In recent years, the environment for domestic investment attraction has been changing. While some universal, nationwide tax incentive policies remain in place, local "subsidy-based" policies tied to investment amounts, taxes, and output value are gradually being phased out.
In the long run, this is a positive development that effectively breaks down market fragmentation, builds a unified national market, and reduces internal friction. However, in a sense, the only cards left in local governments' hands for investment attraction are now "investment" and "application scenarios (market)". This undoubtedly benefits first-mover cities that have already formed scaled industries, while latecomer local governments face new challenges in developing the semiconductor industry locally.
Two Major Challenges Facing Local Investment
In fact, for many latecomer cities, it is still too early to talk about "marketization" and "application scenarios." The local industrial chain is incomplete, and application scenarios are relatively limited. Unless there are strong state-owned enterprises with significant capabilities in relevant fields, local governments can play very little substantive role in this regard. Moreover, even if a locality has leading enterprises, mismatches between government investment attraction needs and the qualified supplier requirements of these leading enterprises can make it difficult for both parties to align their steps.
Therefore, the primary focus for most localities remains on "investment."
Semiconductor investment has been booming in the past two years, with many semiconductor companies successfully going public. Stories like "how much money a certain local government made by investing in semiconductor projects" often become popular topics, as if everyone were Warren Buffett or Duan Yongping. However, the real barrier to entry for semiconductor investment is much higher than it appears; it is an extremely professional task.
In the past, relying on subsidy policies, local governments played a role more akin to NBA team owners, who could distribute bonuses at the end of the season based on player performance—companies had to meet metrics such as revenue, taxes, and job creation first, after which the government would fulfill its promised subsidies. This allowed them to "not shoot until they see the rabbit," ensuring clear and balanced accounting.
Investment, however, transforms the local government's role from team owner to "scout." The government must make judgments before the company generates output, deciding whether the future enterprise will be an "All-Star" or a "benchwarmer." Furthermore, after investing, the government sometimes assumes part of the responsibility for "cultivating the enterprise"—when companies encounter issues with production lines, suppliers, etc., they seek government assistance. This requires local governments not only to perform investment tasks but also to act as managers.
In a sense, this is not a strength of local governments.
Equity investment itself carries high risk, and the semiconductor industry, due to its high technical barriers and rapid iteration, often entails even greater risks. Even top-tier investment institutions commonly find that out of ten projects invested in during one fund cycle, only two or three ultimately generate profit.
Moreover, local governments face two additional major challenges when making investments compared to general investment institutions.
First, they cannot calculate the overall account simply. A market-oriented fund may invest in ten projects, succeed in two, and lose in eight, yet still be considered successful if the portfolio is profitable overall. However, local governments manage public resources and find it difficult to accept issues of "loss of state-owned assets." Sometimes, succeeding in nine projects and failing in one can easily result in a situation where "10 - 1 = 0."
Second, there is a mismatch in cycles. General investment institutions are relatively tolerant regarding the time horizon for return on investment, whereas government funds often enter assessment and exit phases within 3-5 years. However, the actual return cycle for semiconductor projects typically starts at 10 years. There is a natural mismatch in time scales between the two.
The increased "caution" in local government investment has positive implications, helping to avoid blind project initiation, low-level redundant construction, and major pitfalls. However, in the context of the rapidly evolving semiconductor industry, this caution imposes far higher demands on local governments' familiarity with the industry, professional judgment, and patience than ever before.
The "Relocation Round" Easily Leads to "Weak-Weak Alliances"
As the saying goes, "Those who hear the Way come first or last, and each has their own expertise." Besides the difficulty of investment, local governments face an additional layer of consideration compared to general investment institutions—they must not only select good projects but also ensure those projects take root locally.
For general investment institutions, investment is not geographically bound; the core metric is return. However, for local governments, making money alone is insufficient; they must also drive local employment and supply chain coordination. Therefore, they often form "wanting this, wanting that, and wanting more" demands on enterprises—typically requiring companies to relocate their headquarters, R&D, or production to the locality.
This gives rise to what is colloquially known in the industry as the "Relocation Round": local governments use billions in strategic equity investments as leverage and are willing to absorb existing shares. Some company founders approaching the deadline of valuation adjustment mechanisms (VAM), finding IPOs difficult, introduce local government strategic investors to absorb old shares. However, local governments also require enterprises to relocate, sometimes imposing VAMs on revenue, capacity, and IPO timelines, with buyback clauses if targets are not met.
In the past two years, practices resembling "takeover-style relocations" have emerged, where local guidance funds acquire projects facing exit difficulties at discounted prices, conditional on the enterprise moving its registered address and committing to tax payments for the coming years. This model inadvertently amplifies subsequent exit risks.
However, "relocation" is not easy to execute. With the vigorous development of the domestic semiconductor industry, when we speak with entrepreneurs, we sometimes joke that local investment promotion bureaus watch their leading enterprises even more closely than "they watch their wives," fearing they might be poached by others at any moment.
Consequently, truly worthwhile enterprises rarely relocate, and those willing to negotiate are mostly weaker firms.
First-tier enterprises have many options and do not lack government financing. Some investments seeking to enter impose no additional conditions, so adding another round of local strategic investment is merely icing on the cake, resulting in extremely low willingness to relocate. More critically, the R&D backbone, core customers, and upstream/downstream supply chains of such enterprises are largely accumulated in their original locations. The cost of a full relocation far exceeds the equity premium offered by local governments.
Second-tier enterprises are often the "face" of the industry in their location. Even if the enterprise itself has reasons to move, the current location generally offers nearly equivalent or better terms to retain them, leaving few cases that can actually be swayed.
Thus, those willing to negotiate and movable often fall into the third or even fourth tier.
The problem with this segment of enterprises is: their inherent industrial driving force is limited. Order volumes cannot support a complete upstream production line, talent density is insufficient to attract supporting manufacturers to follow them in relocation, and their technology paths cannot establish regional standards.
When such an enterprise meets a latecomer locality with an incomplete industrial chain, it easily evolves into a "weak-weak alliance." The locality transplants a "semiconductor" sapling but must constantly "water and weed" it. In reality, these enterprises often fail to bring upstream and downstream supply chains to invest alongside them; instead, they face the issue of "transplanted trees dying"—having failed to thrive in regions with better industrial ecosystems, they grow even slower after relocating due to insufficient local application scenarios. After several years, substantial fiscal funds have been spent, but the local industrial chain has not truly grown new branches.
Sometimes, even if relocation is negotiated, it results only in "registered address landing"—core R&D, customers, and talent do not actually move, and the enterprise's main business and revenue generation remain in the original location. The locality ends up holding nothing but a nameplate.
ICwise Viewpoints
Facing the dual challenges of investment and localization, the way forward for latecomer local governments lies in the following key points.
First, enhance professionalism. Semiconductor investment is not an ordinary fiscal project; decisions must be made by teams and mechanisms that truly understand the industry. Collaborating with professional investment funds, establishing industrial expert review mechanisms, and jointly investing with genuine industrial capital are realistic paths to bolster professionalism. The ability to identify good projects and detect fake ones is a prerequisite for success.
Second, accept the spillover effects from major cities and do not immediately target the largest, hottest industrial chains. The semiconductor industry is highly concentrated. Leading manufacturing, design, and equipment companies will likely continue to converge in a few first-tier cities, determined by industrial laws. Rather than forcing competition for head-track sectors, latecomer local governments should proactively capture supporting segments and niche leaders spilling over from first-tier cities. Although these "spillovers" are not flashy, they are often easier to genuinely root locally.
Third, shift from "external relocation" to "internal activation." Compared to grabbing projects from outside, building up the local business environment, innovation ecosystem, and talent policies is often a steadier and more sustainable path. Work such as innovation entrepreneurship competitions, local project discovery, and talent recruitment must be done solidly, especially guiding existing traditional local enterprises toward industrial upgrading and transformation. Enterprises that grow naturally within the locality are inherently rooted there and do not face "exit" issues.
Fourth, find semiconductor interfaces within local advantages. Starting from scratch in semiconductors poses an extremely high barrier for latecomer local governments. However, if they can find connection points between existing local industrial advantages and geographical endowments and semiconductors—for example, environmental protection, energy, chemicals, or specialized manufacturing—to capture corresponding materials, equipment, or application segments in the semiconductor supply chain, it becomes a less resistant entry path. When local traditional industries can serve as downstream applications for high-tech industries, the roots of the industrial chain are essentially planted.
Epilogue
Finally, there is a higher-level and more fundamental issue—the assessment mechanism.
Once local governments embark on the path of fund-based investment attraction and equity investment, they are essentially doing the work of investors. Therefore, the assessment of latecomer local governments perhaps should change along with the nature of this activity.
In the future, rather than focusing on the profit or loss of each individual project, success could be defined by the overall account: if state-owned assets achieve preservation and appreciation, and the local industrial chain develops, it is considered relatively successful. As industrial challenges escalate to this stage, local governments have been pushed into the position of investors. Thus, their evaluation methods should upgrade accordingly.