SpaceX Is a Smaller IPO Than a 1987 Japanese Telecom Giant. That's Not the Comfort It Sounds Like.

In February 1987, Japan's state-owned telephone monopoly, Nippon Telegraph and Telephone, went public in an offering worth roughly 7–10% of the entire Japanese stock market. The Nikkei absorbed it without a tremor. NTT shares doubled in the first two months of trading. Three years later, the same stock had fallen more than half from its peak, and the overhang dragged through the rest of the decade.

SpaceX's expected US listing is a different scale: about 2.5% of current US stock-market capitalization, according to UC Berkeley economist Barry Eichengreen. Even adding the forthcoming OpenAI and Anthropic IPOs leaves the combined cohort smaller, in proportional terms, than what Tokyo digested in 1987. On raw size, the "markets can handle it" reading is straightforward.

The mechanism behind NTT's absorption is less reassuring. Japanese banks and insurance companies were effectively required to load up on the shares. The offering carried limited free float, regulators and peers expected institutional participation, and the weight of NTT in the Nikkei made underweighting it tantamount to a public bet against Japan. In Eichengreen's telling, those institutions did not buy NTT because they had independently concluded it was the right investment. They bought it because their mandate, their index, and their career risk left them no defensible alternative.

The structural fit to today's mega-IPO cohort is partial. Index funds still must buy a stock when it enters their benchmark. The same mechanism that pushed NTT onto Japanese bank balance sheets now pushes SpaceX (or its successor benchmark listings) onto passive vehicles worldwide. But the issuer is different. NTT was a regulated monopoly with pricing power and a captive customer base; it was treated as a one-way bet on Japan's transition from manufacturing to information. The forthcoming AI and space listings are competitive private companies selling into markets with at least two or three credible alternatives at every layer. The forced-buyer problem applies; the monopoly-rent cushion does not.

The macro backdrop is worth naming. The Bank of Japan had cut rates from roughly 5% in 1984 to about 2.5% by 1987, and the Nikkei rose about 14% in 1985 and 43% in 1986, per Eichengreen's account. That kind of monetary tailwind does not announce itself in advance, and it can reverse without warning. That is one of the three lessons Eichengreen draws from the comparison.

The second lesson is the one most often skipped. Institutional momentum is bidirectional. The same banks and insurers that absorbed the initial NTT tranches eventually became the sellers, and the float returned to the market over years rather than quarters. NTT peaked in April 1987, fell about 25% by April 1988, shed a further 37% by April 1989, and traded through the 1990s at one-third to one-half of its bubble-era valuation, according to Eichengreen. A listing that "works" on day one can still be a multi-year overhang, and the buyers whose participation made it work are the ones who eventually supply the supply.

The third lesson is the one Eichengreen puts last and the public discussion tends to skip: today's technology dominance is not permanent. Railway stocks, electrification champions, and telecoms all looked like one-way bets at their respective peaks. Each cohort's dominance faded on a horizon longer than the typical investor's holding period.

The proportional-size comparison is arresting. As reassurance, it is incomplete without four structural questions answered: who is required to buy, what is the free float, what is the issuer's competitive position, and what happens when those forced buyers eventually become voluntary sellers. When a commentator says markets can absorb a mega-IPO because Tokyo absorbed NTT, those are the questions that turn the analogy from reassurance into a usable model.

Eichengreen, a long-tenured Berkeley economic historian, makes this argument in a July 2026 Project Syndicate column syndicated via interest.co.nz. His faculty page lists the longer body of work behind the analogy. The argument is an opinion essay, not a forecast. The structural reading it forces is what the proportional-size conversation currently lacks.