Why AI and Defense Are Soaked in Capital While Food and Water Go Thirsty

In a healthy innovation economy, capital should generally align with society’s most urgent needs and the best risk-adjusted opportunities. However, that is not the case right now. Venture capital has shifted dramatically toward artificial intelligence and defense technology, with deal sizes and valuations increasingly resembling the speculative frenzy of the late 1990s dot-com bubble. At the same time, critical issues that will impact human well-being for the rest of the century, climate resilience, water scarcity, and food security, remain chronically underfunded relative to their significance.

Food and water prices are already flashing warning signals: the FAO Food Price Index rose 6.9% year-over-year in 2025, while in the U.S., the all-food CPI climbed 23.6% between 2020 and 2024. These are not abstract risks; they are eroding living standards today and reshaping global stability.

The AI frenzy: Growth or bubble?

Artificial intelligence has absorbed more venture funding than any other sector in history. In 2024, AI and machine learning startups attracted $131.5 billion globally, representing a 50% increase over the previous year. In the U.S. alone, $109.1 billion in private AI investment dwarfed China’s $9.3 billion, making this not just a market race but a geopolitical one.

Generative AI captured $33.9 billion of that total in 2024, even though industry revenue was just $16 billion in the same year, with optimistic forecasts placing it at $85 billion by 2029. That mismatch, more than $100 billion invested annually into a market worth a fraction of that today, is the classic marker of bubble dynamics.

Meanwhile, the underlying economics are strained. Training large models can cost tens of millions of dollars per run, with flagship Nvidia H100 GPUs priced at around $25,000 each. The infrastructure burden is equally daunting: McKinsey estimates global data-center capex will reach $6.7 trillion by 2030, with $5.2 trillion tied to AI workloads alone.

This is not an argument against AI itself. It is an argument against capital piling into AI at valuations and burn rates that the real economy cannot justify on venture timelines.

Defense: A boom with long clocks

Defense technology has also become a magnet for venture dollars. In 2024, the sector drew $9.1 billion across 228 deals, and in June 2025, Anduril raised $2.5 billion at a $30.5 billion valuation, one of the largest private rounds on record.

The sector’s appeal is clear: rising global tensions, trillion-dollar defense budgets, and governments signaling a willingness to work with startups. Yet the monetization gap is stark. A Wall Street Journal review found that the U.S. government invested only $22 billion in contracts with the top 100 national-security startups, less than half of the $53 billion those companies raised in venture capital.

Defense cycles are notoriously slow. Regulatory approvals, testing, and procurement stretch over years, often decades. For LPs expecting quick exits, this is a mismatch between the patience of the problem and the impatience of the capital.

The opportunity cost

This is not just a moral lapse; it is a strategic blind spot. Water underpins food, and food systems underpin climate stability. Agriculture alone accounts for 70% of global freshwater withdrawals and about one-quarter of greenhouse gas emissions. If water systems falter, food systems collapse, and with them, climate goals and national security alike.

The WHO and UN project that by 2025, half the world’s population could be living in water-stressed regions, while the FAO reports that 295 million people already faced acute food insecurity in 2024, up nearly 14 million from the year before. Droughts alone are draining $307 billion annually from the global economy, and water scarcity is projected to shave up to 8% off global GDP by 2050. These are the kinds of systemic risks that make AI GPU shortages look trivial.

But climate tech already gets billions!

Climate tech did indeed attract $56 billion in the year to Q3 2024, according to PwC. But that money is spread across energy, mobility, industry, buildings, and carbon management. Within that pie, water solutions accounted for just 1-2% of climate-tech capital in 2024, and agri-food investment has been in steep decline since 2021.

If you accept the World Resources Institute’s math, that feeding nearly 10 billion people sustainably will require a 50% increase in production alongside significant emissions cuts, then today’s capital stack looks dangerously misaligned. Food and water systems are not niche. They are undervalued and undercapitalized, despite being the foundation of survival.

A correction is inevitable. The FAO Food Price Index is up nearly 7% year over year, utilities worldwide are imposing record double-digit water tariff hikes, and food prices globally have risen 35% over the past five years. Scarcity is becoming the most expensive commodity of the 21st century.

The deep irony

Here lies the irony. The very technologies absorbing venture capital depend directly on the resources being neglected.

Data centers alone consume vast amounts of water for cooling. On average, a facility uses about 1.8 liters of water per kWh of energy consumed, and a typical 1 MW data center can draw 25.5 million liters annually, enough to supply 300,000 people for a day. In regions already facing water stress, the growth of AI infrastructure is a direct competitor with households and agriculture.

And defense is no different. Armies still march on their stomachs. Troops need food to fight. If agriculture falters under water scarcity and climate stress, even trillion-dollar defense budgets cannot paper over the shortage of calories required to sustain militaries, let alone civilians.

The engines of AI and defense ultimately rest on the stability of food and water systems. Ignoring that foundation is not only shortsighted, it is strategically self-defeating.

The market test

AI and defense are flush with capital, not because their economics are proven but because their narratives are irresistible. A flashy demo or a new defense contract can move billions overnight. Food and water innovations, by contrast, are incremental, technical, and slow-moving, but they deliver tangible, compounding returns. Precision irrigation reduces costs year after year. Advanced seeds increase resilience for decades. Water reuse and leak detection save billions in wasted resources.

These are not hype cycles; they are the real economy.

What rebalancing should look like

If venture capital is serious about long-term returns and resilience, reallocation is essential. That means:

1. Late-stage growth equity in precision irrigation with the same intensity we see in AI coding agents.
2. Billion-dollar debt facilities for municipal leak detection on par with data-center expansion.
3. Deep pools of patient capital for seed genetics, soil health, and water reuse startups, not just for another round of model training runs.

Unlike AI hype or defense cycles, demand for food and water is not optional, cyclical, or speculative. It is as inevitable as biology.

The closing reality check

The point of venture capital is not to finance press releases; it is to finance future cash flows. Right now, those cash flows are imperiled less by a shortage of chatbots or defense drones and more by the slow-motion crisis of droughts, floods, and soil degradation.

If LPs and GPs continue mistaking spectacle for opportunity, the correction will be painful. A crowd of AI and defense unicorns will discover their customers are slower than their burn, while undercapitalized food and water systems face the climate’s full brunt.

The arbitrage is hiding in plain sight: invest where demand is guaranteed, externalities are positive, and scarcity is pricing itself into the global economy. Venture capital exists to deliver returns to LPs, not to chase mirages. The boom-or-bust bets on over-inflated bubbles in AI and defense are fool’s gold; steady, grounded investments in water, food, and climate resilience will win the day.

Paul Streater is the Chief Operating Officer at Accumont, a global food and agriculture innovation consultancy, and an Operating Partner at Natural Ventures, a food and water security investment fund. He leads strategy, investments, acquisitions, and investor networks for clients worldwide. Paul serves as a board observer to several ag-innovation companies and advises on food and agriculture for the Abu Dhabi Sustainability Week. He writes at the intersection of capital, technology, and climate-resilient agriculture.
See all author stories here.