Is the Air of 1998 Coming Back? Reading the AI Market from the "Credit Market"

1) Is the "AI Boom" a Repeat of the Dot-Com Era? First, the "Similarities"

When observing the AI market, many people are reminded of the dot-com era from 1999 to 2000. An analysis by WSJ (introduced by Seeking Alpha) points out the similar "rise in enthusiasm" compared to that time. A symbolic example is Cisco, a dominant player during the dot-com era, which has "reached the level" it was at during the peak of March 2000 once again. The point here is not that Cisco is a bad company. Even excellent companies can have stock prices that deviate from reality, which is a "lesson from the market." Motley Fool Community

Indicators showing overvaluation also support this association. For instance, the CAPE (Shiller P/E Ratio) is in the 40s, historically a high level (often reminiscent of the dot-com period). Multpl

Furthermore, the "tectonic shifts" in investment are also similar. During the dot-com era, the excessive investment in communication infrastructure (fiber optics) became a talking point. The WSJ analysis recalls that over "$100 billion" was invested in new communication networks at the time, with parts of it "lying dormant for nearly a decade" due to insufficient demand. Now, data centers and computing resources (GPU/servers) are playing that role—with investment figures even reaching "trillion" levels, to the extent that economists discuss their contribution to GDP growth. Motley Fool Community

2) However, "This Time is Different" is Also True. The Biggest Twist is "There are Profits"

This is why the AI market in 2025 cannot be simply labeled as "Dot-Com 2.0." Fed Chairman Powell clearly drew a line in a press conference, stating "it's different from back then," and noted that companies leading AI investments have "revenue," "business models," and "profits," without naming specific companies. AI investments are led by companies with "revenue," "business models," and "profits". Multpl

This statement spreads easily on social media because it hits the "heart of market sentiment." On LinkedIn, many posts cite Powell's point, arguing that "AI spending is different from dot-com speculation." LinkedIn

In other words, AI is not in a phase where it is bought only for the "story," and at least large companies have cash-generating power. This alone makes the foundation different from the startup proliferation period of 1999.

3) Yet, the Reason Anxiety Doesn't Disappear: Twist ① "Supply-Driven Expansion"

On the other hand, the bubble warning theory remains. For example, investor Michael Burry expresses concern that data center construction and GPU orders are accumulating as "supply-side enthusiasm" rather than "demand." Business Insider

The fear here is the illusion that "increased investment = strong demand." Construction progresses, equipment is installed, and the supply chain thrives. The numbers alone suggest a good economy. However, if there is no accompanying "who" and "how much" will pay to use it, capital investment could become "premature consumption."

4) Twist ② "Credit (Debt) Market" Returns to the Forefront

Another "twist" this time is that the temperature of the bond and credit markets, rather than stocks, is more likely to influence the sustainability of AI investments.

As AI investments expand, the need for financing also increases. Here, companies that can cover costs with cash are strong, but as reliance on external funding grows, market perceptions change. In fact, Greg Jensen of Bridgewater warns of the danger as big tech increasingly relies on "external capital" for AI expansion funding (also noting that funding surged from 2024 to 2025 according to UBS figures). Reuters

Additionally, Mark Marek, CIO of asset management firm Siebert, describes the "trillion-dollar promise" of AI as clashing with "the inconvenient math of debt and CAPEX (capital expenditures)," arguing that we are returning to an era where balance sheets matter more than stories. We are returning to an era where balance sheets matter more than stories. blog.siebert.com

"Stocks can buy dreams, but bonds require interest and principal to be repaid." Structurally, as credit perspectives become stricter, AI investments become more tied to the "cost of funds."

For reference, analyses in the VC community also note that corporate bond issuance by hyperscalers significantly increased in 2025 (note: this should be read as an aggregation and interpretation of public information, not primary data). Tomasz Tunguz

5) Twist ③ "Individual Speculation" Has Returned

One of the "final fuels" that drove the frenzy of the dot-com era was individual speculation. In posts quoting excerpts from WSJ analysis in investor communities, it is noted that the phenomenon of "individuals dominating trades and betting on small, loss-making stocks" is once again prominent. Motley Fool Community

What's interesting here is the difference in small-cap indices. For example, the S&P SmallCap 600 has profit requirements (a simple "profit filter") such as "positive earnings in the most recent quarter and the cumulative four most recent quarters," while a broader collection of small-cap stocks includes many loss-making companies. Schwab Brokerage

In short, when funds are more likely to gather in stocks that "have more dreams because they are not profitable," the market becomes even more unstable.

6) SNS Reactions: Divided Between "This Time is Different" vs. "Copy & Paste"

What's interesting this time is that reactions on social media are clearly divided (examples of some posts are shown below).

• "This Time is Different" Group
  Many posts positively view AI investment as "capital investment in the real economy," citing Powell's point of "having profits" and "having business models." LinkedIn

• "History Repeats Itself" Group
  In a LinkedIn comment, the atmosphere of "valuation doesn't matter" in 2000 is compared to the AI boom in 2025, stating "History isn’t repeating. It’s copy-pasting." The stronger the expression, the more easily it spreads on social media. LinkedIn

• "AI is Real, the Market is Different" Group (Intermediate)
  Like Burry, who says "the timing of the collapse is unpredictable," a tone of caution about supply-driven expansion (the "accumulation" of data centers and GPUs) is spreading. Business Insider

• In Community-Based SNS, "Individual Issues" are Explored in Depth
  In investor communities, discussions tend to focus on "mechanisms of collapse" such as (1) valuation, (2) excessive infrastructure investment, (3) individual speculation, and (4) chain reactions during credit contraction (margin calls, etc.). Motley Fool Community

7) Conclusion: AI Might "Win," But That Doesn't Mean the Market Will "Win"

The lesson from the dot-com era is simple. Even if the technology wins, the price investors paid may not be salvaged.
And the "twist" of the AI market lies precisely there.

• This time, large companies have profits (i.e., a strong foundation). Multpl

• However, as investments grow larger, and supply-driven expansion, credit market fluctuations, and the resurgence of individual speculation overlap, the market shows a different face. Business Insider

• Ultimately, the question is the speed at which "computing resources created by capital investment" are converted into "actual demand payments."

The possibility that AI is a revolution and the possibility that the AI market is a bubble can coexist. This is why discussions on social media heat up. Ultimately, the key to understanding the AI boom in 2025 lies more in "the flow of funds (cash, debt, credit)" than in "the correctness of the technology."