Startups
A recap of the AI workshop jointly held by SouthBridge and Genspark on August 1st. From university students to opticians and F&B workers, people using AI properly for the first time created association website demos, web games, and business plans on the spot with 8,000 credits.
Nexon is establishing investment corporation Nexon Partners and will invest 250 billion won over five years in early-stage game developers alongside the Ministry of Culture, Sports and Tourism and Kona Venture Partners. This is the first large-scale public-private fund targeting the seed to Series A stage, where game startups are most vulnerable.
Samsung C-Lab Outside's 9th cohort is once again drawing AI and robotics startups into its fold. The moment a founder accepts seed funding and infrastructure from a conglomerate, what exactly do they gain — and what do they silently give away? The two faces of the corporate accelerator ultimately circle back to the question of Korea's startup ecosystem's capacity for self-sustaining growth.
Sovereign AI and 'AI top three' have risen to the level of national strategy, yet what they mean and how to measure them varies from person to person. What happens when the terminology outpaces the policy?
The agent gold rush is cooling. The layers that evaporate above the model and the ones that survive are diverging. What must Korean startups build on top of American models to protect their margins? The problem is not the courage of founders — it is the coordinates of the ecosystem.
An industry built on connecting each other's strengths — seed money of 200 million won each for more than ten domestic fabless startups… AI [data centers] that prioritize the use of domestically made chips…
When you follow the trail from a fragrance warehouse and a Seongdong-gu café to a survey of small factories, it becomes clear that generative AI is not an 'adoption' question for Korea's small-business owners and SME manufacturers — it is a question of 'what works and where does it breaks.' Bank of Korea data puts those boundaries into numbers.
Reading the 2026 IPO rebound as recovery would be a mistake. Statistics only capture companies that survived, while those that quietly died have been erased from the denominator. Down rounds haven't ended—only companies that could withstand down rounds remain on stage.
The 2024 consensus that 'thin wrappers die' was reading the value chain backwards. Margins flow not to models but to those who control entry points and workflows. Which layer Korean companies stand on will determine the next cost structure.
Korean startups have begun choosing Japan and Southeast Asia as their first overseas markets. This is not a retreat, but a structural shift in the order of expansion. The problem lies not with founders, but with an ecosystem still holding onto outdated maps.
As AI drives up capital efficiency for solo founders, the seed round is evaporating entirely. It's not the price that's collapsing—it's the structure of the fundraising market itself. The lessons taught by digital assets are repeating themselves.
The Fund of Funds is the invisible backbone of Korea's startup ecosystem. But the moment the state becomes an investor providing capital, that money both reads and rewrites market signals. Have we ever defined this duality in institutional language?
Open source began as a shared industrial commons, but through license wars, it has solidified into a new moat of trust and operations. Is Korea a country that borrows this infrastructure, or one that builds it?
Software once sold per seat is now shifting to per-outcome pricing. When the era of humans viewing screens ends, the very unit of billing changes.
The narrative of 'a hundred billion won in solo revenue with AI' claims to have eliminated employees. But in reality, it has merely moved employees outside the accounting books. What has disappeared is not labor, but the name of labor.
Startups raising capital and users simultaneously with tokens and points instead of equity are on the rise. The real question isn't the coin price, but who governs a company whose equity is distributed to users and how.
The real enemy of youth startups in Busan is not lack of capital. Founders somehow manage to hire their first employee. Companies collapse at the second hire. And the people who could fill that position have already boarded the KTX to Seoul.
Korean indie developers launching without publishers in 2026 battle not against graphics, but against invisible accounting: refund timers and discount cycles. Break-even points are determined not by download counts, but by friction in design.
A capital structure that measures semiconductor materials and robotics by software timelines. Reading the mismatch between capital recovery periods and technology maturity through the time units of industrial accumulation.
In a country where large corporations don't acquire startups, IPO is the only exit. We examine how that single narrow gate quietly makes founders more conservative in their risk-taking, from an ecosystem structure perspective.
In 2026, the era of buying channels to achieve growth has ended. As Big Tech drives up acquisition costs and AI churns out infinite content, the center of gravity in growth is shifting from acquisition to retention, from channels to product. This shift is restructuring the cost base of Korean startups.
Stablecoins are not new technology but a new plumbing for dollar liquidity. What do Korean remittance and commerce startups gain and lose on this rail opened by regulation?
The moment burnout is translated into an individual's resilience problem, structural failure is written off as one entrepreneur's weakness. Who designed this ledger?
Companies that grade models are occupying a deeper position than companies that build models. The moment trust becomes a bottleneck, evaluation becomes not a tool but an industry.
Reading the offerwall and reward app-tech market solely through user numbers or ad rates misses the core issue. The real point of contention in this market lies in whether issued points are debt, who holds those balances, and who guarantees their value. The structural questions that once confronted the virtual asset market are now directly facing Korea's reward economy.
Korean founders who directly import Silicon Valley playbooks fail differently with the same strategies. The problem isn't the strategy—it's that we copy the answers without translating the questions.
Live ops studios operate a single game forever. That model converts the risk of game entrepreneurship into endless operating costs. Eternal operation means eternal design debt.
The vocabulary Busan uses to attract global talent consists of rent, settlement conditions, and quality of life. These are all the grammar of the catch-up era, calling people with costs. But costs can be replicated anywhere, while the depth of trial and error accumulated in one region cannot be replicated.