Why Stop Steel Dumping While Ignoring Software Dumping?

Machbase is a venture company that has spent the past 13 years developing a new kind of software—a time-series database—from the ground up, gathering every resource we could and working to create a market for it.

Recently, while preparing a proposal for an airport corporation, we noticed the time-series database requirements in an earlier specification.

The original document named two foreign time-series database products, “InfluxDB” and “TimeScaleDB.” Unsurprisingly, Machbase was not listed.

Our absence from that list is not the issue. How much marketing capacity can a small domestic software company really have?

The problem is that those foreign software products are free.

Under those conditions, every proposal from a domestic systems integrator—whether a large corporation, a mid-sized company, or a small business—will propose that free foreign software.

There is no room for Machbase to enter the competition.

Fortunately, after learning that a domestic time-series database existed, the airport corporation prepared the RFP using a generic description rather than vendor names. That gave us an opportunity as well.

But does something not seem wrong here?

When low-priced Chinese steel enters the country, the state moves swiftly to initiate anti-dumping proceedings.

• Korea Trade Commission, Ministry of Trade, Industry and Energy — Anti-Dumping System

Yet when an American software company backed by hundreds of billions of won from major Silicon Valley venture capital firms distributes its software for free, no one says a word.

In other words, when cheap imported steel encroaches on the domestic market, the government immediately considers anti-dumping duties. Trade remedies are triggered when goods imported below their normal value cause injury to a domestic industry and a causal relationship between the low-priced imports and that injury is established. The Korea Trade Commission explicitly lists imports of dumped goods, material injury to a domestic industry producing like products, and a causal link between the two as requirements for an anti-dumping investigation. When steel is sold below cost, the state steps in to protect the industry. Why, then, is the distribution of entire software products for free—undermining the foundations of a domestic industry—not examined by the same standard?

Of course, applying the existing anti-dumping framework directly to software is not straightforward. Anti-dumping rules were originally designed for physical goods crossing borders. They require comparisons of prices, import volumes, origin, normal value, and like domestic products. Software, however, is downloaded or delivered through the cloud. It is simultaneously a service and intellectual property, sometimes a platform, and sometimes an entire data ecosystem. The WTO also distinguishes several modes of trade in services, including cross-border supply, consumption abroad, commercial presence, and the movement of natural persons. Digital technology has fundamentally changed how goods, services, and intellectual property are produced and distributed across borders.

The central issue, however, is economic impact, not legal form. Consider a situation in which giant foreign technology companies use enormous capital resources, cloud infrastructure, government research support, and ecosystem power to supply core software for free or at extremely low prices. At first, this clearly looks beneficial to users. Businesses and public institutions save on purchase costs. Developers can readily adopt new technologies. Budget managers can report immediate savings as an achievement. Over time, however, domestic software companies lose maintenance revenue, reduce their R&D staff, and eventually leave the market. What remains is not free software. It is structural dependence on a particular platform, cloud, API, and data ecosystem.

This is not fundamentally different from the logic of dumping that we have long seen in manufacturing. The unit in which it operates is different. Steel dumping is measured in tonnes; software dumping operates at the ecosystem level. Steel dumping may reduce factory utilization. Software dumping weakens a country’s ability to build and retain its own technological capabilities. Steel dumping becomes visible relatively quickly through indicators such as inventory and prices. Software dumping becomes apparent much later, through the collapse of standards capabilities, talent, data resources, operational expertise, and maintenance systems. That makes it more dangerous: the roots of the industry wither out of sight.

The simple belief that “free is good” is therefore dangerous. A price of zero does not mean a cost of zero. Costs do not disappear; they move elsewhere. Technical support, security updates, incident response, cloud migration, data migration, and retraining staff tied to a particular technology all return later like bills coming due. What was initially free may become the most expensive choice. Competition law has long addressed this kind of structure. Low prices generally benefit consumers. But if a dominant firm can drive competitors out through below-cost pricing and then recover its losses over time through excessive prices or platform power, consumers ultimately suffer. Free software can certainly stimulate innovation across an industry. Equally, major firms can use free distribution to eliminate competitors and later reap the benefits of market power.

This does not mean that all open-source or free software should be condemned as dumping. Open source is an important foundation of innovation that has supported the modern software industry. From Linux and databases to development tools and AI frameworks, it has underpinned technological progress worldwide. The issue is not whether the source is open; it is the effect on the market. Genuine technology sharing, community-led innovation, and mutually beneficial ecosystems should be protected. But the situation changes when free distribution is combined with state subsidies, aggressive strategies backed by major venture capital to dismantle a particular software sector, monopolistic cloud infrastructure, closed APIs, control of de facto standards, and strategies to capture public markets. This may be a strategy for market domination rather than technology sharing. It may be a new form of dumping that undermines the domestic industrial base rather than innovation.

The following table shows the funding raised by well-known overseas software companies that have positioned themselves as open-source businesses.

Rank	Company	Flagship products	Total funding (USD)	Approximate KRW equivalent
1	Cloudera	CDH, Cloudera Data Platform	$1.1 billion	KRW 1.5 trillion
2	Confluent	Confluent Platform	$900 million	KRW 1.2 trillion
3	HashiCorp	Terraform, Vault, Consul	$349 million	KRW 480 billion
4	MongoDB	MongoDB Enterprise	$311 million	KRW 430 billion
5	InfluxData	InfluxDB Enterprise	$239 million	KRW 330 billion (a competitor of Machbase)
6	Elastic	Elasticsearch, Kibana	$162 million	KRW 220 billion
7	Cockroach Labs	CockroachDB	About $630 million	KRW 860 billion (would rank around third based on more recent funding)
8	Redis Ltd.	Redis Enterprise	About $350 million	KRW 480 billion (similar to HashiCorp based on more recent funding)
9	DataStax	DataStax Enterprise	About $340 million	KRW 470 billion
10	Percona	Percona Server, Percona XtraDB	About $170 million	KRW 230 billion

Backed by these enormous sums, these companies have presented their products as free software for more than a decade, effectively carrying out software dumping worldwide. Over time, they accelerate commercialization by changing policy: moving from open source to commercial offerings, changing licenses, or strengthening Enterprise features.

• Notice of the commercialization of previously free Hadoop through Cloudera

• The 2023 controversy over the end of open-source clustering in InfluxDB / The 2025 shift toward commercialization of on-premises offerings

Existing trade rules already contain the seeds of regulation that could curb such practices. While anti-dumping measures respond to price discrimination, countervailing duties address situations in which financial contributions or price support from governments or public bodies benefit particular companies or industries and cause injury to a domestic industry. The Korea Trade Commission likewise explains countervailing duties in terms of financial contributions, income or price support, and benefits limited to specific companies or industries.

This is where a new discussion must begin in the digital era. If technologies subsidized by foreign governments, free software cross-subsidized by major platforms, loss leaders used to sell cloud services, and free AI models and databases intended to secure market power materially damage domestic industry, they cannot be viewed solely as consumer benefits. The difficulty of directly applying traditional anti-dumping duties does not justify doing nothing. It is time to discuss a new policy category: “digital dumping” or “injury to the software industry.”

A policy response does not necessarily mean imposing tariffs. Software requires far more sophisticated instruments. Public procurement should not consider only whether software is free. It should also assess total cost of ownership, availability of technical support, security accountability, the accumulation of domestic technological expertise, data sovereignty, and the risk of long-term dependence. National R&D and public infrastructure projects should require arrangements that allow domestic companies to participate in maintenance, expansion, training, and the development of local technological capabilities, even when free foreign software is used. For critical infrastructure software, the availability of domestic alternatives, the ability to verify source code, and sovereign control over incident response should be assessed separately. The question is not simply whether software is cheap to use today. What matters more is whether, in five or ten years, we will still be able to repair, operate, and improve it ourselves.

At the WTO, tariffs on electronic transmissions are no longer a distant debate. Since the adoption of the Work Programme on Electronic Commerce in 1998, members had renewed the moratorium on customs duties on electronic transmissions roughly every two years. However, at the 14th Ministerial Conference in Cameroon in 2026, the moratorium expired without a multilateral agreement. Some members, including South Korea, announced interim implementation of a plurilateral e-commerce agreement and began pursuing a separate path. This clearly demonstrates that digital trade has moved from the margins of the trade agenda to its center.

• WTO MC14 Post-Briefing Note – E-commerce

• What is the World Trade Organization e-commerce moratorium that stops tariffs on digital downloads and streaming?

• Announcement by the Office of the United States Trade Representative (USTR)

South Korea is a manufacturing powerhouse. But the manufacturing industries of the future cannot move forward without software. Factories, power plants, railways, batteries, semiconductors, defense systems, robots, and AI infrastructure all run on software. Software is no longer an accessory to industry. It is its brain, nervous system, and operating system. Leaving such a critical area entirely dependent on free foreign supply may save money in the short term, but it is not a long-term industrial strategy. Today’s budget may be smaller. But if the price is the loss of technological sovereignty and the industrial ecosystem, the choice is far from cheap.

The criteria for fair trade must change as well. In the past, inspecting the goods inside a container was enough. From now on, we must also examine clouds, APIs, data, AI models, databases, operating systems, and development platforms. A zero on the price tag does not mean zero harm to the market. The most powerful form of dumping may arrive without any visible price at all. Behind it may lie far greater costs: control of standards, data dependence, the erosion of technical talent, and the exit of domestic companies.

We prevent steel dumping not to protect a handful of steel companies, but to preserve the industrial base itself. Software is no different. The software industry is no longer peripheral; it is foundational to national competitiveness. We must now ask:

When is free software innovation, and when is it digital dumping?