Abstract/ TLDR: There is already important conversation on AI and society: how AI will change work, education, science, inequality, and government. This paper asks a more direct question: What should a society do when it acquires an extraordinary windfall of productive capability?
The AI windfall creates the opportunity for enormous wealth and capability. But there is no guarantee that this benefit will be realized, or that it will translate into productive capacity. This is the resource curse: without the institutions, coordination, and allocation mechanisms to convert a windfall into productive capacity, AI will create far less value than its potential.
AI infrastructure is being built at extraordinary scale, speed, and cost. We make two claims about what comes next. First, the highest-return uses of that infrastructure are at the frontier: critical foundational markets that are deeply underserved and where the potential for new value is greatest. Second, reaching that frontier requires a shift from the traditional venture model, based on disruption, toward a systems approach and interdisciplinary collaboration from the outset. This requires a national approach to setting priorities, developing policy, and aligning capital and entrepreneurs to direct the AI windfall toward the most important opportunities, and unlocking markets that work for the many.
The background
Last year, I wrote a post asking readers to imagine AI as an analog equivalent: a gold mine. My argument demanded a deeper respect of the game when it comes to AI and the massive effort it will take to translate its potential and the massive investment into enduring economic value. What If AI Was a Gold Mine?
Last week, I read that Jase Auby, CIO of the Teacher Retirement System of Texas, compared investment in AI with the buildout of the U.S. interstate highway system under President Eisenhower, as well as earlier infrastructure booms such as canals and railroads.1 This illustration expresses my idea with more elegance: If we build enormous capacity, what will we use it to produce? What are the best use cases? Where will the cash flows come from? And how will these applications strengthen the economy beyond the companies supplying the technology? How do we know?
The dominant narrative is that AI investment is building the infrastructure of the next economy. That may prove correct. Infrastructure is the beginning of the economic argument. We are already having important conversations about AI and labor, power, productivity, education, science, governance, and society. What is less discussed is how to translate this enormous expansion of capability into a lasting economic dividend.
The Philosophy
Countries that are endowed with a valuable natural resource face the risk of a resource curse: assuming that the resource itself is the dividend. Societies that have converted a windfall into lasting prosperity have had to do something harder. They have designed and implemented complex, but integrated strategies, institutions and systems through which the resource could generate broad value over the long time.
America's natural resource is different. It is our ability to consistently produce technological innovation. Historically, technology did not advance this quickly, nor was it at the center of every industry. The other domains of education, regulation, tax policy, labor markets, infrastructure, and institutions had time to catch up.
AI presents a different reality. It is one of the largest technological revolutions we have experienced, happening at extraordinary speed, while absorbing enormous amounts of capital and resources—including energy, water, and land. The surrounding systems cannot wait for the technology story to play out. They must advance simultaneously, otherwise, the gap between technological and social progress will widen further. The scale and scope of return on investment must move beyond the ROI for investors on companies’ capitalization tables towards global return since all of us are investors in AI now.
Building the Highway System
The highway analogy is useful because building infrastructure, earning a return on it, and creating broad economic benefit are three different achievements. Construction creates jobs, demand for materials, and a physical asset. The larger dividend comes when the highway connects people and goods to productive destinations: businesses, employment, services, and markets. Its value depends on what develops around it and what becomes reachable because it exists.
The same is true of AI infrastructure. Failure can happen in several ways:
Value can concentrate: A small group captures most of the economic gains with no broad lift to market participation or return on public investment.
Value can accrue elsewhere: The infrastructure succeeds, but the benefits accrue neither to local markets nor to those who invested.
Product-market fit can fail: Technical capability exists but doesn’t translate to any useful products at sufficient scale.
Returns can remain low: The technology works, but the market opportunity and productivity lift are insufficient to justify the capital deployed, or the markets that would create the highest-value uses never materialize.
Resources can be misallocated: Energy, water, land, talent, and capital are committed to uses that turn out to be less valuable than critical alternatives.
Building the infrastructure is just the beginning.
From infrastructure to Dividend
If AI is as consequential as its advocates claim, an investment of this magnitude demands robust architecture for realizing its value. This goes beyond the race to build more capability and capacity to how those are directed toward productive uses at scale.
At Fairbridge, we believe the highest-value use of this technological windfall is rebuilding foundational systems of health, economics, and a livable planet. These are areas that affect billions of people worldwide, have been historically underinvested and underserved, and place enormous strain on resources. AI can now change that equation.
Entrepreneurial innovation and market mechanisms provide a way to build new products and services that address unmet needs better, boosting agency and utility and efficiency. Applied to foundational systems, technology can create structural pathways for participation in economies and markets that have remained inaccessible or poorly served. This is the technology dividend.
This has been done before: Lessons from the winners
After uncovering a windfall of oil, Norway established a robust strategy for converting the petroleum revenues into national wealth.2
What Norway did | How it worked | The lesson |
Found the means and partners to extract | Assembled the technical expertise, capital, and operating partnerships needed to turn the resource into productive output. | An endowment becomes an economic opportunity only when you build the capability to produce from it. |
Captured a public share of the value | Petroleum taxation and direct state participation generated public revenues. | The distribution of a windfall is a design choice. |
Converted revenues into lasting assets | Established its petroleum fund in 1990 to accumulate financial wealth for the future. | A temporary boom should create capabilities and assets that outlast it. |
Diversified beyond the source of wealth | Invested the fund internationally rather than concentrating it in the domestic petroleum economy. | Do not make the whole system dependent on the activity generating the windfall. |
Disciplined deployment | Introduced a fiscal rule in 2001 linking spending over time to expected real returns, currently estimated at 3%. | Deployment should follow durable capacity, not the excitement of abundance. |
Made allocation accountable | Petroleum revenues enter the fund; transfers through the budget require parliamentary decisions. | Stewardship requires institutions, responsibilities, and oversight. |
The transferable lesson is that producing value and a lasting dividend requires orchestration across different institutions, capabilities, and interests. Botswana is another good example. Diamond revenues, prudent management, and rules-based institutions helped finance development. But continued dependence on diamonds remains a vulnerability especially as diamond prices are falling.3
Other places like Zimbabwe and Congo are cautionary tales. Mineral endowment has not consistently translated into broad-based development. Political conflict, weak institutions and regulation, policy inconsistency, and difficulties capturing mineral rents have constrained that conversion.4 Saudi Arabia is currently struggling through this exercise with projects like Neom City that are stalling. The distinction is not geography, but superior capability to systematically unlock value.
The Opportunity is big: The Good Commercial Sense test
At this year's Clinton Global Initiative, Bill Clinton urged participants to “step into the void created by retreating governments” following cuts to foreign assistance. He also acknowledged that philanthropy and other sectors cannot completely replace the capacity being withdrawn.5 Our argument is not that AI can replace public funding, or that companies can substitute for government. It is more ambitious.
We believe that entrepreneurial innovation is an underutilized asset to tackle the most difficult problems. We believe it can unlock new structural pathways for billions of people to participate in the economy with better outcomes. Our north star is the eventual end of subsidy as the mechanism required to make essential solutions available. We want innovative solutions that people can choose and can afford and are happy to pay for because they deliver superior customer value. The process that makes this possible requires coordination of 5 functions:
Prioritization: Societies have to decide what they want to win in, where they have competitive advantages, and what capabilities they want to build. These choices are revealed through budgets, investment priorities, and institutional attention. The current administration has prioritized defense spending, nuclear energy, and other areas while making significant cuts to healthcare—including Medicaid and the ACA ($1T), food security/SNAP ($187B), and higher education through changes to student loan programs ($60B). Whatever one thinks the top priorities should be, where government does not prioritize, it can still enable other players through policy.
Designing the Playing Field: Policy and regulation can offset the budget cuts by determining what can be built, who can use it, who carries the risk, and what gets rewarded in foundational systems. A healthcare tool must fit into clinical responsibility, delivery, and reimbursement. A financial product must fit within the applicable securities, banking, consumer-finance, or payments regulatory framework. An AI tutor is not automatically an educational pathway. It must connect to learning, credentials, institutions, and work. Better financial information is not access to credit. Better health information is not access to care. Rules and policy are the means to an end, the design of a playing field that aligns incentives, draws talent, boosts production, establishes accountability, and promotes broad participation.
Produce and Distribute: Turning the technology windfall into durable companies requires exceptional talent, aligned capital, and networks. Founders tackling foundational problems often need the capacity building that goes along with it, especially mentorship, access to good teammates, and access to stakeholders who understand the problems on the ground deeply. These networks provide social license, insight, and pathways to distribution. The goal is to produce companies that can wield technology to build useful markets at scale.
Tax and Reinvest: Investment in AI is not just private capital. The buildout also requires energy, water, land, infrastructure, public research, education, and other resources that could have been deployed elsewhere. This means the public is also an investor. When done well, taxation is an excellent mechanism for national portfolio construction, reinvesting some of the economic benefit from technological advancement back into other important areas that had to be deprioritized to fund AI.
Esther Dyson’s recent post makes an interesting proposition, suggesting to foster health through the redistribution of AI wealth.6 This is not a moral position, but a claim that if Esther is right that increasing people’s income is one of the most significant contributors to better health, and if health is a priority for us, then it is not strange to propose redistribution as an economic allocation decision.
The same logic applies elsewhere. Education, scientific research, infrastructure, household income, environmental resilience, or other priorities can all become allocation decisions. The AI dividend is then maximized when it’s managed expertly to build the foundation of resilience.
The commercial takeaway
What we may have in front of us is a huge windfall of capability that continues to demand enormous investment. Like any critical infrastructure investment, there’s high risk of loss of principal. There’s also enormous opportunity to translate this technological infrastructure into a long-lasting dividend that delivers high global return. This is not primarily the work of disruption. It is the work of bringing together world-class experts in the critical institutions of government, finance, law, and civic life to rally around entrepreneurs and build strong institutional foundations and businesses in areas that matter. The coordination required to translate infrastructure investment into a system that compounds value over the long term is simultaneously the biggest challenge and opportunity today.
In Summary
AI has given the hardest access problems renewed attention because of its potential to transform the fundamentals of the economics of serving them. But the opportunity becomes real only when technical capability is translated into compelling products and services that deliver superior value at scale and over time. And realizing that opportunity requires a new way of building—one that demands complex coordination.
Sources
Jessica Hamlin, PitchBook, Texas Teachers’ CIO warns AI capex boom echoes past busts, Sep 21, 2026.
Norway Ministry of Finance, The Norwegian Fiscal Policy Framework.
Mining Weekly, IMF recommends reforms as Botswana faces mining-induced fiscal pressure (IMF 2025 Article IV consultation), Sep 29, 2025.
Bright Mutandwa, Leveraging Zimbabwe’s mineral endowment for economic transformation and human development, University of the Witwatersrand, 2018.
The Chronicle of Philanthropy, Bill Clinton calls on leaders to fill gaps left by ‘retreating governments’, Sep 2026.
Esther Dyson, How to foster health: redistribute AI wealth, Substack.
Join our Mission
Please share this post with builders, investors, catalysts, and policy people in your networks!
We are always seeking to partner with others who are advancing social progress through entrepreneurial innovation. Please reach out to me or at [email protected].
And if you are a founder, you can apply for funding.
Thanks to our sponsor this week. Greenfield Robotics is an example of what AI can make possible in a critical industry...
Can Robotics Make Regenerative Farming Scalable?
Greenfield Robotics is on a mission to give farmers alternatives to herbicide-dependent weed control. BOTONY is the beginning of a broader robotic farming system designed for a wider range of applications in the field.
The Reg A+ offering is now live for investors who want to be part of what comes next.
This Reg A+ offering is made available through StartEngine Primary, LLC, member FINRA/SIPC. Please read the Offering Circular and related disclosures before investing. This investment is speculative, illiquid, and involves a high degree of risk, including the possible loss of your entire investment.