An older acquaintance apparently has around ¥10 million saved while still in their mid-30s. The immediate reaction: that is genuinely impressive.
What stood out even more was how casually they paid for lunch and dinner, about ¥7,000 total. The explanation was basically: “I do not spend much normally, so it is fine.” That sentence may explain the ¥10 million better than the balance itself.
1. The statistics trigger one very unserious thought: where did everyone’s money go?
J-FLEC’s 2025 survey reports the following for single-person households, including households with no financial assets:
| Age group | Average | Median |
|---|---|---|
| 20s | ¥2.55M | ¥0.37M |
| 30s | ¥5.01M | ¥1.00M |
| 40s | ¥8.59M | ¥1.00M |
| 50s | ¥9.99M | ¥1.20M |
| 60s | ¥13.64M | ¥3.00M |
| 70s | ¥14.89M | ¥5.00M |
For singles in their 30s, the average is ¥5.01M and the median ¥1M. The statistically irresponsible first thought is: “Seriously, what is everyone spending their money on?”
If ¥10M is comparable under the same definition, it is about twice the single-household average and ten times the median. But household type and the exact definition of the acquaintance’s “savings” are unknown. For households of two or more people in their 30s, the average is ¥10.96M and the median ¥3.11M. So ¥10M is substantial, but an exact “top X%” claim would be invented precision.
2. Two corrections from the conversation
First, ordinary deposits are not automatically excluded. J-FLEC distinguishes mainly by purpose: money accumulated for investment or future use is included, while money reserved for everyday payments, direct debits and routine withdrawals is excluded. The issue is the job assigned to the money, not merely the account label.
Second, the earlier table at ages 20, 25, 30, 35 and so on was interpolation. J-FLEC’s official age data are basically decade groups. Those 5-year points were not official statistics. For a 35-year-old comparison, use the official 30s group.
Detailed numbers receive a mysterious authority buff. Sometimes the detail came from the source; sometimes we added it ourselves.
3. “I normally do not spend much” is the coolest part
The useful lesson is not “never spend anything.” It is cut spending you do not care about, then spend comfortably where you actually care.
Routine leakage stays low, yet about ¥7,000 for shared meals is no crisis. That is not someone who never spends; it is someone who chooses where to spend.
There was also no “I paid for you” pressure. Some free meals arrive with an invisible psychological invoice. This one did not.
About ¥7,000 paid. Psychological transaction fee: ¥0. Excellent payment network.
4. Then the freed-up ¥7,000 goes to AI
Here the story becomes questionable.
Because lunch and dinner were covered, roughly ¥7,000 survived in the budget. The sensible response would be “Great, save it.” Instead: “Maybe I will spend the amount I got treated to on AI.”
The acquaintance’s everyday restraint creates spare capacity; that pays for food; the meal protects ¥7,000 of someone else’s cash; that money may flow into an AI service and eventually keep a GPU busy somewhere.
Acquaintance → restaurant → surviving ¥7,000 → AI → GPU
They thought they were buying food. They may have become an indirect sponsor of compute. Capitalism has taken the scenic route.
5. Is AI spending an investment? Status: unconfirmed
An AI subscription does not magically become self-investment when the card is charged. If barely used, it is simply another fixed expense.
It starts looking like investment when it leaves results: time saved, better research or decisions, faster learning, more or better output, new revenue or opportunities, or hypotheses tested that would otherwise be too slow or expensive.
If AI gets asked “What should I eat?” 500 times and nothing else happens, that is entertainment. Entertainment is allowed. But renaming every subscription “self-investment” to remove budget guilt is dangerous accounting.
The most accurate conclusion remains: “It might help the future. Still unconfirmed.”
6. Conclusion
The lesson is not “spending less always makes you better.”
Do not let money leak everywhere unconsciously. Then spend naturally where you genuinely care.
About ¥10M in the mid-30s is strong. Naturally covering roughly ¥7,000 of meals is also strong. The person receiving the meal then trying to reincarnate the ¥7,000 as an AI subscription is the suspicious final step.
Whether AI becomes an asset or merely another subscription will be judged by what remains afterward, not by how much it was used.
7. Data notes and sources
- J-FLEC, Public Opinion Survey on Household Financial Behavior (2025): https://www.j-flec.go.jp/data/kakekin_2025/
- 2025 classified data for single-person households: https://www.j-flec.go.jp/wpimages/uploads/per12501.xlsx
- 2025 simple tabulation for single-person households: https://www.j-flec.go.jp/wpimages/uploads/shukeit25.pdf
- The age table includes households with no financial assets and uses official decade groups, not an official value for age 35.
- No percentile is estimated because household type and the exact scope of “savings” are unknown.
