Companies Gaining Attention Amidst Intensifying AI Competition: Listings, Investment, and New Strong Players

Anthropic is pursuing a listing to raise $100 billion with a valuation of $2 trillion, and news has emerged that OpenAI is also preparing for an IPO. The reason AI companies are successively pursuing listings is to be recognized with the maximum corporate value based on high growth rates, secure large-scale funds, and solidify market dominance.

Securing funds to gain an early lead in the AI market is essential, and funds raised through listings will be reinvested in R&D and infrastructure expansion to strengthen future competitiveness.

Based on Annual Recurring Revenue (ARR), Anthropic has surpassed OpenAI, emerging as a new strong player in the AI market. Anthropic launched Claude Opus 3.5, ushering in the agent AI era, which led to an increase in market share.

In response, OpenAI launched new products such as GPT-4o, o1 AI model series, and ChatGPT Work. OpenAI is rapidly increasing its user base and recovering market share through strategies to secure computing capacity and improve token cost-effectiveness.

The fierce competition between the two companies is driving the advancement of AI model performance and cost efficiency, acting as a factor in stimulating the overall AI market.

When used via cloud APIs, open-weight models cost only about one-fifth of frontier AI models. This high cost-effectiveness is a significant factor in AI model selection.

In terms of performance, the gap between open-weight models and frontier AI models is narrowing, and driven by this trend, the token share of open-weight models, including Chinese AI models, has risen to 50%. This demonstrates a trend of companies switching to cost-effective AI solutions.

Cloud companies are reducing the model usage fees they previously paid to frontier AI models like Anthropic or OpenAI. This has become possible by adopting open-weight models or their own self-developed AI models.

As frontier AI model usage fees decrease, cloud companies are expected to realize even greater profits. Developing proprietary AI models or utilizing open-weight models has the potential to significantly increase cloud companies' net operating profit after tax.

Consequently, as AI model competition intensifies, a structure is being created where cloud companies can generate greater profits through reduced model usage fees and the provision of their own solutions.

SpaceX recently announced its goal of securing 10GW of data center capacity by the end of 2027 in its earnings report. This is a large-scale expansion plan, approximately seven times the 1.4GW as of the second quarter.

This data center expansion will require enormous financial investment. It is estimated that approximately $30 billion to $60 billion will be needed per GW, and SpaceX's large-scale investment will serve as a major driver accelerating the growth of the AI infrastructure market.

SpaceX's aggressive investment once again emphasizes the importance of data centers, key infrastructure in the AI era, and raises expectations for further growth in related industries.

Microsoft has stated its principle of executing capital expenditures (Capex) only within the cash flow generated from its operating activities. This demonstrates its commitment to making investments within a manageable scope while maintaining the company's financial health, rather than indefinite spending.

This cash-flow-based investment principle reflects Microsoft's conservative investment stance, which avoids indiscriminate investment and emphasizes efficiency. This is interpreted as a strategy to promote stable corporate growth while carefully managing funds necessary for AI infrastructure expansion.

Capital expenditure (Capex) by global big tech companies is expected to increase further in the future. In particular, it is projected to increase by approximately 70% year-on-year in 2026 and 30% in 2027.

If the recent trend of rising memory prices is factored in, the actual investment scale could be even higher than anticipated. This indicates that investments in data centers and related facilities for AI infrastructure will continue to expand steadily.

These large-scale investments by big tech companies are essential for expanding infrastructure for AI technology development and will be a major driving force for the growth of the entire AI industry ecosystem in the long term.

Hyperscalers refer to global big tech companies that can invest cash flow generated from non-core businesses into data center capital expenditures (Capex). Beyond the traditional definition, the category is expanding to include companies with self-financing capabilities.

These companies include US big tech firms such as Amazon, Microsoft, Google, and Meta, as well as Chinese companies like Alibaba and Tencent. They are leading the AI era by building and operating vast data center infrastructure.

The redefinition of hyperscalers highlights the importance of companies that can undertake large-scale investments based on their own financial strength in the AI infrastructure market, demonstrating their role as a core driver of the AI industry.

The MAGS ETF adopts an equal-weighted investment strategy instead of the typical market capitalization-weighted approach. This strategy involves investing roughly 14.8% equally across the 7 stocks included in the fund.

The equal-weighted approach is used to prevent concentration in a few large-cap companies and to capture growth opportunities across all included stocks. Rebalancing occurs quarterly to adjust the weighting of each stock, allowing for flexible responses to changing market conditions.

Unlike the ACE ETF mentioned earlier, the Tiger US Tech Top 10 ETF adopts a market capitalization-weighted approach. This means that companies with larger market caps are given a higher investment weighting, a strategy focused on investing in large technology companies that lead the market.

This ETF invests broadly in overall big tech companies, not just hyperscalers, to construct its portfolio. The market capitalization-weighted approach is more sensitive to market trends and is suitable for investors seeking to track the performance of large technology stocks.

Both hyperscalers and neoclaud companies have a business model of building data centers and selling them in a cloud format. In other words, their core business is essentially the same.

However, there are differences between the two groups in terms of scale and funding methods. Hyperscalers operate large-scale data centers and have the capability to self-finance data center construction costs through operating cash flow.

Neoclauds, on the other hand, are classified as smaller companies entering the cloud market. They often partner with companies like Nvidia to secure funding for data center investments.

For US-listed ETFs, the Neoclaud ETF uses a strategy to increase its weighting in key stocks such as Navius and CoreWeave by combining direct investment in underlying shares with Total Return Swaps (TRS). TRS is a total return swap contract, a derivative product that allows investors to gain the returns of a stock without actually holding the shares.

This allows the ETF to effectively increase its investment weighting in specific stocks even with limited funds. Domestic products, on the other hand, exclude hyperscalers and primarily include data center infrastructure companies in their portfolios.

This weighting strategy enables flexible investment tailored to market conditions while capturing the growth potential of neoclaud companies to the fullest extent.

Many former Bitcoin mining companies are successfully transitioning to the cloud and data center market by leveraging their power grid infrastructure. As the growth potential of the mining business faced limitations, they opted for a strategy of utilizing their existing power infrastructure to operate AI data centers.

These companies are rapidly transforming their industries based on established power grids, gaining the advantage of early market entry. This move reflects an awareness of the importance of data centers as core infrastructure in the AI era and an effort to efficiently repurpose existing assets to secure new growth drivers.

This transition goes beyond a mere business model change; it exemplifies how existing mining infrastructure can be re-evaluated as a crucial resource in the age of AI.

There is extreme polarization among companies within the software sector. Legacy software companies like ServiceNow and Salesforce saw their stock prices drop by more than half, recovering slightly recently but still reporting sluggish performance. These companies are facing difficulties, with negative year-to-date returns.

In contrast, cybersecurity companies like Palo Alto Networks and AI-benefiting companies such as Datadog and Snowflake are performing strongly. They provide essential security and data analytics solutions in the AI era, maintaining high growth rates.

This polarization suggests that as the impact of AI technology on industries grows, traditional software companies may lose competitiveness if they fail to adapt to new technological changes. Conversely, companies closely associated with AI are leading the market, recognized for their high growth potential.

Palantir's stock began to move after a surge in quarterly new contract volumes with US companies post-Q4 2024. New contracts have continued to increase significantly in the most recent quarter, maintaining this growth trajectory.

This increase in contracts is expected to translate into future revenue growth as it converts into backlog. Notably, Palantir's 12-month forward EV/EBITDA multiple has declined from around 200x to 60x. This means that while revenue growth is high at 92% quarter-on-quarter, the multiple burden has relatively decreased.

Palantir's multiple expansion is limited due to the emergence of strong competitors. Anthropic and European AI company Mistral have begun to enter the software industry, and Palantir has frequently shown signs of countering them in recent earnings announcements.

With big tech companies like Amazon and Microsoft also entering related businesses, Palantir is no longer a sole provider. This intensifying competition suggests that while Palantir may see stock price increases based on performance, recovering its past monopolistic-driven multiple will be challenging.

The entry of competitors can expand the market pie, but it can also dilute the unique premium of individual companies. Palantir faces the ongoing challenge of maintaining its competitive edge through innovative technology and services.