[Edaily Marketin Song Seung-Hyeon Reporter] As policy funds, such as the National Growth Fund, are being injected on a large scale into artificial intelligence (AI) companies, concerns are growing among limited partners (LPs) about an AI valuation bubble. There have even been cases where the formation of funds focused on AI investments has fallen through, as private investors have hesitated to match the valuations driven up by policy funds.

According to venture capital (VC) and investment banking (IB) industry sources on the 22nd, a VC firm recently attempted to form a project fund to invest in AI companies but was reportedly rejected by its limited partners (LPs). The LPs reportedly refused to invest, asking, “Is there nothing else besides AI?” and judging that the valuations of AI companies were already excessively high, showing signs of a bubble. AI, which had been considered the most favorable sector for fundraising, has instead become a source of caution for private LPs.

Industry observers attribute this sentiment among LPs to the rapid concentration of investment flows toward AI in recent years. According to THE VC, a startup investment data platform, while total domestic startup investment in 2024 fell by 19.7% year-over-year, investment in the AI sector rose by 41% to 966.6 billion won. The share of AI investments, which had remained between 8% and 9% from 2021 to 2023, rose to 16% in 2024 and 23.6% last year. In the first half of this year, investment in AI and robotics surged 485.2% year-over-year to 2.685 trillion won, and its share of total investment expanded to 34.3%.

Analysts note that this concentration is accelerating as large-scale investments from policy funds—such as the National Growth Fund—have joined the steady flow of private capital shifting toward AI, even during the investment winter. Funds from the National Growth Fund have been heavily concentrated in the so-called “Big Three of AI”—Rebellion, Furiosa AI, and Upstage.

Rebellion completed a 640 billion won pre-IPO round last March, securing a valuation of approximately 3.4 trillion won. Of this amount, 300 billion won came from the National Growth Fund (250 billion won) and the Korea Development Bank (50 billion won). Furiosa AI is currently conducting a pre-IPO round worth 700 to 850 billion won based on a valuation of approximately 3 trillion won, with the National Growth Fund and the Korea Development Bank (KDB) allocating 400 billion won. Of the 560 billion won raised by Upstage, 130 billion won came from the High-Tech Strategic Industries Fund and KDB. In major funding rounds for leading AI companies, policy funds have accounted for as much as half of the total, playing a leading role in establishing their valuations.

The problem is that their performance does not justify these valuations. Last year, Rebellion posted revenue of 32 billion won and an operating loss of 120.5 billion won. Furiosa AI recorded revenue of 5.7 billion won and an operating loss of 62.5 billion won last year. Upstage posted revenue of 24.8 billion won last year, yet its market value at the time of its IPO is estimated to be as high as 5 trillion won. This translates to a market value 100 to 500 times its revenue. This is why critics point out that even though policy funds participate in funding rounds alongside private investors, the influx of massive capital all at once has left little room to validate these valuations.

As government funds drive AI investment, various institutional investment programs appear to be following suit. Korea Venture Investment’s first regular contribution to its fund-of-funds this year totaled 1.63 trillion won, a 63% increase from the previous year, with the AI convergence and deep tech sectors accounting for 47% of that amount. The Korea Post set a condition for its AI venture fund investment in the second half of the year requiring fund managers to invest at least 200% of the capital contributed into the AI value chain.

The investment industry is concerned about whether the valuations that AI sector companies have achieved at the unlisted stage can be sustained in the public offering market. While early-stage investors can still realize a profit even if the public offering price falls below their recent investment cost, late-stage investors who entered at high valuations cannot even be sure of recouping their principal.

An investment industry official stated, “Policy funds have driven up valuations, making it burdensome for the private sector to follow suit,” adding, “If private capital enters at already inflated prices and the IPO price is set below that level at the time of listing, the losses will fall entirely on private LPs.” He continued, “The fact that LPs are asking, ‘Is there nothing besides AI?’ is ultimately because returns are uncertain,” adding, “It seems we’re starting to see a trend where they want to diversify their portfolios into sectors other than AI.”