Busan-Jinhae’s New Industrial Strategy Meets a Jobs Divide

BJFEZ is seeking to add AI, data and advanced energy to its priority industries. In 2024, revenue across its existing strategic sectors rose 44 percent while employment fell 9.9 percent, exposing the distance between attracting investment and converting it into operating capacity and local jobs.

The Busan-Jinhae Economic Free Zone is preparing to redraw the industrial map it has used for the past five years. In September, the authority submitted a new set of proposed strategic industries built around advanced shipbuilding, maritime transport and logistics, future mobility, next-generation energy, and AI and data convergence. The proposal still requires review by the Ministry of Trade, Industry and Energy and approval by the national Free Economic Zone Committee, with a final decision expected later this year. It would replace a framework adopted in 2021, when Busan-Jinhae was assigned four priority sectors: integrated logistics and transport, smart transportation equipment, advanced materials, parts and equipment, and bio-healthcare.

The proposal connects the zone to industries receiving substantial national policy support, including artificial intelligence infrastructure, advanced energy and cleaner transport. Busan-Jinhae can also point to projects that give some of those ambitions a physical presence, from Microsoft’s data-center investment to manufacturers seeking roles in new energy supply chains. The economic record left by the previous strategy, however, does not move in a single direction.

The free economic zone as a whole continued to expand in 2024. The number of businesses operating inside it rose from 2,266 to 2,442, total employment increased from 62,645 to 63,144, and annual investment also moved higher. Those figures do not describe an industrial district in broad contraction. They show a zone that was still adding companies and, on balance, workers.

A different pattern appeared inside the group of businesses classified as the zone’s strategic industries. Their number barely changed, rising from 481 in 2023 to 484 in 2024, while their combined revenue increased 44 percent, from about 6.25 trillion won to 9.0 trillion won. Employment moved the other way. The same group reported 13,826 workers in 2024, down from 15,353 a year earlier, a decline of 1,527 people, or 9.9 percent.

A growing zone, a shrinking strategic workforce

The fall in strategic-industry employment needs a narrower reading than the headline number invites. Busan-Jinhae did not lose 1,527 jobs overall; total employment across the free economic zone increased by 499. The contraction occurred within the businesses counted under a policy classification designed to identify the industries the zone was trying to develop most deliberately. A region-wide employment trend and the employment record of a selected industrial portfolio were moving in different directions at the same time.

The composition of that portfolio was highly uneven. Of the 484 businesses classified under the four existing strategic industries in 2024, 268 were in smart transportation equipment and 144 in integrated logistics and transport. Advanced materials, parts and equipment accounted for another 70 companies, while bio-healthcare had only two. More than 85 percent of the businesses in the strategic-industry group therefore sat in transportation equipment and logistics, even though the four sectors appeared as parallel categories in the administrative framework.

An aggregate decline can conceal very different conditions across those industries. A reduction in employment at a handful of large manufacturers would have a different meaning from smaller cuts spread across logistics operators, just as rapid revenue growth at a few capital-intensive companies could lift the total without producing comparable hiring. The publicly released statistics do not provide the four-sector employment breakdown needed to distinguish among those possibilities. They identify where the divergence occurred, but they do not identify its cause.

Another official dataset points in a similar direction without resolving the attribution problem. Foreign-invested businesses in Busan-Jinhae employed 14,974 people in 2024, down from 16,202 a year earlier, while their combined revenue rose by 10.1 percent. The foreign-invested group and the strategic-industry group are different populations, even though some companies belong to both, so the two employment declines cannot be added together or treated as evidence of the same corporate decisions. The second dataset shows only that stronger revenue and weaker employment also appeared in a separate group of businesses operating inside the zone.

Company-level records provide fragments of the industrial pressure underneath the aggregates. Korens EM, one of the companies tied to Busan’s transition toward electric-vehicle components, had originally been associated with an investment plan of roughly 352.6 billion won and 370 direct jobs. By late 2024, reported cumulative investment was about 145.8 billion won, while employment had fallen from roughly 205 people in September 2023 to 164. The company attributed part of the gap to changes in its expected business with BMW and the broader slowdown in the electric-vehicle market.

Government statistics do not disclose how much Korens EM contributed, if at all, to the 1,527-job decline in the strategic-industry aggregate. Its own trajectory nevertheless documents pressure inside the transportation transition that Busan had been trying to accelerate: investment remained below the original plan, headcount declined, and the supplier cluster envisioned around the company developed more slowly than expected. The case can illuminate conditions inside the broader portfolio without serving as an explanation for the aggregate employment loss.

The larger statistical pattern remains more consequential than any single company. Between 2023 and 2024, the number of strategic-industry businesses was essentially unchanged, their revenue increased sharply, and their workforce contracted by almost one in ten. Stronger sales, changes in the mix of firms, higher productivity, outsourcing or adjustments concentrated at a few large employers could all contribute to such a pattern. The published tables are not detailed enough to determine which of those mechanisms mattered most.

The 2026 redesign is arriving after a period in which two valid measures of the same regional economy moved differently. Busan-Jinhae as a whole added companies and workers, while the businesses inside its strategic-industry classification generated substantially more revenue with fewer employees. The investment projects cited alongside the new strategy introduce a separate measurement problem because an agreement, a construction project, an operating facility and an actual local job can represent different stages of the same investment.

What counts as an investment success?

In announcing the proposed industrial overhaul, BJFEZ cited 24 major investment-attraction achievements associated with its strategic industries. Microsoft, Renault Korea, Shinsung ST and Prestige Biopharma’s Innovative Discovery Center were among the examples identified by the authority. The material reviewed for this article does not include a complete list of all 24 projects, the period over which they were counted, the criteria for inclusion or a standardized account of the stage each project had reached.

Without that breakdown, the 24 cannot be treated as 24 completed industrial projects or 24 sources of current employment. Public records for individual investments show several distinct stages: agreements can be followed by capital commitments, construction, completion, operation and hiring, while some projects change schedule or direction before reaching all of them. Busan-Jinhae has identifiable projects at each of those stages.

Prestige Biopharma’s Innovative Discovery Center is among the clearest cases in which an investment progressed into an operating facility. The project was originally associated with roughly 211.9 billion won in investment and a plan for 116 jobs, many of them for researchers with advanced degrees. The center was completed in October 2024 and opened the following month. In a May 2025 interview, Prestige Biopharma IDC chief executive Ko Sang-seok said the Busan facility employed 110 people, roughly 90 of them with master’s or doctoral degrees.

The original 116 positions were a project target rather than a same-date employment benchmark, so the later headcount should not be converted into a formal completion rate. The record does show a facility moving through construction into operation with a workforce close in scale to the employment originally associated with the project. That is a different economic stage from an agreement whose jobs remain dependent on future construction or production.

LX Pantos shows an earlier stage of the process. The company’s New Port Eco Logistics Center was announced with roughly 100 billion won in investment and a plan for 188 new jobs. Construction began in June 2025, giving the project a physical footprint beyond the original agreement. Busan Port Authority has said full operations are expected in 2027, which means the 188 positions still belong to the project’s future operating model rather than the current employment base.

DP World’s planned Busan New Port logistics center adds the complication of time. The company committed about $50 million to the project and advanced through formal agreements and foreign-investment procedures, with an initial operating target in the second quarter of 2026. By July 2026, DP World’s Asia-Pacific leadership was describing an opening in the first quarter of 2027. The investment remains active, but the period between attracting the capital and receiving its operating and employment effects has lengthened.

Projects requiring land, permits, design, construction and specialized equipment routinely develop over several years. The timing becomes important when an investment announcement is placed beside a current employment number, because the two measurements may describe different points in the project’s life. Capital attracted in one year can remain in construction while companies already operating elsewhere in the zone are hiring, holding headcount steady or reducing it.

Shinsung ST illustrates a larger change in execution. In 2024, the battery-component maker agreed to invest 46.3 billion won in a smart factory in Busan’s Mieum Foreign Investment Zone and outlined plans to employ 60 local workers. The project was also tied to a broader plan to consolidate operations in Busan. Yet Shinsung ST’s year-end 2025 corporate disclosures continued to list its principal Korean operating sites in Changwon, Chilgok and Haman, with no Busan facility among the main locations, while the company was directing new capital toward a production base in Kentucky.

The company disclosures do not show that the Busan project was formally cancelled. They do show that the factory and consolidation plan had not materialized on the timetable originally associated with the agreement and that significant new investment was meanwhile being directed to the United States. The original agreement remains evidence of an investment commitment; it is not evidence that the planned Busan factory and 60 local positions were already operating.

Several logistics projects provide additional points along the same progression. Mitsui-Soko Korea moved beyond an agreement for an additional Busan New Port warehouse to the arrival of foreign capital, while Naigai’s expanded logistics operation progressed to an opened facility. DWL Global and Access World also established physical logistics centers after earlier investment commitments. Their development histories reinforce the need to date the economic effect of each project rather than count every investment event as though it represented the same stage.

Renault Korea is different again because its Busan factory is already a large operating manufacturer rather than a new entrant. A 2024 agreement envisaged about 118 billion won in investment and 200 new positions as the plant expanded its ability to produce electrified and hybrid vehicles. Busan described the production-base investment as a three-year program running through 2026. Public records reviewed for this article do not establish that the 200 planned positions had already been added to the plant’s workforce.

Across these projects, investment has already produced research facilities, logistics capacity and manufacturing upgrades, while other commitments remain in construction or have moved beyond their original schedules. The employment effect appears later in the sequence and can be altered by project timing, automation, production volume and changes in corporate capital allocation. A count of investment achievements therefore captures a real part of the zone’s development without measuring the amount of operating capacity or current payroll created by those projects.

Investment announcements and strategic-industry employment can consequently move in different directions without either dataset being wrong. New facilities can be under construction while established employers reduce headcount, and a capital-intensive project can raise investment sharply without creating a comparable number of operating jobs. The 24 investment achievements and the 2024 employment decline describe different parts of the industrial system and should not be used as substitutes for one another.

The proposed strategy extends this measurement problem into industries whose economic structures differ substantially. AI infrastructure can require large amounts of capital while employing relatively small operating teams. Advanced-energy manufacturing can enter a new market through certifications, component orders and existing production lines before a large local end market develops. Evidence of both processes is already visible in western Busan.

AI arrives first as infrastructure

Microsoft’s data-center investment gives the AI and data-convergence category a substantial physical base. The company completed a second Busan data center in 2024 after opening its first facility in the region several years earlier, and further cloud infrastructure investment has been discussed with the city. This is operating infrastructure connected to a global cloud provider rather than a prospective industrial project awaiting construction.

The direct employment profile is much thinner than the capital footprint. Microsoft has described its Busan-area data-center workforce as roughly 50 people. Data centers derive much of their economic value from computing capacity, networks and the digital services they support rather than from large operating staffs, so the employment figure by itself does not capture the value of the investment. It does set a boundary on claims about direct job creation.

Regional economic effects would also have to appear outside the data-center payroll. Microsoft’s Data Center Academy provides one channel: the first cohort graduated 28 trainees, and six entered internships at the company’s Busan data-center operations. A practical training lab has since been added. Public material reviewed for this article does not show how many participants moved into permanent jobs, how many local companies became suppliers to the facilities or how extensively firms inside BJFEZ adopted Microsoft cloud and AI services as part of their own production.

Supplier spending, contracting, skills formation and the use of computing services by other companies can all carry the effects of a data center beyond its own workforce. Measuring those connections requires evidence different from the investment value of the facility or its direct headcount. Publicly available records currently describe the infrastructure more clearly than the network of local economic relationships around it.

Busan is also developing computing infrastructure aimed more directly at existing industries. A shipbuilding and maritime edge-AI data center planned for the Myeongji and Noksan area is designed to support demonstrations using high-performance GPUs and domestic neural-processing units. The project could connect computing capacity to maritime and manufacturing applications, but the current public record does not yet show a measurable group of local firms whose employment, sales or productivity can be attributed to that infrastructure.

The AI and data-convergence proposal therefore begins with assets that already exist alongside economic effects that are less completely measured. Microsoft’s facilities establish that Busan can host major computing infrastructure. The extent to which that capacity is entering the production systems, supplier relationships and employment structures of companies elsewhere in the zone remains much less visible.

Advanced energy grows out of an older industrial base

The next-generation energy category begins from a different industrial base. Busan does not yet host a large commercial SMR industry, but it does contain manufacturers, engineering companies and component suppliers built around conventional nuclear power, heavy industry and precision manufacturing. Several of those companies are already attempting to carry existing capabilities into advanced-reactor markets.

Taewoong provides the clearest commercial example found in the public record. The large forging and steel company operates in western Busan and holds qualified-supplier status with Korea Hydro & Nuclear Power. In 2026, it won an order for a critical component used in the rotating-plug system of TerraPower’s Natrium project in Wyoming, placing a Busan manufacturer inside the supply chain of an advanced reactor under development in the United States.

One commercial order does not establish the size of a regional SMR industry. It does establish that an existing Busan manufacturer has converted nuclear manufacturing credentials into business linked to an advanced-reactor project. The evidence therefore reaches beyond technical potential even though it remains far short of demonstrating a mature cluster.

Other companies occupy earlier stages. Youngpoong Industry, an official BJFEZ tenant in the Hwajeon industrial area, holds Korea Hydro & Nuclear Power supplier qualifications and KEPIC certification and has developed equipment for nuclear facilities. Public records reviewed for this article did not show a specific commercial SMR order for the company. Orion ENC, another BJFEZ tenant whose registered business includes nuclear-reactor-related manufacturing, has moved into SMR development work through a feasibility and technology study with Seoul National University.

Existing manufacturers are relevant to the energy strategy before a large local SMR market exists. Their commercial progress can appear through nuclear-quality certification, qualification for supplier lists, engineering work and component orders. Those indicators differ from the land, construction and payroll measures that are more useful for evaluating a new logistics center or research facility.

Busan’s recent nuclear-industry planning documents identify the same underlying constraints. They describe companies with nuclear-equipment experience while also pointing to shortages of specialist workers, qualification costs, limited access to markets and barriers facing smaller suppliers attempting to enter nuclear supply chains. Proposed support focuses on certification, component development, testing infrastructure, workforce development and access to domestic and overseas customers.

Regional statistics remain less precise than the policy agenda. A recent Busan research report listed 39 member companies in the city’s nuclear-industry association; the larger figure of 112 referred to Busan, Ulsan and South Gyeongsang combined. Some certification figures repeated in regional analyses also originate in older surveys. No current public table consolidates how many BJFEZ companies already meet nuclear-quality requirements, how many hold commercial SMR work, how many people they employ in that work or how much revenue advanced-reactor business generates.

The available evidence therefore supports a narrower description than either “potential industry” or “established SMR cluster.” Nuclear-qualified suppliers already operate in and around BJFEZ, and at least one Busan manufacturer has entered an overseas advanced-reactor project. A broader group of firms remains at earlier stages of certification, engineering development and market entry.

What the public record still cannot show

The limits of the available data become clearer when the old and proposed strategies are placed on the same timeline. The government’s 2024 survey shows that employment across BJFEZ’s existing strategic industries fell by 1,527 people while revenue rose sharply, but the published tables do not show how the loss was divided among transportation equipment, logistics, advanced materials and bio-healthcare. Without that breakdown, the largest employment movement in the dataset cannot be assigned to a particular strategic industry.

The investment record has a similar gap. BJFEZ cites 24 major investment-attraction achievements, but the complete list and a standardized status table are not available in the material reviewed for this article. Projects reconstructed from separate records range from operating facilities with confirmed employees to construction projects, delayed openings and agreements whose original schedules have not yet been realized.

The proposed industries introduce new denominators that are also difficult to see. Nuclear suppliers can be identified through company records, certifications and contracts, but the current number of commercially active SMR suppliers inside BJFEZ is not published as a consolidated statistic. AI infrastructure and its direct workforce can be identified more readily than the local companies supplying it, using its services or changing production because of it.

The missing tables do not erase the outcomes already visible in the record. They limit the claims that can be attached to each indicator. Capital investment measures money committed or spent, company counts measure the presence of firms, revenue measures output, and employment captures one part of the effect on the local labor market. Movement in one measure does not establish equivalent movement in the others.

The existing BJFEZ data already show that separation. Strategic-industry revenue rose sharply in 2024 while the number of firms barely changed and employment fell. Several large investments are now operating, others remain under construction, and some employment commitments extend beyond the dates on which the investments were first announced. AI and next-generation energy likewise begin from different assets and produce different kinds of early evidence.

In western Busan, Taewoong has already moved from nuclear qualification into a commercial order tied to TerraPower’s Natrium reactor project in Wyoming. The order covers a critical component for the reactor’s rotating-plug system.