The distance between these two groups is the distance between the pilot and the production line. A pilot flourishes under conditions no enterprise can carry to scale: a single team, data prepared by hand, a person within reach to correct the system’s errors, and a vocabulary the room already shares. A chef may cook by instinct for a table of four; a hundred kitchens require the recipe to be written down. A restaurant franchise is a recipe to which everyone has agreed.

Ask a company how many customers it has, and the honest answer has always been that it depends on who is counting. Sales counts accounts, finance counts legal entities, and support counts logins. For years, a person stood between the data and the decision, quietly reconciling the three. That labor made disagreement a tolerable economy. An agent removes the person. It will answer with complete assurance, drawing on whichever number it encountered first.

Most boards still decide by the number. The more prudent have begun to ask what stands behind it, and they call the question governance: who answers for a figure, what it includes, and whether it means this quarter what it meant last. They would not say semantic layer, and the vocabulary is unnecessary. The instinct is what matters. A board’s credibility rests on the figures it repeats, and defining them first is how that credibility is preserved. Data teams have practiced precisely this for some thirty years, largely uncelebrated, through budget cycles that rewarded whatever shipped first. The work every budget deferred, the definitions, lineage, and ownership, is the work the leaders are now found to have done.

A pilot is a stage of a strategy, and stages end. Ten pilots resting on ten definitions of a customer amount to ten small successes and no capability. A model may be licensed in an afternoon; a definition takes as long as the disagreement it resolves. Each institution will absorb the frontier’s next release at the speed its definitions permit, and that speed was set in the years when no one was asking.

Line up the major estimates and a pattern holds with unusual consistency. Acemoglu’s model finds AI adding about one percent to American output over ten years. The International Monetary Fund, a multilateral body with no commercial exposure to either outcome, estimates global growth could rise by up to eight tenths of a point annually, a real but bounded contribution. Goldman Sachs, whose business includes financing the capital expenditure this technology requires, estimates a seven percent addition to global output over the same decade. Anthropic is more careful than the other three, publishing three scenarios built on different assumptions about adoption and autonomy rather than a single prediction, with its own literature stating plainly that the future remains open. That carefulness rarely survives the trip into a board memo, where the scenario’s headline number, thirty-two percent in four years under the most favorable assumptions, does the same work a forecast’s number would have done anyway. The size of the estimate has tended to track the estimator’s proximity to the technology’s commercial success, a property of forecasting from inside the institutions built to profit from the outcome, and one worth naming plainly before any of these numbers enters a strategic plan.

Each of the four rests on assumptions, about adoption speed, about which tasks yield to automation, about how quickly displaced labor finds new work, that no one currently has the data to settle. What can be assessed is the authorship behind each number. A forecast produced by an entity with revenue riding on adoption answers a narrower question than the one usually asked of it: what conditions would make its business case true.

Latin American boardrooms tend to receive a narrower version of this range than their counterparts elsewhere, because the adversarial half of the conversation rarely travels with the optimistic half. In the United States, Acemoglu’s skepticism runs in the same outlets that carry Goldman’s optimism, and a board can weigh one against the other inside a single news cycle. The region has comparatively few independent economists publishing competing estimates of AI’s local effect, and comparatively many consultancies and vendors whose material arrives pre-filtered toward adoption. The gap is access to the argument against the information, at the moment a decision is actually being made.

The number a Latin American board repeats back to itself this year says less about artificial intelligence than about which report happened to reach the room first. Unevenly distributed scrutiny is a more durable disadvantage than any gap in compute or capital, because it compounds in every decision made before the range finally narrows.

I learned early which branches were reaching for light and which were only reaching: a sucker rising straight from the root with more ambition than any real hope of fruit in it, a limb crossing back over itself out of nothing but habit. I learned the difference between a cut that helps a tree and a cut that only satisfies the hand holding the shears, before I learned much else worth knowing.

Then came years spent attending to other things, things that answered faster, and the tree kept growing the way trees do when no one asks anything of them: outward, indiscriminate, more wood than fruit. Distance did not kill it. Neglect rarely moves fast enough to look like harm while it is happening.

What I found on returning was a thing heavier with thorned deadwood than with anything a hand would want to reach for, its shape barely legible as the tree I still called mine.

Pruning, done honestly, is closer to negotiation than to correction. Some branches come away clean, and the tree barely registers the loss. Others bleed sap for days over a cut that seemed, at the time, like the obvious one to make. There is no way to know in advance which kind of cut you are making, and some seasons the fruit has been smaller for it.

A tree does not stay pruned. Whatever shape it is given this season drifts back toward wildness by the next, and the hand that wants a harvest must return on the tree’s schedule, not its own. I used to treat tending as a task with an end point, a box closed once and shelved. The tree has been patient in teaching me otherwise, more patient than I deserved, given how long the shears sat unused.

Last season it gave more fruit than it had in years, small and imperfect and entirely worth the trouble, and I stood beneath it longer than the harvest required. It never asked to be forgiven. It only kept making room, branch by branch, for the hand that finally came back.

The two facts are not separate. When a general-purpose technology becomes cheaper to run, the savings rarely stay idle. They are reinvested in more use, more reach, and more dependence. The steam engine did not slow the coal mines. It made steam economical in mills, ships, and workshops that had never justified it before, and consumption rose with every improvement. Each reduction in the price of a token enlarges the set of things worth doing with tokens, and capital moves to the edge of that newly viable set.

What Microsoft published alongside the numbers matters more than either figure. Mustafa Suleyman, who leads that division, laid out the architecture behind the reduction: every model in a product should be substitutable, and substitution is only possible when harness, context, memory, and action space are built independently of any single model family. Satya Nadella then repeated those four terms to shareholders on the earnings call, and that is where an engineering preference becomes a corporate doctrine. A blog post signals intent. Stated to the market, it becomes a commitment to build this way and to be judged by it. The doctrine identifies where the value has moved: the model is now the disposable component.

A price collapse is never general. It reaches only what can be swapped out without loss. Everything that survives substitution appreciates in the same movement, because the cheap component now depends entirely on it. The fall in the price of intelligence is therefore a rise in the value of everything that directs it: the record of what an institution has decided and why, the standard by which it judges an answer correct, and the accumulated context that makes a general system useful in one business rather than in any business. Models are bought. Systems are accumulated. Only one of the two responds to a discount.

The region consumes intelligence heavily and owns very little of it. We account for 14 percent of the world’s visits to artificial intelligence platforms and rank third globally in downloads of generative applications, while capturing 1.12 percent of global investment in the technology against 6.6 percent of global output. Those are the numbers of a large, active, and undercapitalized market, and cheaper intelligence makes a market larger. Consumption follows price. Ownership follows accumulation.

Peru has written careful law requiring these systems to explain themselves, but it has yet to build the systems that would do the explaining. The rules are addressed to a machine still owned somewhere else. Microsoft now sells the layer those rules describe, offering through Foundry the architecture it built for itself, and the offer will be attractive and, on price, correct. A supplier can ship the infrastructure. What gives it value is the institution’s own record inside it, written slowly, decision by decision, in the only place it can be written. That part has no vendor and no discount.

The savings are real, and they have already been claimed. They settle wherever something was built to receive them, which is how a 90 percent reduction in the price of intelligence passes through the region that consumes it most heavily and leaves no deposit. Peru will be able to buy the architecture within the year and still not produce the explanation its own law demands, because that explanation is not a feature anyone ships. It is a record of decisions, and only the institution making them can write it. The distance will widen fastest in the years the technology is most affordable, and those years have already begun.

The numbers deserve their full weight. Hyperscalers committed more than $600 billion in capital in 2026. Venture capital concentrated $239 billion in AI companies in a single quarter. BlackRock projects an additional $5 to $8 trillion in AI infrastructure through 2030. The physical dimension of this buildout connects the intelligence economy to the geology of the Andes and the hydroelectric capacity of the Amazon basin. Copper demand is rising 110,000 metric tons annually from data centers alone. Lithium markets are moving from surplus to deficit as battery storage scales to match AI’s electricity appetite. Latin America holds precisely the reserves the buildout requires. The region is not reading this as the structural matter it is.

This is where the argument turns uncomfortable. Latin America has been in this position before.

In the early years of this century, the region’s commodities fed China’s industrialization. Copper prices tripled. Soy exports multiplied. Governments collected royalties, sovereign funds accumulated assets, and the terms of trade moved in the region’s favor for nearly a decade. What the region did not build was the institutional capacity, the industrial base, or the knowledge economy to participate in the value being created upstream. When the supercycle ended, the royalties ended with it. The structural position was unchanged.

The AI buildout is not a commodity supercycle in the classical sense. The technology is newer, the pace faster, the geopolitical stakes sharper. But the structural position Latin America is being assigned is recognizable: supplier of the inputs, consumer of the output. Seventy percent of the leading foundation models are built in the United States. Twenty-five percent in China. The Inter-American Development Bank projects regional growth at 2.1 percent in 2026, below the emerging market average, even as the region supplies the physical substrate of the most consequential capital cycle in modern history. The region has no equity stake in what is being built on top of its copper.

The region does have AI activity. Brazil’s startup ecosystem is producing companies of genuine scale. Its regulatory framework is the most advanced in the hemisphere. Chile and Mexico lead regional AI readiness indices. Peru is developing AI skills at one of the fastest rates in the region, according to the IDB. These are not trivial facts, and they are not the point. Activity at the adoption layer is not the same as participation at the production layer. Training workers to operate AI systems is not the same as building the systems. Acquiring a US data-intelligence company, as Nubank did, is not the same as developing the intelligence capability in-house. The distinction matters because value concentrates at the production layer, and it has never done otherwise.

What makes the present moment consequential is not the size of the buildout. It is the pace at which the value chain is consolidating. Foundation model leadership is concentrating in a handful of firms. Infrastructure is concentrating in five hyperscalers. The window to position as a producer rather than a consumer is measured in years, not decades. Sovereign AI programs in France, Germany, Saudi Arabia, and India are being designed precisely because their governments understand that the transition from AI consumer to AI producer becomes structurally harder the longer it is deferred. These are not technology investments. They are bets on where in the value chain their economies will sit when the buildout is complete.

Latin America is not having that conversation at scale. The region is having a regulatory conversation, a talent conversation, and a startup conversation. All of those matter. None of them address the foundational question: in the economy organized around intelligence, what will Latin America produce?

The copper and the lithium will be extracted regardless of how that question is answered. The issue is what gets built alongside them.

Peru is moving now, compressed into different infrastructure. The National Artificial Intelligence Strategy 2026–2030 commits to targets with timelines that close as they unfold. Forty percent of companies deploying AI meaningfully by 2030. Critical sectors including healthcare, finance, education, and justice integrated by September 2026. Not aspirations. Requirements with enforcement dates. The regulatory sandbox is operational. The risk framework is drawn. The compliance timelines are immovable.

The current position reveals the distance to travel. Fourteen percent of Peruvian companies deploy AI meaningfully. The region averages 40 percent. Peru observes the gap. Peru is moving to close it. The adoption concentrates where velocity has already gathered: retail at 31 percent, media and communications at 29, finance at 26. These are not theoretical deployments. They are production points where the compression is already happening, where companies are learning what works and what breaks.

The stakes are domestic. According to the Inter-American Development Bank, if executed correctly, AI doubles Peru’s GDP by 2030 and raises productivity to 7.3 percent. That is not marginal improvement. That is structural transformation. That is why the timeline cannot extend beyond what Peru can sustain.

The compression creates friction at every layer. Universities must translate applied research into deployable systems, not in academic years but in quarters. Companies must integrate AI into operations while training workforces in parallel. Government must enforce regulation while the technology is still being deployed, creating rules for systems that are learning as they operate. The private sector must commit capital before the proof is complete. Risk and speed move together.

Kubitschek paid the cost in visible ways: $1.5 billion in debt, 43 percent inflation, institutional strain that outlasted his term. Peru will pay similar costs. Organizational disruption. Talent migration as companies compete for scarce expertise. The friction of institutions moving faster than they are built. The difficulty of governing what you are simultaneously deploying. These are not problems to solve. They are prices to pay.

The distinction is not in avoiding cost. It is in choosing which cost to pay. Speed has a price. Delay has a steeper one.

By year-end 2026, Peru must produce evidence: significant funding rounds in AI infrastructure, startup exits that required Peru’s regulatory environment to succeed, research breakthroughs that could only happen where governance and velocity moved together. These are not hopes. They are metrics. They are how the transformation measures itself.

This is a compression. Decades collapsing into five years. Institutions built while velocity increases. Governance established as deployment happens. The cost will be visible: organizational strain, difficult transitions, the friction of speed. But also the possibility of transformation that delay would never permit.

Whether Peru recognizes this moment for what it is, whether it moves like Kubitschek moved, understanding that some transformations cannot be deferred, that speed itself becomes a form of discipline when stakes are high enough, will determine whether this strategy becomes history or remains a statement.

A structure is an agreement among its parts. Each link consents to carry the next, each stone to bear the one above it, and the whole stands because the agreement is kept in silence. When a single link withdraws its consent, nothing dramatic announces the change. The line simply stops trusting itself. One learns, in the space of an afternoon, that the body had been an act of cooperation all along, and that cooperation can be revoked without notice.

For years I had studied a harder grammar than health, the kind learned on the mat, where pressure is a language and to yield is a sentence another man finishes for you. That grammar asks the structure to hold under someone else’s full weight, and the structure can no longer be asked. The door is not locked. It is only, this season, a door I am not permitted to walk through.

And so the rooms grew quiet in more than one sense. The family is away. The mat is empty. The apartment keeps a stillness I did not choose and cannot give back, the kind that gathers on the furniture in the late afternoon and asks, without insistence, what I mean to do with it. There is no one here to perform recovery for, no audience to the small negotiations of a body that has learned caution. There is only the long evening, the city going on beneath the window, and the tenant that has moved into the structure and shows no intention of leaving on my terms.

I want to be exact, because the temptation is to ennoble this, and it does not deserve ennobling. Pain is not a metaphor. It does not arrive carrying meaning the way a guest arrives carrying wine. It is a fact, and its first labor is subtraction. It takes the ease. It takes the motion one never had to consider. It takes the oldest illusion, that the body is a thing we command rather than a thing we have been quietly negotiating with from the beginning. What remains when the subtraction is finished is not wisdom. It is only what was load-bearing from the start, standing in the emptied room, finally visible.

There is a temptation older than self-pity, and it is the wish to make the world answer for the wound. It is the easy path, easy in the way any downhill is easy, asking nothing of a man but that he stop resisting the slope. Pain will tell him that, having come to him unfairly, it has left him owed, and that the debt may be collected from anyone within reach. Turned that way, it hardens into a kind of doctrine, a permission to do harm, so sure of its own grievance that it mistakes ruin for justice.

I do not imagine myself above this. I feel its pull as plainly as any man, and on the worst evenings more plainly than I would choose to admit. Whatever I have learned was not learned by being stronger than the pain. It was learned by watching, without flattery, what pain has made of better men than me, and by recognizing that I am made of the same material, bound by the same limits.

The other direction is uphill, and the climb shows. The same fact that subtracts can also gather, though only if a man does the gathering himself. The hours the mat once took belong now to quieter work, and a structure forbidden its old motion holds very still, and a still thing can be worked with great care. What I had set aside for a more convenient season turns out to need this one. I am building in the evenings the injury emptied, sharpening the things that ask nothing of the body to be made sharp. The wound did not give me this. It cleared the ground on which it became possible, and I have stopped pretending the two are strangers.

None of this is resolution. The line has not been restored to its old trust, and I have no honest reason to expect the pain to end on a schedule I would choose. There is no turn here that closes the account, no redemption that makes the breach worth it. I would undo it if undoing were on offer. It is not, and that is nearer the truth of the thing than any consolation. The season is what it is. The family is away, the structure is compromised, the mat is empty, and the only part still within reach is what gets made in the silence.

That is the part no one can do in my place, and the part the tenant, for all its occupation, cannot reach. What a man builds when the ordinary roads are closed, when there is no one to applaud the building and no reward waiting to be collected at the end of it, is the truest thing he will ever learn about himself. The tenant will stay as long as it stays. I have decided, without sentiment and without any promise of being repaid, to make good use of the company.

Colombia approved CONPES 4144 in February 2025. Eight ministries coordinating capacity. Talent, research, applied innovation. Not a document. A commitment of capital. Brazil announced Rio AI City in April, before Peru’s regulation had fully taken hold. 1.8 gigawatts by 2027. Infrastructure anchoring research and capital. Mexico moved through AWS and Microsoft. These were not pronouncements. Capital deployments. Signals to the global market: we are building the infrastructure layer.

The regulation Peru produced is more sophisticated than any of theirs. The thinking clearer. The framework more precise. And yet sophistication does not accelerate time. Peru learned what careful preparation sometimes teaches: being right about what should govern this moment is not the same as moving fast enough to determine what actually will.

On May 1st, Peru announced its National Artificial Intelligence Strategy. The framework was complete. The thinking had been done. The region, having already moved, was watching to see whether announcement would become implementation.

History shows this pattern. India’s technology sector became globally preeminent in services delivery while the foundational research infrastructure consolidated elsewhere. Research depth, talent retention, capital for indigenous innovation: these moved to where opportunity had already concentrated. By the time the gap became visible, the asymmetry had hardened into structure. Argentina followed this same path. Fintech execution at global standards. Talent migrated to where infrastructure already existed: North America, Europe. The places where capital and opportunity had already consolidated. In both cases, the sector thrived. In both cases, the region’s capacity to innovate beyond execution, to build rather than build upon, remained constrained. The execution excellence was real. The constraints were structural.

This choice point has appeared before. Peru faces it now with compressed time. The 18-month consolidation window is measurable and real. Cloud infrastructure shows the pattern: when first movers commit capital and infrastructure, talent and research follow within that narrow band. After that window closes, reversal becomes difficult. The gap hardens. The tier that consumes becomes permanent.

What determines the outcome is visible. The timeline is unforgiving. The private sector must respond to the announced strategy with matching capital within four months. Universities must consolidate applied research into industry partnerships within five months. By year-end, Peru produces evidence that the framework enabled production: a significant funding round, a startup exit, a research breakthrough that required Peru’s conditions to develop.

The conditions remain. Peru has governance frameworks that India and Argentina lacked at this moment. It has adoption momentum. It has institutional coherence. What Peru possesses that they did not is clarity about the pattern, and time enough to act on that clarity. Whether it does is the question of the next twelve weeks.

Peru had reason to watch this moment carefully. The country had spent the previous year building governance frameworks: Law 31814, risk categories, regulatory sandboxes. The machinery was in place. The question was whether it could become something more than constraint. Could governance become a platform?

Ten months have passed. What Peru has observed in that interval reveals something essential about how Latin America is choosing—or failing to choose—its place in an AI economy that will not wait.

The Latin American AI Index 2025 describes a region organizing itself into three tiers of capability and intention. This is not a hierarchy of sophistication. It is a fracture.

The Pioneers—Brazil, Mexico, Chile—advance on multiple fronts simultaneously. Research infrastructure thickens. Governance matures. Capital accumulates. They sustain complexity because they invest across dimensions. Their scores exceed 60. Their distance from the rest widens each quarter.

The Adopters occupy a different position. Peru, Colombia, Argentina move with velocity on adoption and governance, but their scores cluster between 40 and 60. They use AI at scale. Citizens deploy it naturally. Companies integrate it. The adoption is real. But the capacity to create—to research, to build on foundations, to move from consumption to production—advances more slowly. The gap between what they adopt and what they produce widens. This is not a problem of access. Models are available everywhere. This is a problem of institutional depth.

The Explorers struggle to keep pace. Eleven countries cannot yet cross the 50-point threshold in digital infrastructure. The region captures 1.12 percent of global AI investment while representing 6.6 percent of global GDP. For these countries, the frontier recedes faster than they can approach it.

What Peru observes from its position in the Adopter tier is clarifying: the three velocities are hardening into a structural division. The window for movement between tiers is narrowing.

Peru’s governance framework arrived at the moment of maximum opportunity. The AI law is sophisticated. The regulatory sandbox exists. The international positioning is clear. Institutional leaders understand the stakes.

And yet adoption races ahead while production capacity consolidates slowly.

Peru leads the region in traffic to AI platforms. Citizens embrace the tools. Companies experiment. But this dynamism has not translated with equivalent force into private investment in AI development, into indigenous research capacity, into the kind of foundational work that turns tools into competitive advantage. High adoption without production infrastructure creates a particular vulnerability: Peru becomes excellent at deploying what others build, but does not build.

The governance advantage that Peru engineered becomes a platform for management of adoption rather than a platform for acceleration of creation. This is not failure. It is asymmetry. And asymmetry, extended over time, becomes dependency.

When open weights became available, the response across LatAm was not uniform.

In the Pioneer tier, the infrastructure absorbed new models and immediately moved to applied problems. Research groups integrated them. Companies built on them. The openness accelerated already-existing momentum.

In the Adopter tier—where Peru sits—the response was more complicated. The models were available. The regulatory framework permitted their use. But the gap between access and productive application revealed itself quickly. Having the tool is not the same as having the institutional capacity to deploy it at scale, to adapt it to local contexts, to build on it with confidence.

In the Explorer tier, the models arrived into environments where infrastructure and capital could not yet absorb them. Access without capacity is a different kind of constraint.

The open weights moment, rather than equalizing, illuminated the existing fragmentation. It showed which countries had the institutional depth to move fast, which had adopted rapidly but lacked production capacity, and which could not yet convert access into action.

Peru, in the Adopter position, faced a choice it did not fully recognize. The governance framework could become what it was designed to be—a platform for responsible innovation, an anchor for production capacity, a magnet for talent and capital. Or it could remain what it was becoming—a compliance infrastructure for managing adoption of what others built.

Ten months after open weights, the answer is not yet written. But the choices are becoming visible.

Peru could consolidate its governance advantage into a genuine platform for production. This requires translating the regulatory legitimacy into actionable incentives: for research, for applied innovation, for the kind of long-term investment that builds indigenous capacity. It requires universities and the private sector aligning around a shared goal: moving from adoption to creation.

Or Peru could optimize what it already does well: manage the adoption of frontier models responsibly, serve as a intelligent consumer of global AI, lead the region in governance maturity without the burden of innovation capacity.

The first path is harder. It requires sustained institutional coherence. It means resisting the pull toward quick wins. It means building talent pipelines and research infrastructure in an environment where capital still flows more easily toward applications than toward foundations.

The second path is safer. It plays to Peru’s existing strengths. It consolidates the regulatory advantage without overextending into domains where the region still lacks depth.

The region watches Peru’s choice because Peru’s choice will echo. If Peru can translate governance into production capacity, it becomes a proof point. Other Adopters will follow. The three velocities might remain stratified, but the middle tier will have agency. If Peru consolidates governance as a management function without expanding production capacity, the fracture hardens. The Pioneers pull further ahead. The Explorers fall further behind. The middle, no matter how sophisticated its governance, remains positioned between consumption and creation—benefiting from neither as fully as it might.

The opportunity that open weights represents is real. Lower infrastructure costs. Democratized access. Communities forming around shared foundations. For a region with capital constraints and talent challenges, this matters. The SaaS revolution of the 2000s showed that shared infrastructure can accelerate innovation exponentially. Open-weight AI models can do the same.

But the window compresses. As the frontier accelerates, as the Pioneers consolidate their advantages, as capital concentrates around established players, the opportunity to move between tiers narrows. The region has perhaps eighteen months to two years—no more—to show whether it can translate adoption velocity into production capacity, whether governance can become a platform for creation, whether the middle tier can hold or will simply manage its own decline relative to global leaders.

Peru has the conditions to lead this translation. It has governance frameworks. It has adoption momentum. It has institutional coherence that many neighbors lack. It has the visibility to demonstrate that regulatory sophistication and innovation capacity can reinforce rather than compete.

The question is whether Peru will choose to use it.

But the same period shows relative decline in the dimensions that determine long-term value creation. Research, development, and advanced human talent registered backward movement. The result is structural asymmetry. Peru has built regulatory frameworks faster than it has built the ecosystems to deploy them effectively. It has created the architecture for responsible AI adoption without yet creating the architecture for responsible AI creation. The gap is not accidental. It reflects where investment has concentrated and where it has not.

The technical enablers paint a mixed picture. Data availability has improved significantly and now exceeds regional averages. Critical infrastructure, by contrast, advances more slowly. The effect is an environment that facilitates daily deployment of existing AI tools but constrains deeper technological development. Peru leads the region in traffic to AI platforms—users adopting rapidly, naturally, at scale. Yet this adoption does not translate proportionally into private investment, business development, or productive scaling. Users deploy more. Creators build less.

This pattern matters because adoption without production capacity creates dependency. Peru can adopt every frontier model and every open-weight offering. But if the region does not develop the talent and research infrastructure to understand those models, adapt them, and build upon them, it remains a consumer of tools rather than a participant in their creation. The governance advantage created by Peru’s early regulatory moves—and it is an advantage—only compounds the urgency of this gap.

The diagnostic from the ILIA Index is clear. Peru advances with strength in governance and adoption while still constructing the foundations for innovation capacity. The challenge ahead is not regulatory. Regulation is done. The challenge is strategic: converting advanced regulatory frameworks into effective incentives for investment in talent, applied research, and technological entrepreneurship. This requires a deliberate reorientation of how the state, the private sector, and universities allocate capital and attention.

The opportunity is equally clear. Governance, properly deployed, can become a platform for attracting the capital and talent required to move from adoption to innovation. The regulatory legitimacy Peru has earned can be leveraged to demonstrate that the country is a credible place to invest in AI development, not just AI deployment. Universities can use the framework as foundation for building research capacity. The private sector can use it as assurance that long-term investment in this space will be supported by stable, sophisticated policy.

If Peru consolidates governance as a foundation for productive capability development, the progress visible in the index can translate more consistently into sustained economic and technological impact. The frameworks are in place. The adoption is happening. What remains is the work of directing investment and talent toward the deeper work of creation. That translation is not automatic. It requires strategy.