place was still an orchard. What truly made it legendary was its almost obsessive seriousness about process: when planar transistors reached their limit and electrons leaked downward like a faucet that could not be shut tightly, it was the first to step forward—first bringing hafnium, a previously rare rare-earth element, into the wafer fab for the first time to replace the leaky "gate"; then changing the transistor from "lying flat" to "standing up" (known in the industry as FinFET). With every step forward, there was no ready-made machine to buy; it could only first invent a process never seen before and then have equipment makers build the corresponding machines accordingly. Once it had paved the way, equipment giants such as Applied Materials, Lam Research, and KLA were drawn in like iron filings to a magnet, conducting joint R&D around its process needs. The technology used today by every foundry in the world is precisely what it pioneered back then.

This points to a conclusion: entering uncharted territory depends not on fighting alone to grab the next point, but on the chain leader taking the lead and pulling equipment, materials, and design close to traverse it together, building the ecosystem first.

In this brand-new "uncharted territory" today, whoever first builds a complete collaborative ecosystem will have the chance to overtake by changing lanes. Optics Valley's "Phoenix Tree Plan" is precisely intended to support this ecosystem.

Three new tracks, each one staked out

He Tingbo of Huawei delivered a keynote speech titled "Exploration and Practice of New Semiconductor Pathways."

For the past 60 years, the global chip industry has been driven by Moore's Law—making transistors ever smaller. But below 3 nanometers, leakage, heat, and soaring costs mean the returns from continuing to shrink dimensions fall off sharply. In May 2026, Huawei proposed the "Tao (τ) Law": no longer fixating on geometric miniaturization, but instead systematically compressing signal propagation delay through logic folding, 3D stacking, and compute-in-memory, so chips can run faster.

Changing lanes means the rules of the game are completely rewritten.

Three new tracks thus become mandatory questions: compute-in-memory, optoelectronic integration, and advanced packaging.

The original intent of the "Phoenix Tree Plan" was precisely compute-in-memory. And these three tracks have long been intertwined: through advanced packaging, compute chips, high-bandwidth memory (HBM), and co-packaged optics (CPO) engines can be packaged into the same package body, forming an integrated "compute + storage + optical interconnect" solution. The industry is ushering in a key inflection point of deep "compute-storage-optoelectronic" integration.

At the end of September this year, China's first 3D advanced packaging mass-production line went online in Optics Valley.

On all three new tracks, Optics Valley is staking out its position—

Compute-in-memory combines the warehouse and the kitchen, so data doesn't have to be moved back and forth. Nationwide, the cities that simultaneously have a storage chain leader enterprise, a 3D integrated manufacturing line, and an advanced packaging mass-production line—and can push compute-in-memory from blueprints to mass production—are few and far between. Relying on the chain leader's core hybrid bonding process, Optics Valley is opening up the entire chain from chip design to 3D integrated manufacturing to packaging and testing.

Optoelectronic integration uses light instead of electricity to transmit data. Optics Valley is one of the world's largest optical communications industry bases; China Information and Communication Technologies Group (China InfoComm), YOFC, and HGTECH have deep roots here, and central China's first silicon photonics chip pilot line has been completed and put into use.

Advanced packaging assembles chips with different functions like Lego bricks. China's first 3D advanced packaging mass-production line has gone online in Optics Valley, becoming a national lighthouse. AI chips from leading design firms elsewhere have already come here specifically for mass-production validation.

On September 22, the Wuhan Municipal Party Committee convened a special meeting and for the first time proposed building integrated circuits into the city's "primary industry."

According to the plan, by 2030, Optics Valley will have built a world-class compute-storage integration industrial base, China's AI chip manufacturing highland, and a source of innovation for compound semiconductors.