Is GMO Internet (TSE:4784) Overvalued Amid Its Turnaround and Global Expansion?

In the high-stakes arena of internet infrastructure, GMO Internet (TSE:4784) has emerged as a compelling case study in the tension between valuation realism and growth potential. The company, which has recently returned to profitability, is trading at a price-to-earnings (P/E) ratio of 51.7x as of November 2025-well above both its Japanese Media industry average of 16.3x and the estimated fair P/E ratio of 41.3x. This premium reflects investor optimism about its AI-driven infrastructure initiatives and global expansion, but it also raises a critical question: Is GMO's valuation justified by its fundamentals, or is it pricing in a future that may not materialize?

Valuation Metrics: A Tale of Two Narratives

GMO's valuation metrics tell a story of extremes. Its P/E ratio of 51.7x is not just high-it is substantially higher than its peers. For context, the median P/E for the Software industry in 2025 is 24.87, and GMO's P/E of 55.08 (TTM) places it in the lower percentile of the sector, meaning it is priced higher than 74.46% of its competitors. Meanwhile, its EV/EBITDA ratio of 6.8x appears modest at first glance, but this metric masks a deeper issue: GMO's EBITDA margin of 24.3% is exceptional for a company in its stage of growth. When adjusted for this margin, the company's EV/EBITDA multiple becomes less favorable, especially when compared to AI infrastructure peers trading at 40x–50x revenue multiples.

The disconnect between GMO's metrics and those of its peers is stark. For instance, Palantir Technologies (NYSE: PLTR) trades at over 700x forward earnings, while Nvidia (NASDAQ: NVDA) and AMD (NASDAQ: AMD) command P/E ratios above 50x and 45x, respectively. These valuations are driven by their dominance in AI hardware and software, but GMO's AI initiatives-such as its "GMO GPU Cloud" for generative AI and robotics-remain unproven at scale. The company's EV/Revenue multiple of 1.6x is low for a firm with such ambitious AI aspirations, yet its P/E ratio suggests investors are betting on a dramatic acceleration in earnings.

Growth Projections: Can GMO Deliver?

Analysts project GMO's revenue and EBITDA to grow at 8.4% and 16.5% annually from 2023 to 2025, respectively. These figures are credible given GMO's 7% LTM revenue growth and 24.3% EBITDA margin. However, the company's PEG ratio of 1.8x-a measure of valuation relative to growth-suggests the stock is overvalued even when accounting for these forecasts. For comparison, SaaS companies in 2025 trade at a median P/E of 6.1x EV/Revenue, while AI infrastructure startups command multiples 5–10x higher. GMO's valuation sits in a gray zone: it is not as speculative as a pre-revenue AI startup but is priced for a level of growth that may require execution beyond its current capabilities.

The company's global expansion adds another layer of complexity. GMO is leveraging its Z.com brand to enter overseas markets and has invested heavily in AI infrastructure, including GPU cloud services according to valuation multiples. These moves align with a broader trend: global AI infrastructure spending is projected to reach $1.5 trillion in 2025, and Japan's government has allocated $135 billion to AI and quantum computing initiatives. Yet, GMO's international ambitions face headwinds. Its EBITDA of $457 million in 2025 is impressive, but scaling this to global markets will require significant capital and operational discipline.

Sector-Specific Dynamics: A High-Stakes Game

The AI infrastructure sector is a double-edged sword. On one hand, it offers explosive growth potential for companies that can establish defensible moats. On the other, it is a capital-intensive, high-margin space where valuations are often disconnected from traditional metrics. GMO's EV/EBITDA of 6.8x is reasonable for a mature infrastructure play but appears undemanding for a firm with AI ambitions. By contrast, late-stage AI startups in core infrastructure trade at 40x–50x revenue, reflecting their perceived strategic value. GMO's focus on GPU cloud services and robotics positions it to benefit from these trends, but its current valuation does not fully reflect the risks of competing with global giants like NVIDIA or Microsoft.

Conclusion: A Calculated Gamble

GMO Internet's valuation is a paradox. It is expensive by traditional metrics but relatively modest for a company with AI-driven growth aspirations. The key question is whether its management can execute on its global expansion and AI infrastructure bets without overextending its balance sheet. If GMO can maintain its 24.3% EBITDA margin while scaling revenue at 8.4% annually, the current valuation could prove justified. However, if GMO's AI initiatives underperform or global expansion proves costlier than anticipated, the stock's 25.7% drop in the past month may be a harbinger of further volatility.

For investors, GMO represents a calculated gamble: a high-margin, cash-generative business with a speculative edge. The market has priced in a turnaround story, but the real test will be whether GMO can deliver the growth to justify its lofty multiples.