In my previous
column, I argued that artificial intelligence is quietly restoring geography as
a competitive advantage. For much of the Internet era, distance became
increasingly irrelevant. Computing migrated into distant cloud data centers,
information traveled around the globe in milliseconds, and economic activity
became less dependent on physical proximity. The emergence of AI, and
especially the coming generation of physical AI systems, appears to be
reversing part of that equation.
Intelligence is
moving closer to where decisions are made, making data centers, fiber networks,
electrical grids, and low-latency infrastructure strategically important once
again. I have gotten feedback to that article from readers that questioned my
thesis. Some have made a reasonable argument that truly low-latency use cases
will build compute power directly into edge devices because even a local data
center may be too slow a feedback loop for real time sensing and actuation.
These readers are
not wrong – but I maintain that local infrastructure is likely to provide
economic advantage that we may not fully understand, a trend we have seen
repeat itself with past technology infrastructure, from electrical grids to
wireless networks. And I believe there will be many use cases that are not able
to handle the added cost and energy requirements of local compute power,
necessitating local hyperscaler capacity. Regardless of whether the thesis is
right or wrong, there is a different question that I want to dig into this
week.
What happens after
data center infrastructure arrives?
Over the past week,
I’ve found myself returning to a sentence from the conclusion of that article: Infrastructure
creates opportunity; people create value. I wrote this as a tidy way to end
the piece, and as a reminder that the future economy won’t be shaped by AI.
More accurately, it will be shaped by people who use AI. We must not forget
that people remain at the center of every technological revolution.
In the current
debates, communities across the country are asking whether they should recruit
hyperscale data centers, expand electrical capacity, or invest in digital
infrastructure that positions them for the next generation of economic growth.
Those are important questions, and I continue to believe that communities
willing to invest thoughtfully in infrastructure generally outperform
communities that define themselves by scarcity. But they miss perhaps the most
important question to ask.
The more
consequential question is whether your community will know what to do once the
new technology arrives. That distinction has followed every major technological
transition in human history.
One of the
recurring themes I provide in my conference presentations is that while we
often divide history into political eras or national boundaries, economies are
more often organized around technologies. Bronze reshaped civilizations because
it created stronger tools and more capable armies than stone. Iron displaced
bronze. Steam reorganized manufacturing. Electricity transformed industry and
cities. The transistor reorganized information. The Internet reorganized
communication.
Every one of those
technological transitions altered where wealth was created and how societies
organized themselves around it. Artificial intelligence will be no different.
The communities that prosper in the Data Economy will not simply be the ones
with the fastest networks or the largest data centers. They will be the ones
that learn how to organize themselves around this new technological foundation.
That has led me to
think about another kind of infrastructure. If the first wave of AI investment
has been physical infrastructure, I believe the second wave must focus on what
I would call institutional infrastructure.
When we hear the
word infrastructure, our minds naturally go to roads, bridges, electrical
grids, substations, water systems, and fiber optic cable. Those investments
create capacity. They make certain kinds of economic activity possible.
Institutional
infrastructure does something different. It activates that capacity. It is the
collection of organizations, relationships, and experiences that help people
recognize opportunity, acquire new skills, meet collaborators, commercialize
research, start companies, find customers, and ultimately create economic value
from the physical infrastructure surrounding them.
There are numerous
recognizable examples of institutional infrastructure:
●
Accelerator programs
●
A university commercialization
office
●
Coworking communities
●
Entrepreneur meetups
●
A mentor network
●
Angel investment groups
●
An AI literacy initiative
None of these
organizations or activities individually transforms a regional economy. Yet
together they create the connective tissue that allows ideas to move between
universities, entrepreneurs, investors, manufacturers, and customers. They are
every bit as foundational to a modern innovation economy as roads were to the
industrial economy or broadband was to the Information Age.
I’ve come to
appreciate successful outcomes from strong systems of institutional
infrastructure through my own work over the years. If you want a nearby
example, consider Wilson, North Carolina. Long before gigabit internet became
commonplace, Wilson made the strategic decision to build a municipal fiber
optic network. But the fiber itself was never the strategy.
Fiber was the
foundation that created economic capacity for Wilson and the surrounding
region.
Over the years, the
community layered institutions onto that foundation to activate that capacity.
They launched the Exchange, a municipal operated coworking space. They invested
in an accelerator program [full disclosure – I lead the team that operates that
accelerator]. Wilson began hosting an annual innovation conference, which will
hold its 10th anniversary event on October 22.
Wilson even created
a brand for the community to better understand the economic potential their
fiber network could unlock. Gig EAST, representing Entrepreneurship, Arts,
Science, and Technology, tells the story of Wilson’s future economic prosperity
and quality of life resultant from technological and institutional
infrastructure investments.
In short, Wilson
has activated human capacity around the core technological shifts leading to
the Data Economy. And growth in the city is booming.
But it has not
happened overnight. Economic development headlines often reward immediacy.
Large corporate recruitment is attractive because it tells a story of
near-instant impact. It produces renderings, ribbon cuttings, construction
announcements, and projections of thousands of future jobs. Those announcements
fit neatly within election cycles and budget cycles. They provide visible
evidence that something has happened.
Unfortunately those
announcements rarely actually deliver the promised outcome, so smart
communities take a more diversified approach. Go whale hunting for corporate
relocations if that’s a political must. But hedge the bets on your economic
future with an entrepreneurial strategy.
Remember, building
an entrepreneurial economy follows a different timeline. It begins with a
researcher deciding to commercialize an invention. A founder working out of a
coworking space. A first customer. A second employee. A local mentor
introducing an entrepreneur to an investor. A startup surviving long enough to
hire ten people instead of two.
For years,
investments in institutional infrastructure can appear almost insignificant.
Then, almost imperceptibly, they begin to compound. The remarkable irony is
that this slower path is, by a wide margin, the less expensive one.
Communities
routinely commit incentive packages worth hundreds of millions of dollars to
recruit a single employer. Yet for a fraction of that amount, sustained over
many years, a community can build an ecosystem that repeatedly creates new
employers rather than hoping to attract one. To give you an idea, investing in
a high quality startup accelerator is about the same cost as installing a
single traffic light. Or securing a new garbage truck. Or paving a few blocks
of sidewalk.
For perspective,
Raleigh’s first coworking space, HQ Raleigh, began with a small investment from
four forward-looking residents. In 2013, a few startup founders came together
at that coworking space to launch Pendo. Nine years later, Pendo built a downtown
high rise to house its ever growing team.
The challenge of
institutional infrastructure is not economics. It is that it requires patience
while not having a defining moment. There wasn’t a singular milestone in the
early Pendo story that would have boosted a political candidate’s campaign. The
golden shovel moment came years later. But it came because of patient
institutional infrastructure that has built across the Triangle.
Infrastructure has
never been sexy. It requires steady investment and maintenance. No one expects
a bridge to be built once and then ignored for the next fifty years. Roads
require resurfacing. Electrical grids require modernization. Broadband networks
require upgrades as technology evolves. We understand instinctively that
infrastructure is not a project. It is an ongoing commitment. Institutional
infrastructure deserves to be viewed through the same lens.
Entrepreneurial
ecosystems require maintenance. AI literacy must evolve as the technology
evolves. Accelerator programs should change as markets change. Networks must be
continuously renewed as experienced founders become mentors and the next
generation takes their place.
If we accept that
maintaining physical infrastructure is a legitimate public responsibility, we
should not be surprised that maintaining institutional infrastructure requires
the same long-term commitment. That does not mean taxpayers should carry the burden
alone. In my last article, I argued that if hyperscale AI companies require
extraordinary new investments in electricity, water, and grid capacity, they
should help finance the expansion of those resources rather than shifting the
costs onto local residents.
I would extend that
principle one step further. If artificial intelligence is becoming the
foundational platform upon which future businesses will be built, then the
companies building that platform also have a vested interest in ensuring
communities know how to use it.This is not philanthropy. It is market
development.
Every entrepreneur
who builds an AI-native company becomes a customer. Every manufacturer that
adopts AI becomes a customer. Every small business that learns to integrate
intelligent systems expands the market for cloud computing, models, software,
and digital services. The fortunes of local communities and the fortunes of the
largest technology companies are far more aligned than either side sometimes
recognizes.
Perhaps communities
should begin asking a different question when negotiating major AI
infrastructure projects. Not simply, “How many jobs will this data center
create?” But, “How will this investment help our community create the
next generation of companies?” And will that hyperscaler vendor commit a
tiny percentage of the data center project budget to fund the next decade of
institutional infrastructure for the community in which it is built?
I would argue these
massive tech companies should provide funding for accelerators and coworking
spaces and meetup groups. They should conduct AI literacy initiatives. They
should support university partnerships, commercialization programs, startup
competitions, and entrepreneur support organizations. A seven figure annual
commitment barely dents the budget of a data center project, but can have an
outsized positive impact on the community in which it sits.
Physical
infrastructure determines whether intelligence can exist in a place.
Institutional infrastructure determines whether that intelligence creates local
prosperity. The Data Economy has arrived. The question is not whether
artificial intelligence will reshape the economy. The question is whether we
will build (and fund) the institutions that allow our communities to shape it
in return.