The Global AI Bubble Risk and the Fate of Indonesian Tech Stocks

In the capital markets, acronyms like “FAANG,” the “Magnificent Seven,” and “MANGOS” have defined distinct eras of leading the United States (U.S.) tech stocks.

FAANG emerged during the 2013–2017 era as a moniker for internet platform giants Facebook, Apple, Amazon, Netflix, and Google. By 2023, market dominance shifted to the Magnificent Seven (Apple, Microsoft, Amazon, Alphabet, Meta, Nvidia, Tesla), a cohort driven primarily by cloud computing and AI chips.

However, since late 2025, the Magnificent Seven’s performance has diverged, prompting Wall Street to search for the next-generation tech leaders through new acronyms like MANGOS (Meta, Anthropic, Nvidia, Google, OpenAI, SpaceX), the FAB 10, or the AI Big 10.

Behind this revolving door of nicknames, two things stand out lately: U.S. tech stock prices keep hitting record highs, and massive data center construction is springing up around the world. Both are really two faces of the same phenomenon—the enormous euphoria surrounding artificial intelligence (AI).

Global data center capital expenditure is projected to top US$ 1 trillion in 2026, while five U.S. tech giants—Microsoft, Alphabet, Meta, Amazon, and Oracle—plan to spend roughly US$ 725 billion this year on AI infrastructure [1]. Some call this the most significant technology revolution since the birth of the internet; others wonder whether this is in fact a giant “economic bubble” simply waiting for the moment it pops.

And the most important thing, how far could the fallout drag down Indonesian tech stocks?

A History of Euphoria Culminating in a Bursting Bubble

5 Stages of Economic Bubbles

Capital market history repeatedly shows nearly the same pattern. Whenever great euphoria emerges around a new technology or asset, accompanied by the belief that “this time is different”, prices tend to soar far beyond their fundamentals before correcting sharply once expectations fail to materialize as fast as promised.

In the 17th century, the price of a tulip bulb in the Netherlands briefly matched the price of a luxury house before collapsing within weeks.

The 1960s–1970s in the U.S. had the “Nifty Fifty”, a group of blue-chip stocks believed worth buying “at any price”, before a sharp correction during the 1973–1974 recession.

In 2000 it was internet stocks’ turn to collapse: the Nasdaq plunged roughly 78% from its March 2000 peak to its October 2002 trough, and the S&P 500 fell about 50%, taking seven years to recover.

Eight years later, U.S. property prices collapsed in 2008 after years of being seen as something that would “never fall nationally”, triggering a credit crisis that brought down the global financial system.

Although no one can predict exactly when a market bubble will pop, history teaches us that great euphoria is almost always followed by a correction, sooner or later. The same warning signs are now emerging in the debt-fueled expansion behind AI infrastructure.

AI Bubble Concerns That Demand Attention

1. Extreme market concentration. The Magnificent Seven now account for roughly 35% of S&P 500 market capitalization, matching the concentration level of the seven largest stocks at the peak of the 2000 Dot-Com bubble, according to an Oliver Wyman analysis [2].

   
   

2. Valuations at extreme levels. The Shiller P/E Ratio (CAPE) briefly topped 40 in 2025 (39.8 in December, versus a historical average of 17.7), a level previously seen only right before the Dot-Com crash.

   
   

3. Funding that circulates within a closed loop (circular deals). Nvidia had planned to invest up to US$ 100 billion in OpenAI, with part of that used by OpenAI to buy infrastructure from Nvidia itself. OpenAI also signed a US$ 300 billion contract with Oracle, which then bought hundreds of thousands of Nvidia chips to fulfill it. Critics worry this closed-loop cash flow creates the appearance of rapid revenue growth that may not reflect genuine end-user demand [3].

   
   

4. A gap between data center capex and monetization. Throughout 2025, new AI-based services generated only about US$ 25 billion in global revenue, compared with data center infrastructure spending exceeding US$ 250 billion—meaning only about 10% of total capital spending has so far translated into revenue. Goldman Sachs has warned that annual profits of more than US$ 1 trillion would be needed by 2026 for this investment to break even, far above the consensus estimate of roughly US$ 450 billion [4].

   
   

5. Skeptical voices are coming from within the industry itself. According to various sources, OpenAI CEO Sam Altman has himself acknowledged an element of bubble in the AI sector. JPMorgan CEO Jamie Dimon has warned of potentially large losses in an AI-triggered crash, while Bridgewater Associates founder Ray Dalio sees similarities with the Dot-Com bubble pattern.

These worries are not merely theoretical. Throughout 2026, markets have already shown several times how quickly confidence in the AI narrative can crack. Sharp sell-offs wiping out market capitalization within days are evidence that the risk of a sharp correction is real, not just a theoretical possibility.

Counterargument: The AI Euphoria May Not Be an Economic Bubble

On the other hand, quite a few parties argue that the AI hype does not yet—or may not even—meet the criteria of an economic bubble. Several key points include:

1. Real profitability, not just promises. Unlike the dot-com era, when many companies were not yet profitable, the Magnificent Seven now post a combined net margin above 25% (versus an S&P 500 average of just 13%). Nvidia recorded trailing-12-month profit of roughly US$199 billion with a net margin of about 55%, trading at a P/E of around 31–32 times — far below Cisco’s 472-times ratio at the peak of the dot-com bubble in 2000.

   
   

2. Financed by operating cash flow, not speculative debt. Microsoft, Alphabet, Meta, and Amazon have so far financed most of their AI capital spending from their own operating cash flow, rather than new share issuance or speculative debt as many dot-com-era startups did.

   
   

3. Genuine demand behind the infrastructure. Nvidia claims to hold roughly US$1 trillion in confirmed AI chip demand through 2027, equivalent to more than five times its fiscal year 2026 revenue, with its Blackwell chips reportedly sold out through mid-2026 [5].

   
   

4. Major institutions lean cautious rather than alarmed. Goldman Sachs says the market is “not yet” in an AI bubble, Morgan Stanley calls the concerns “off the mark,” and JPMorgan concludes AI has not met the classic criteria of a financial bubble [6][7].

Market sentiment can reverse within days before recovering again. High volatility does not mean a bubble has definitely burst, but it is also no guarantee that the risk has disappeared.

What a Potential AI Bubble Means for Indonesia’s Tech Sector

The major players in Indonesia’s technology sector are currently dominated by digital consumer apps and e-commerce services, rather than chip manufacturers or AI infrastructure providers. Therefore, the correlation between the global AI bubble and the local stock market is not particularly significant.

From a fundamental perspective, GOTO just made history by posting its first-ever quarterly net profit in the first quarter of 2026, worth Rp171 billion, supported by revenue that jumped 26% year-on-year to Rp5.34 trillion.

Indonesian internet stocks for 2026 are driven by an overall improving profitability trend, including the recovery of EMTK/Vidio and the turnaround at Bukalapak. Through 2025, the IDX Technology Sector index also surged 138.35%, the best performance on the Indonesia Stock Exchange.

Even so, some stocks’ price performance remains vulnerable to technical dynamics such as liquidity and free-float composition, a reminder that its short-term direction is driven more by domestic market mechanics than by global AI sentiment.

For investors in Indonesia, a disciplined approach grounded in understanding business models is far more important than following either the euphoria or the momentary panic.

References

[1] AI Giants Add $350 Billion Debt as $725 Billion Spending Surges, https://finance.yahoo.com/technology/ai/articles/ai-giants-add-350-billion-195553515.html

[2] What happens if the AI bubble bursts?, https://www.oliverwyman.com/our-expertise/insights/2026/jan/impact-ai-bubble-burst-on-global-financial-markets.html

[3] A Guide to the Circular Deals Underpinning the AI Bloom, https://www.bloomberg.com/graphics/2026-ai-circular-deals/

[4] AI may generate only half the profit needed to justify the investment, Goldman analyst warns, https://finance.yahoo.com/news/tech-companies-may-only-half-115338788.html

[5] Nvidia sales opportunity for Blackwell, Rubin chips more than $1 trillion by 2027, https://www.channelnewsasia.com/business/nvidia-sales-opportunity-blackwell-rubin-chips-more-1-trillion-2027-5999816

[6] How are we evaluating the AI boom?, https://am.jpmorgan.com/us/en/asset-management/adv/insights/market-insights/market-updates/on-the-minds-of-investors/how-are-we-evaluating-the-ai-boom/

[7] AI infrastructure pullback offers buying opportunity, Morgan Stanley says, https://seekingalpha.com/news/4618637-ai-infrastructure-pullback-offers-buying-opportunity-morgan-stanley-says

---

Disclaimer: This content is created for educational purposes or service promotion, and does not constitute a recommendation to buy or sell any specific Securities. Any risks arising from investment decisions made based on the information in this publication are the sole responsibility of the respective audience. PT KAF Sekuritas Indonesia is licensed and supervised by the Financial Services Authority (Otoritas Jasa Keuangan / OJK).