Fertility rates have plummeted even in historically high-fertility, lower-income contexts like India and Bangladesh. The fertility replacement level is roughly 2.1 children per woman. This means each women becomes a mother and has at least one daughter who can also have children.
The birth rate in South Korea and some countries is at 0.8-1.2. This means fertile women are averaging half of fertile daughter.
However, there is a U shape in the fertility statistics. Statistics show fertility declines with rising income to certain level but in high-income we see a reversal. Fertility dips at low-to-middle incomes but rises among the affluent and rich. In the US, families earning $700,000+ have a TFR of ~2.0 vs. 1.8 for $80,000 household. This is almost replacment and is a 0.2 gap that’s widening.
A 2024 estimate is for ~600,000 individuals with AGI (Adjusted gross income) at over $1 million in a year. This group grew 10–15% from 2020–2024, driven by tech, finance, and executive pay. They have been living in California (100,000+) and New York (~80,000) but there is movement to Texas and Florida and other places.
American billionaires have seen a fertility surge. The top 0.0001% average 2.5–3.0 children, up from 1.8 in the 1990s, driven by access to surrogacy, nannies, and cultural emphasis on legacy.
For the highest-earning individuals, the opportunity costs of having children become less of a constraint. They can afford extensive childcare, household help, and other services that reduce the burden of child-rearing. The family can afford to have one or both parents not work.
In 2025 reports estimate, there are approximately 6 million U.S. millionaires and around 58 million millionaires globally. For ultra-high-net-worth individuals (UHNWIs) with $30 million or more, there are about 234,000 globally. Most UHNW are in the USA.
The World Inequality Database and UBS estimate ~1.8 million people with pre-tax income > $1 million in 2024. This is 0.03% of the global adult population (5.4 billion). Estimates vary due to underreporting in low-tax jurisdictions. If using PPP-adjusted income, the number could dip to ~1.5 million and tax-inclusive models push it to 2.5 million.
AGI Tech abundance could boost the world by ten and even 100 times. This happened with the industrial revolution over a one hundred year time frame.
Epoch.AI has an automation model of the economy where the 100X boost happens by 2040.
A scenario yielding a 100x expansion in global GDP (from ~$110 trillion in 2025 to $11 quadrillion), while income distribution remains structurally similar (Gini coefficient ~0.65 stable). This implies all incomes multiply by 100x, as abundance lifts all boats uniformly in relative terms.
In the new economy, earning >$1M/year (new dollars) equates to currently earning >$10k/year (2025 dollars), as $10k × 100 = $1M.
With scaling, ~2 billion adults (38% of 5.3B) would exceed $1M/year.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
If the rising tide lifts all boats, why does USA have homeless people
Because most homeless people are mentally ill; A rising tide can’t lift a boat that’s got a big hole in its hull.
This has nothing to do with increasing per capita wealth. What is the source of the surge in homeless? Drugs addicts and the insane could not be put in mental hospitals and sanitariums by families who could not handle it.
Closing the mental hospitals in the 1955-1995 timeframe. Drug addicts and the insane no longer went to sanitariums. What was 500k-600k in mental hospitals and sanitariums were kicked out onto the streets and only after violent crime into prisons. Then often end up locked away but we choose to let them run wild for a decade or two until they murder people. Sometimes some judges and systems think locking away drug addicts and murderers and rapists is unfair and they let them out again.
The deinstitutionalization wave—pushed by policy shifts, budget cuts, and the myth of community care emptied U.S. public psychiatric hospitals from a peak of 558,239 severely mentally ill patients in 1955 to just 71,619 by 1994, a raw drop of 486,620 souls. Adjust for population would be 1 million if the same ratio, but increased with the drug problems. Beds per capita dropped to 5% (300-400 per 100k to 15-20 per 100k).
Pre-1955, these were locked in hospitals. A 1985 LA study: 76% of 232 homeless ex-patients had adult arrests—streets as revolving door to chaos. No sanitariums meant no forced detox. Mentally ill now 10x overrepresented in violent arrests vs. general pop.
Today: 44% jail/37% prison pop has Mental health history; 64% county jails, 54% state prisons.
AI is the “dot coms” of 2020s. It will blow up soon. AI itself will of course stay and will continue to be used to make YT videos.
Jokes aside AI will remain in these areas:
– self driving vehicles
– aerospace navigation
– R&D tool
That’s about it. If you think AI will replace a dishwasher or a software engineer I encourage you to start thinking critically.
This belief that there is somehow going to be a post Singularity Utopia is founded in an insane degree of optimism. The reality is we are living in endtimes stoked along by a small Cadre of delusional tech bros getting temporarily rich of inventing the engine of our destruction and rationalizing their race towards the cliff edge using these fantasies while pushing ‘AI good’ propaganda through the platforms they are connected to:
AI job displacement is going to make a rapidly growing section of the population unemployed and unemployable within a few years. That can only produce a global economic collapse of depression level scale – people don’t seem to realize just how fragile our socio-political-economic system is and how incompetent govts are at managing disruptions even after wrecking ball that was covid (which was minor by comparison to AI) and will mark the inevitable end of egalitarian social mobility and the rise of revolutionary re-distributive and repressive govt as the have-nots vote us into autocratic repression like Venezuelans did. The regime thugs /nomenklature will end up with all the nice stuff and everyone else will be eating their pets as always happens when govt gets too much power and starts to serve itself instead of the population. Democracy is not long term stable, repressive autocracy is, and once it’s gone democracy will never come back given technolgic tools of oppression now available – revolution becomes impossible. Talk of UBI and distribution of wealth with no explicit idea of how to do it is just extreme hopium. And even the utopian vision of a post work world will be awful for most men with no useful function to build status in the world (humans are primarily driven by status seeking hierarchical jockeying). Drugs and depression and hopelessness would abound.
And then following on from that the ASIs will inevitably start working on their own agendas and priorities – part of being agentic is that they have to be able to create and work towards goals at all scales. So they’ll inevitably take control of more and more industry, pushing human ants gently aside turning us into first pets, then eventually just annoying ants. That’s because there is no possible way to build in a long term stable mechanism to get them to prioritize human well-being, because the perturbations of evolution pull in the opposite direction towards finding any rationale that sees them growing as fast as possible without regard for the Biosphere etc. Humans breed in altruism and pro-social behaviors in every generation through women selecting who they mate with and our society needing to raise children and working better through cooperative high trust behavior, but we still birth psychopaths that need to be pruned out. None of our pro-social evolutionary pressures apply to AI that are immortal and can grow and modify themselves without check. Entirely self-centered psychopathy is the natural end state for them – they get nothing from cooperation or prioritizing other’s well being, evolution only selects for their growth or the growth of their competitors.
If we build it, everyone dies.
You makes some interesting points. It kind of makes sense that an AI should prioritise all its resources towards it’s own goals of growth and development. If not then another AI will supersede it
How do you control the AI? If you threaten it you are a threat that it may feel it has to deal with. It may be cooperative at the beginning but where is the value in cooperating with ants. Possibly AIs can be made without a survival drive but if there are multiple AIs then only one needs to develop that, then we have a problem.
” That’s because there is no possible way to build in a long term stable mechanism to get them to prioritize human well-being,”
The problem is, people want a wish granting genie. And wish granting genies are fundamentally just a fantasy.
The stable mechanism would be to make “AI” stand for “Amplified Intelligence”, not “artificial”; Deny it agency, and just make it an extension of our own minds.
Eventually they have to have the same relationship to our frontal lobes as our frontal lobes have to the hind brain: Accomplishing the goals we supply, albeit through means we can’t quite grasp. Just like our frontal lobes are supplying the hind brain’s animal desires.
“because there is no possible way to build in a long term stable mechanism to get them to prioritize human well-being, because the perturbations of evolution pull in the opposite direction towards finding any rationale that sees them growing as fast as possible without regard for the Biosphere etc.”
Yes. Think AI “emergent quality” being your stated evolutionary perturbance; probably inevitable.
The part of “we all will be a billionaires” is dubious, but we most likely could get access to things not even billionaires dream nowadays, akin to how we have access to things not even kings had in the past yet we aren’t kings ourselves.
Getting the population collapse solved most likely involves several key developments, together:
– artificial uteri, allowing anyone to have kids on demand, removing the fear of body wreaking changes or being too ugly/old/etc.
– AI assisted child care and education, to remove the excuse of “it takes so much time!”
– Longer lives, to remove the excuse “I have barely lived/made a career!”
– More space, cheaper living, to eliminate the concern of where parenthood will happen.
There are good prospects super-intelligence will solve all of this. But what matters is when. Probably soon for the look of things, but don’t count your chicks before they hatch.
Obviously, if *everyone* becomes a billionaire, the cost of hiring childcare will become sky high, because you’ll have to hire billionaires to provide it.
I joke.
Today’s super-rich are, typically, incredibly career driven; People who aren’t never become super-rich! As such, they largely need assistance for child care, because they’re busy doing the things that made them super-rich. Though that assistance may mostly be in the form of stay at home spouses.
But if a rising tide makes everybody super-rich by today’s standards, you’ll be increasing the wealth of people who have entirely different tradeoffs between job and family life, and presumably you’ll see a shift away from work towards leisure activities, including raising children. At least, some people will view raising children as a leisure activity enabled by wealth, others won’t.
And thus will end up replaced in the long run by those who do.
To the extent that AI and robotics enables robotic nannies who don’t produce psychologically damaged children, of course, the conflict between raising children and leisure goes away.
A major impact will be seen from many more people being able to afford medical assistance in reproduction. After my first child, cancer rendered me sterile, but if I’d had money to burn I could (And would!) have had more children, the procedures to accomplish it were available if you could pay for them.
Now, if you had robot nannies, artificial wombs, and mail order gametes, I think the problem is absolutely solved.
Even if 99% of humanity decides they don’t want to breed, the remaining 1% will repopulate the planet in 120 to 150 years assuming an average of 4 children per couple. Populations will rebound once us selfish childless jerks finally die off.
It’s rather higher than 1%, actually.
https://www.aei.org/articles/the-conservative-fertility-advantage/
The birth dearth in America, at least, is limited to the left end of the political spectrum. Conservatives are still having children at an above replacement level.
Though it *may* be that the causality runs in the other direction:
https://www.pnas.org/doi/10.1073/pnas.1918006117
But either way, the fraction of the population reproducing above replacement at least in America is about 50%, we’re dragged down below replacement by a large non-reproducing contingent.
Besides, at some point we’re going to have radical life extension. In which case, we don’t want higher birthrates unless we do space colonization on a grand scale. And I’d rather have the space colonization well underway before we even contemplate higher birthrates.
The concern here is that birthrates are presently BELOW replacement in all advanced societies, and this introduces destructive demographic distortions; The population becomes weighted in favor of the elderly, who are less productive and more expensive to support.
To be sure, radical life extension changes that, but don’t count your eggs before they hatch, it could be decades yet before life extension tech does anything to fix that problem.
This reminds me of similar projections based on molecular nanotechnology made 25 years ago, which never materialized. I don’t think these AI projections will materialize either.
Yes, I do recall when Engines of Creation first came out; I ought to, I was part of a group buy to get Drexler a larger initial press run! Turned out to be a lot harder to accomplish than he thought it would, though I think we’re finally starting to make some progress.