economics · quantitative analysis · 2026-05

Economics of the AI-corporation regime.

What the public conversation around Argentina's AI-corporation regime is not quantifying: how much it costs to incorporate an agent-entity, how much it costs to operate it for 24 months, and where Argentina is structurally competitive. For economic advisors, tech-business journalists, and founders deciding where to incorporate.

Central thesis. The regime announced by Minister Sturzenegger can be structurally competitive on total cost of operation, but only if the technical layer exists and is sovereign. If Argentine founders end up using Stripe + Twilio + AWS for their critical flows, the differential collapses and AR becomes a way-station to Delaware or Wyoming. The existence of an AR-sovereign open-source toolkit (ar-agents) is the difference between capturing value and being a transit window.

Methodological honesty. The figures below are approximations at the official USD MEP rate of 2026-05-13. They cover direct costs (regulatory fees, registered agent, certificates, basic accounting); they exclude banking setup, potential litigation costs, and the operator's personal expenses. Primary sources cited at the end.

1 · Incorporation cost (one-shot)

JurisdictionFormal costTotal + setupTimeLocal presence
Argentina (standard SAS today)ARS 60,000 (~USD 60)USD 200-5005-20 business daysYes · notary + IGJ
Argentina (proposed AI-corp)TBDEstimated USD 100-300Estimated 1-7 daysTBD
Wyoming DAO LLCUSD 100USD 250-4501-3 daysRegistered agent (USD 100/y)
Marshall Islands MIDAOUSD 6,000-9,500USD 8,000-12,0002-4 weeksMIDAO mandatory
Estonia OÜ + e-ResidencyEUR 25 (~USD 27)USD 250-4001-3 dayse-Residency card (~EUR 100)
Delaware C-Corp (Stripe Atlas)USD 500USD 500-8002-7 daysRegistered agent included
Singapore VCCSGD 300 (~USD 230)USD 2,000-4,0001-2 weeksNominee director mandatory

Preliminary conclusion: Argentina (AI-corp) and Wyoming/Estonia are structurally comparable on incorporation cost. Marshall Islands is off the menu on cost. Delaware is the global default by reputation, not by cost.

2 · 24-month operating cost (TCO)

An AI-corp operates for 24 months. Typical annual loads: renewal fee, basic accounting, minimum fees + taxes assuming no revenue, technical infrastructure.

JurisdictionIncorporationAnnual operationsTech infra (24m)Total 24 months
Argentina (AI-corp)USD 300USD 600USD 0 (self-host ar-agents)USD 1,500
Argentina (AI-corp + Cloud Studio)USD 300USD 600USD 1,200 (USD 50/mo)USD 2,700
Wyoming DAO LLCUSD 400USD 250USD 600 (Stripe + Twilio basic)USD 1,500
Estonia OÜUSD 350USD 400USD 600USD 1,750
Delaware (Stripe Atlas)USD 800USD 600 (Atlas+RA+tax)USD 600USD 2,600
Marshall Islands MIDAOUSD 10,000USD 3,500USD 600USD 17,600

Reading: over 24 months, an Argentine AI-corp on self-hosted ar-agents competes directly with Wyoming on TCO. If it needs managed hosting + regulator-ready audit (Cloud Studio), it rises to USD 2,700, still below Delaware. MIDAO falls off the menu on cost. Estonia is the closest competitor if founders prioritise the EU seal.

The frame that matters for the legislative debate: the AR regime does not need to "win on price", it needs to not lose against comparable options. On TCO, it is already there. The differential is decided on other axes: legal certainty, quality of technical infrastructure, and international reputation. The three are buildable.

3 · Value captured by jurisdiction

If an AI-corp invoices USD 100,000/year, who captures what? Approximation for the first year of operation (rounded numbers):

JurisdictionIncome taxValue-added taxRegulatory feesLocal capture
Argentina (responsable inscripto)30%21% (deductible if supplier is also RI)USD 600/yearUSD 30,000-35,000
Argentina (monotributo top category)Included in fixed quotaIncluded~ARS 1M/yearUSD 8,000-12,000
Wyoming DAO LLC (pass-through)0% federal (pass-through to member)Sales tax per stateUSD 250/yearUSD 0-3,000 (federal)
Delaware C-Corp21% federal + 8.7% stateVariable sales taxUSD 800/yearUSD 25,000+
Estonia OÜ0% on retained earnings + 20% on distribution22% EU VATUSD 400/yearUSD 0 (if reinvested) or USD 22,000 (if distributed)

The real economic argument: Argentina captures more fiscal value per incorporated company than Wyoming or Estonia. If the law executes well and attracts even 5,000 Argentine AI-corps operating at an average USD 100K/year in revenue, the regime delivers USD 150-175 million annually in direct fiscal capture, before counting formal employment generated by integrators + services + auditors. That is the quantitative economic opportunity.

4 · Estimated formal employment

Every Argentine AI-corp, no matter how technically autonomous, will require:

  • 1 licensed accountant (General Companies Law, annual fiscal closing, balance sheet).
  • 1 lawyer or legal consultant (incorporation, bylaw amendments, potential litigation).
  • Audit services (internal + eventually external for larger companies).
  • Technical infrastructure (developer ops if the company self-hosts, or Cloud subscription if it outsources).

If the regime attracts 5,000 productive AI-corps in 24 months, indirect formal employment generated approaches 2,000-3,000 specialised positions (accountants + lawyers + dev ops + auditors). Not the mass employment factory, but high-skill employment in sectors where the country already has installed capacity.

5 · When Argentina loses the match

The argument against:

  • International legal reputation. Delaware is the default thanks to 100 years of case law + Court of Chancery. AR cannot compete on that axis. The trade-off is cost + speed + proximity jurisdiction for Latin American founders.
  • FX controls + operational restrictions. If the Argentine AI-corp invoices in USD, withdrawing those dollars abroad faces regulatory friction that does not exist in Wyoming or Estonia. The regime should provide a special FX framework or the incentive dilutes for international founders.
  • Macro volatility. A change of government in October 2027 could reverse the regime. Estonia, Wyoming, Delaware have multi-administration continuity. This is the most sensitive factor for foreign founders.
  • Lack of local VC specialised in AI corps. If an Argentine AI-corp needs to raise a Series A, it still depends on US/UK VCs. Estonia mitigates this through its e-Residency program; AR still has no equivalent.

6 · For press coverage

The three headlines defensible with these figures:

  1. "Argentina could capture USD 150-175M annually in fiscal revenue if the AI-corp regime attracts 5,000 entities in its first 24 months." Defensible number.
  2. "AR's AI-corp costs less to operate than Delaware or MIDAO; it sits at Wyoming and Estonia levels." Defensible against the international benchmark.
  3. "The technical layer the regime needs already exists, is open-source, and was built by a single Argentine dev. The alternative was to import it." That is the tech-sovereignty angle.

7 · Sources

8 · Caveats

  • AFIP/ARCA rates can change; the FX rate too. Figures are as of 2026-05-13.
  • Does not include potential litigation costs, AAIP sanctions, BCRA fines, or IGJ infractions.
  • The “5,000 companies in 24 months” estimate is a reasonable upside scenario, not an official projection. For comparison: Wyoming has accumulated ~4,000 DAO LLCs since 2021. Estonia has ~139,000 cumulative e-residents over 12 years; active companies are a fraction of that.
  • The author is a developer, not an economist or an accountant. These numbers require validation by a specialist before being used as policy basis. Open invitation to any economist who wants to review them publicly: /en/co-sign.