Global fertility may be below replacement. Robotics VCs are spending accordingly

Why it matters

VC portfolios have long treated population and labor supply as background variables. If global fertility crossed below replacement in 2026, automation demand becomes more durable while consumer-market growth becomes less automatic. The winning investments will solve labor constraints already visible in customer budgets, rather than waiting for a 2056 population peak to rescue weak economics.

Video version

Jesus Fernandez-Villaverde (@JesusFerna7026) and Patrick Norrick argue that global fertility likely fell below replacement level in 2026, a threshold that would bring the economic effects of population aging and eventual decline forward by decades.

Their new working paper, examined in an August 19th Forbes analysis, lands during a record rush of venture capital into robotics. Investors have already decided that fewer available workers will create durable demand for machines. The paper gives that thesis a stronger demographic foundation while exposing its weakest assumption: a long-term labor shortage does not guarantee that today's robotics startups will reach commercial scale within the life of a venture fund.

Fernandez-Villaverde, the Howard Marks Presidential Professor of Economics at the University of Pennsylvania, studies macroeconomics, econometrics and quantitative economic models. Norrick is a research associate working on demographics and political economy. Their intervention matters because it challenges the population baseline used in corporate plans, government budgets and investment models, rather than merely revising fertility estimates for a handful of aging rich countries.

The paper moves the demographic clock

Fernandez-Villaverde and Norrick fit a single-factor model to fertility data from 236 countries and territories dating to 1950. According to their paper, a common global fertility component peaked in 1978. Country-level trends now drive the continuing decline: fertility is falling in 219 of the 236 places studied, at an average rate of 0.062 births per woman each year, and the authors find no country trend that has leveled off.

That model produces a sharper result than the conventional demographic transition story, which tied low fertility largely to wealth, education and urbanization. The authors find recent declines concentrated in low- and middle-income countries and among poorer, less educated women. Mexico's fertility rate reached 1.51 in 2025, they report, despite per-capita output near 28% of the US level. Tunisia reached 1.45, while Thailand fell to 0.87.

The figures are estimates derived from incomplete and uneven national data, especially in countries without comprehensive birth registration. The claim that humanity crossed below replacement in 2026 therefore remains a model result, rather than a settled population count.

It also departs sharply from the United Nations' 2024 projection, which placed the global population peak in the mid-2080s at roughly 10.3 billion people. Fernandez-Villaverde and Norrick forecast a peak around 2056. That roughly 30-year difference would materially change assumptions about labor supply, consumption, tax bases and the addressable markets available to companies founded today.

The authors attribute the pattern to "modernity itself." Their conjecture is that modern life raises the cost of having a third child while reducing the economic and social cost of having none. Small shifts in family composition then produce large changes in total fertility because children arrive in whole numbers. In one example, increasing childlessness and moving a modest share of families from two children to one drops cohort fertility from 1.8 to 1.3 even though most women's behavior remains unchanged.

That explanation remains a conjecture, and fertility projections have a record of revision. The paper's strongest contribution is the cross-country pattern it identifies: poorer countries appear to be moving through the fertility transition faster, and to lower endpoints, than many population models expected.

Venture capital has already bought the automation hedge

Robotics investors did not wait for demographic consensus. Crunchbase counted $18.8 billion in global robotics startup funding through June 22nd, exceeding the $15 billion raised during all of 2025 and the $14.1 billion recorded in 2021.

The total includes several financings whose scale would have looked implausible for hardware startups a few years ago. Austin-based Saronic raised a $1.75 billion Series D on March 31st at a $9.25 billion valuation. Kleiner Perkins led the round, joined by Advent International, Bessemer Venture Partners, DFJ Growth, BAM Elevate, 8VC, Caffeinated Capital, Andreessen Horowitz, Elad Gil and Franklin Templeton.

Saronic co-founder and CEO Dino Mavrookas spent 11 years as a Navy SEAL, including eight combat tours, before investing in technology at Vista Equity Partners. Co-founder and CTO Vibhav Altekar was an early Anduril engineer who worked on autonomous defense programs. Their backgrounds show why Saronic can raise against a near-term procurement case alongside the longer demographic argument. Navies need autonomous vessels under current budgets and current geopolitical pressure. Saronic does not have to wait for the global workforce of 2056.

Anduril supplied an even larger marker in May. The defense technology manufacturer raised $5 billion at a $61 billion valuation in a round led by Thrive Capital and Andreessen Horowitz. CEO Brian Schimpf said Anduril generated $2.2 billion in 2025 revenue, double the prior year's figure. RuntimeWire subsequently reported that Anduril secured an Air Force production slot for its autonomous fighter jet, moving the manufacturer deeper into a funded Pentagon program.

Those financings show where the robotics money is actually going. Defense procurement, AI advances and government industrial policy provide immediate reasons to fund Anduril and Saronic. Fertility decline adds a structural tailwind, especially for autonomous manufacturing and operations, though it does not explain the current contracts.

Bessemer's 2026 robotics thesis makes the distinction visible. Bessemer cites demographic trends in the US, Europe, Japan and China as a lasting source of demand, then predicts that defense robotics will produce the category's first $50 billion-plus public companies. Its argument rests on buyers with large budgets, urgent capability gaps and high switching costs. Those are commercial conditions, not demographic statistics.

The distinction should shape diligence. A shrinking workforce can improve a robot's theoretical value without solving installation costs, reliability, integration, maintenance, safety approvals or customer payback periods. A 2056 demand curve can excuse weak 2027 sales for only so long.

Demographic time and venture time do not match

Fertility falling below replacement does not immediately remove workers from factories, hospitals or warehouses. Children born in 2026 would generally enter full-time employment in the 2040s. The near-term labor constraint comes from earlier fertility declines, retirement, immigration policy and the age structure already in place.

The Congressional Budget Office's 2026 demographic outlook illustrates the lag in the US. CBO expects deaths to exceed births beginning in 2030 and projects that net immigration will account for all US population growth after that point. Without immigration, the US population would begin shrinking in 2030. CBO separately expects labor-force growth to average 0.4% annually from 2026 through 2029, down from 1.6% during the four years following the start of the pandemic.

That is enough to support automation demand in labor-intensive industries. It is less useful as a standalone explanation for which startup wins. Venture funds typically need liquidity on a much shorter schedule than demographic models. Founders selling labor replacement must show a shortage inside today's customer budget, with a deployment that pays for itself under current wages and operating conditions.

The strongest automation startups will sell into existing bottlenecks: unfilled shifts, dangerous work, costly turnover, constrained production capacity or tasks where human performance is difficult to standardize. Population projections can establish durability after the first market works. They cannot manufacture the first market.

Geography also matters. The same fertility rate can produce different startup opportunities depending on immigration, retirement age, labor participation and capital costs. A country that offsets low fertility with immigration may experience slower wage pressure than a country with similar births and little migration. A manufacturer in South Korea faces a different adoption curve from a hospital network in the US, even if both describe aging as the reason to automate.

Pictet Asset Management estimates robot density in Japan, South Korea and China at about 40 robots per 1,000 workers, nearly four times the global average. Its research associates higher robot density with roughly 0.4% in additional annual GDP growth. The direction supports the automation thesis, although the estimate does not establish that any specific robotics vendor will capture that productivity gain.

Aging creates demand that venture still struggles to finance

The capital response has been far less aggressive in products built directly for older adults. AgeTech covers remote monitoring, fall detection, care coordination, home assistance, mobility, financial protection and tools for unpaid family caregivers. Demand should rise as populations age, yet the category has fragmented buyers and difficult distribution.

An AARP AgeTech Collaborative report says more than $3 billion has flowed into AgeTech startups since 2015, with roughly half directed to technologies and services designed specifically for older adults. The report also says adoption and sustained engagement have failed to keep pace with investment.

The comparison with robotics uses different category definitions and time periods, but the capital preference is difficult to miss. Robotics attracted $18.8 billion during part of 2026. AgeTech accumulated more than $3 billion over a decade in AARP's accounting.

The buyer explains much of the gap. Defense contractors sell to governments with explicit procurement budgets. Factory automation vendors can calculate output, wages and downtime. AgeTech founders often navigate a chain involving an older user, an adult child, a care provider, an insurer and a health system. The person using a product may control none of the purchasing budget. Savings may accrue to another institution years later.

Aging therefore creates a large need without automatically producing venture-scale revenue. The AgeTech companies most likely to break through will attach themselves to a payer with measurable economics: fewer hospital admissions, lower staffing costs, delayed entry into institutional care or reduced claims. Consumer willingness to pay remains useful. Reimbursement and enterprise distribution decide whether the market scales.

Shrinking markets will change the other side of the portfolio

Automation is only half of the venture implication. Persistent low fertility also weakens the default assumption that every market expands with population.

Founders selling childcare, education, entry-level housing or products for young families will face smaller cohorts in many countries. Customer acquisition may become harder as fewer buyers enter the category each year. Consolidation becomes more likely when fixed infrastructure was built for a larger population.

Companies serving older adults, healthcare providers, asset managers and employers may gain demand. Even there, spending power matters. Aging raises medical and care needs while increasing pressure on public budgets and family finances. A growing user count does not ensure a growing pool of profitable customers.

The International Monetary Fund's work on the silver economy estimates that healthier aging could add about 0.4 percentage point to annual global growth between 2025 and 2050 by extending working lives and improving older workers' productivity. Under current policies, the IMF still expects aging and other forces to pull global output growth lower over time. Automation and AI could offset part of that drag, although the IMF's central analysis does not attempt to model the full technological response.

That uncertainty is the investable opening. A founder who enables one nurse, engineer, shipyard worker or technician to produce more can sell into a constraint visible today. A founder whose pitch depends on the exact year of the global population peak is selling a forecast.

Fernandez-Villaverde and Norrick's paper gives investors a reason to revisit the population assumptions buried inside market models. It strengthens the case for productivity software, industrial automation, autonomous systems and care technology. It also demands tighter underwriting. Demography supplies a durable source of demand. Founders still need a customer, a budget and an outcome that arrives before the fund expires.