Looking innovative without paying the cost of being so
Last week I presented a trilogy of analyses, —The Silent Retreat of AI in the Dominican Republic, The Retreat No One Wants to See and The Price of Complacency—, which documented with technical evidence the decline of our innovative capacity and the economic cost of state inaction in the face of artificial intelligence. The…

Last week I presented a trilogy of analyses, —The silent retreat of AI in the Dominican Republic, The retreat nobody wants to see, and The price of complacency—, which documented with technical evidence the decline of our innovative capacity and the economic cost of the state's inaction on artificial intelligence.
The diagnosis was clear: between 2020 and 2025, the Dominican Republic fell seven positions in the Global Innovation Index (from 90 to 97) and three in the Latin American Artificial Intelligence Index (from sixth to ninth place). This drop was not a methodological fluctuation, but the predictable result of strategies without execution, investments without return, and policies without consequences.
According to Tabuga Intelligence data, the country will spend more than US$300 million on imported AI solutions during 2025, while local investment in technological development barely reaches US$10 million. No Dominican university appears in global rankings, we produce four patents a year, and we keep exporting technical talent to countries that do invest in innovation ecosystems. Worst of all: the State, instead of catalyzing the market, competes with it, concentrating resources on its own projects such as CiudadanIA, which exclude MSMEs from the innovation process and perpetuate dependence.
That was the crux of the criticism: the Dominican Republic does not lack resources, but conversion. We invest, but we do not transform; we adopt technology, but we do not produce it. The problem is structural —a dysfunctional ecosystem where talent, capital, and infrastructure do not converge to generate endogenous value.
And right after that assessment, on October 9, the government announced the Center of Excellence in Artificial Intelligence (CEIA) and local production of microprocessors. Both announcements share a worrying pattern: they prioritize narrative over capacity, image over infrastructure, headline over measurable impact.
Copying names, not capabilities
Chile founded the National Center for Artificial Intelligence 2021 in (CENIA), a public-private consortium that brings together five universities, technology companies, and the State under a distributed governance model. CENIA publishes research in indexed journals, trains PhDs in machine learning, and collaborates with similar centers in Argentina and Uruguay under the integration framework proposed by ECLAC. Building it was costly: sustained investment, complex institutional coordination, and accountability against international metrics.
The Dominican Republic, by contrast, announces CEIA: a name that replicates the acronym with one letter of difference, but without the ecosystem that sustains it. There is no mention of university coordination, independent business participation, or regional integration. There is no public budget, no technical governance structure, no measurable impact indicators. The initiative seems designed to compete symbolically with Chile, not to collaborate with the region. We want the prestige of a center of excellence without paying the price of genuine collaboration: trained researchers, peer-reviewed publications, computing infrastructure, and multi-year budget commitment.
Where ECLAC proposes south-south bridges, we prefer the photo of the newly inaugurated bridge. Where other countries spend years building capacity, we want immediate recognition. It is the illusion that an announcement equals an achievement, that naming something makes it real.
Confusing manufacturing with innovation
The announcement of manufacturing microprocessors in the country reveals the same confusion between appearance and substance. NVIDIA, whose name ironically evokes our aspiration, is not a technology leader because it makes chips —Taiwan does that for them— but because it designs architectures, develops software (CUDA) and generates intellectual property valued at hundreds of billions of dollars. Jensen Huang is not rich because he owns factories; he is rich because he owns patents. Manufacturing is maquila; design is wealth.
In the 90, Costa Rica and Colombia bet on attracting multinationals with tax incentives, believing that the physical presence of technology companies would generate knowledge transfer. Intel arrived in Costa Rica in 1997; after two decades, the plant closed without leaving significant R&D centers, only assembly capacity. Colombia received call centers, not innovation labs. Both countries today have patenting rates lower than those of Chile or Argentina, which prioritized developing their own capabilities over attracting passive foreign investment.
Manufacturing microprocessors without design capacity is repeating the Costa Rican mistake with more sophisticated technology. It is low-value-added industrialization disguised as technological sovereignty. It looks like innovation —there are chips, there is technology, there are jobs— but it does not build the future; it rents the present. The real cost of innovation includes training engineers with doctorates in semiconductor architecture, funding years of research in integrated circuit design, and accepting the risk that most prototypes will fail. Nobody wants to pay that cost; we prefer the factory that can be photographed.
The aesthetics of modernity
The announcement has symbolic value, and that is not trivial. It signals that the State recognizes AI as a strategic priority and that critical pressure generated an institutional response. In that sense, there is a political merit: putting the issue on the public agenda and showing that technical criticism can mobilize government responses.
However, the gesture carries the same structural fragility we have denounced. It does not specify how CEIA will coordinate with universities, how it will retain talent, or how it will measure its contribution to GDP. There is no plan for locally manufactured microprocessors to contain Dominican design protected by Dominican patents. It is a reputational move that tries to counter an uncomfortable diagnosis without changing the structure that causes it.
In historical terms, it recalls the old colonial exchange of «mirrors and gold»: a narrative of superficial modernity offered in exchange for symbolic control. The State offers the image of a center of excellence and a chip factory; the private sector and universities cede prominence in the innovation ecosystem. Everyone gets headlines; no one builds capacity. Everyone looks innovative; no one pays the cost of being so.
The real cost includes: funding doctorates for a decade before seeing a return, accepting that 80% of startups will fail, allowing researchers to publish in international journals even if they criticize the government, investing in expensive computing infrastructure that may become obsolete, and submitting to international metrics that will reveal our limitations before showing our achievements. That is the price Chile, Argentina, and Uruguay are paying. We want the recognition without the bill.
What it would cost to do things right
If the government were willing to pay the real price of innovation, three investments would be more effective than a centralized center without a budget or a factory without design capacity:
- Create a competitive fund of US$50 million for applied AI research, with evaluation by international peers and mandatory participation of local MSMEs. Let universities, companies, and research centers compete for resources on technical merit, not political proximity. Let the first three years produce more failures than successes, because that is how research works. Cost: accepting that we are investing in projects that will not generate immediate headlines.
- Integrate CEIA into the CENIA-Argentina-Uruguay-Peru-Panama ecosystem under shared collaboration and comparable metrics. Let the Dominican center be a a regional hub, not an island. This would imply publication standards, researcher mobility and impact benchmarking. Cost: accepting that other countries are ahead of us and that we must learn from them, and that collaboration requires giving up symbolic sovereignty.
- Prioritize design over manufacturing. Instead of investing in producing generic chips, fund the creation of a university-industry consortium to design specialized architectures. Chips for natural language processing in Caribbean Spanish. Chips for optimizing power grids in tropical countries. Let the intellectual property stay in the country. Let Taiwan manufacture, if necessary; we design. Cost: waiting ten years to see commercial results, training an entire generation of semiconductor engineers, and accepting that the first designs will be inferior to international ones.
The indices don't lie
When the new indices are published in 2026, we will have the definitive answer. The Global Innovation Index doesn't reward announcements; it rewards patents, publications, scalable startups and talent retention. The Latin American AI Index doesn't measure how many centers of excellence we inaugurate, but how many models we train, how many papers we publish and how many AI products we export.
The indices will measure whether we pay the cost or merely buy the appearance. They will measure whether CEIA produced doctorates or press conferences, whether chip manufacturing generated patents or assembly jobs, whether we invested in capacity or in communication.
The Dominican Republic wants to be Silicon Valley without having solved Santo Domingo Este. It claims excellence without infrastructure, without retained talent, without scientific output, without intellectual property. We want international recognition without international accountability. We want the photo with the prize without competing in the category.
Real excellence isn't decreed, isn't announced at press conferences and isn't manufactured in technological maquilas. It is measured in indexed papers, built through regional alliances and sustained by talent that decides to stay because the ecosystem retains it. Both CEIA and the announcement of microprocessor manufacturing lack auditable impact indicators and binding regional commitments. They are gestures designed to look innovative without paying the cost of being so.
Time will tell whether we are willing to pay the bill. So far, we have only wanted the discount.