The machine that fails even with fuel

Cuba’s national power grid collapsed for the third time in ten days in mid-July 2026, cutting electricity across the island of 9.5 million people. Power was briefly restored the next morning but blackouts have worsened since; the nationwide failures blur into rolling local outages that have become a defining feature of daily life. The immediate mechanism is straightforward: the grid cannot generate enough power. But the causes beneath that sentence are three separate failures that reinforce each other.

The physical backbone of Cuba’s grid — its 16 thermoelectric plants, eight of which were already offline by early 2026 — is in catastrophic condition. Jorge Piñon, a senior energy researcher at the University of Texas, describes them as “old, broken and tired.” The Antonio Guiteras plant alone logged at least 17 unscheduled outages in 2026, accumulating 293 hours of downtime between January and late May from economizer defects. It was inaugurated in 1988 and has had no comprehensive capital maintenance since 2010. Cuba’s 1960s-era industrial base was never modernized to manufacture thermoelectric equipment domestically; grid architecture remains locked into large centralized plants with no distributed generation fallback. Even if fuel were fully restored tomorrow, the physical asset base cannot carry the load.

On the fuel side, the crisis has a clear trigger. Cuba depended historically on Venezuelan oil. The US military removed Venezuelan President Nicolás Maduro from power on January 3, 2026, and Washington imposed a six-month oil blockade that severed the fuel lifeline. Energy Minister Vicente de la O Levy said the government faces “a total absence of fuel” and lacks access to spare parts. Foreign companies — hotel operators, airlines, miners, shipping firms — have departed; an electric-car importer reported containers stranded in Kingston and China with no delivery timeline.

But the blockade is an accelerant of a pre-existing failure, not its sole origin. Piñon’s characterisation of the plants as “old, broken and tired” predates January 2026. The blockade is necessary but not sufficient for the collapse; the pre-existing plant weakness means the blockade’s independent contribution cannot be precisely isolated. Lift the blockade tomorrow and the grid still fails under load with its current infrastructure.

Summer heat — mid-30s Celsius with 80 percent humidity — converts electricity from convenience to survival necessity: water pumps, phone charging, food refrigeration, sleep. Thermal plant efficiency drops roughly 0.27 percent per degree Celsius of ambient temperature increase, a thermodynamic relationship established in engineering literature. “An hour isn’t enough time to run the pump to get water or to charge phones,” Alberto said during a protest in Havana’s Vedado neighborhood. “People want the government to act right now.”

The monopoly that foreclosed every alternative

The fuel shortage is not just a function of the blockade — it is the product of a decades-old procurement structure that eliminated every alternative before the crisis arrived. Cuba’s energy model depended on a single primary supplier — Venezuela — for subsidized oil priced below market, which destroyed the economic incentive for procurement officials to seek alternatives. The state-enterprise monopoly on energy procurement lacked competitive pressure, procurement expertise, and contractual flexibility to source from non-traditional suppliers. No domestic manufacturing capacity exists because the 1960s-era industrial base was never modernized under the state planning model that treated imported subsidised oil as the solution.

The Cuban government announced 176 measures to expand the private sector and attract investment. As of reporting, these had not been enacted. No implementation timeline, accountability mechanism, or enforcement trigger exists. The US State Department dismissed the measures as “superficial smoke signals” — language that corroborates the finding that the announcement functions as a policy signal, not an operational programme. The gap between announcement and implementation is itself a structural failure: reforms without enforcement mechanisms reproduce the pattern that produced the crisis.

This institutional-economics reading identifies the state monopoly on energy procurement as a root condition. Removing the current blockade without reforming this structure would leave Cuba dependent on whatever single supplier the US allows — reproducing the same fragility with a different supplier’s name.

The fragmentation that blocks a settlement

Even if the procurement monopoly were broken and fuel were available, the crisis would persist because no single institutional actor holds the authority to negotiate a resolution. Executive authority is fragmented across the formal government — the Díaz-Canel administration — the military, and the Castro family network. The revolutionary governance structure intentionally concentrated power in informal networks rather than institutional processes, creating a system that resists adaptation as the original leadership generation ages out. The transition to Miguel Díaz-Canel left the family network as a parallel authority without official mandate.

Without a single, formally mandated negotiating authority, freelance actors occupy the negotiating space. The US leaks details of these talks publicly — an act that itself signals asymmetric information advantage and would not occur if a recognised, credentialed negotiating counterpart existed.

This reading — institutional fragmentation as twin root alongside physical plant degradation — reaches the same terminus as the monopoly reading through a different path: the physical asset base cannot carry load even with normalised supply, and the absence of a singular institutionally backed authority prevents any settlement that could lift the blockade. The two readings survive alongside each other because the source material supports both, and the disagreement reflects which root condition the analyst privileges as more consequential to recurrence prevention.

President Trump added to the fragmentation problem from the other side. In March 2026 he said: “Whether I free it, take it, I think I can do anything I want with it.” The maximalist US posture is unresolved: if US objectives are not achievable through negotiation regardless of Cuban institutional coherence, then Cuban institutional reform is necessary but not sufficient for blockade removal — a reading that weakens the institutional-fragmentation chain’s contribution to recurrence prevention.

The feedback loop that keeps wounding the regime

The crisis has a secondary self-reinforcing mechanism that operates independently of fuel supply. The loop runs: public anger → protests → government response → US-Cuba negotiations (via Raúl Guillermo Rodríguez Castro) → Rodríguez Castro interview → public anger returns. Each iteration narrows the government’s negotiation options while increasing external justification for maintaining the blockade. Even if fuel were restored, the political loop could independently sustain crisis dynamics.

Rodríguez Castro is the 42-year-old grandson of Raúl Castro. He holds no recognised institutional public function. He gave a USA Today interview displaying Hermès sneakers, a Rolex, and a Salvatore Ferragamo bag, saying, “It pains me that many people can’t live the way I do” and adding that he had no interest in politics but “if at some point the revolution needs me to step up, I will do it.” Michel Torres Corona — whose state-television program Con Filo was considered the epitome of Cuban government propaganda — wrote: “To usurp the functions of government, to assume a public role for which no one elected you, to proclaim yourself spokesperson for measures or new directions for the country … would anyone else be allowed to do that?” Academic Julio César Guanche described Rodríguez Castro as someone “without recognised institutional public functions.” Michael Bustamante, chair of Cuban and Cuban American studies at the University of Miami, called the interview “a cry for relevance” and noted “an open question as to who exactly he speaks for, and whether the channel of communication with him is ongoing or not.” Washington had been seeking “a Cuban equivalent of Venezuela’s Delcy Rodríguez” — a credentialed institutional counterpart the regime has not produced.

In the Mitchell-Agle-Wood stakeholder-salience framework, Rodríguez Castro classifies as Dangerous: medium power through family connections, very low legitimacy, medium urgency. The classification itself is a finding: the regime’s only publicly visible negotiation channel is a stakeholder without standing.

The cascading collapse of state capacity

The grid failure does not stop at darkness. Crime, once among the lowest in Latin America, is rising — fights on the streets, break-ins, violent muggings. Police, once ubiquitous, are difficult to find; victims report hours-long waits for response. Blackouts degrade surveillance and police coordination; economic desperation raises crime; police were already withdrawn with deteriorating response times. The state’s capacity to maintain order is degrading, which feeds a perception of state failure that accelerates the political cycle and reduces the government’s leverage in negotiations. The energy crisis feeds back on itself through a state-capacity path.

Prisoners Defenders reported 1,306 political prisoners as of July 9, 2026 — up from 1,260 in April and 1,281 at end of May. Detainees include Héctor Ochoa Vergara, held after participating in a peaceful demonstration against blackouts and water shortages in Ciego de Ávila, and Luis Manuel Otero Alcántara, en route to US exile after completing a five-year sentence but held in an unknown location while his visa was arranged. The rise correlates with deepening crisis; a protest-arrest pattern is plausible but the mechanism is not established in the source.

A Florida federal grand jury unsealed a May 20 superseding indictment of Raúl Castro — age 94 after correction against a June 3, 1931 birthdate; the Guardian’s “95” is an age error — on charges including conspiracy to kill US nationals, destruction of aircraft, and murder in connection with the 1996 Brothers to the Rescue shootdown. The indictment opens a Venezuela-style judicial-extraction pathway. It adds a legal dimension to the existing military-economic blockade, broadening the US pressure toolkit without requiring new executive action. The Florida federal court operates under an independent legal mandate the executive branch cannot control or easily rescind.

Who is at the table and who is not

The power-interest distribution across the crisis reveals a dangerous asymmetry. High power and high interest are concentrated in three actors: the Cuban government (existential interest in regime survival, control of state apparatus and military), the US government (sustained engagement through blockade, military, and legal system), and the Cuban diaspora (political constituency and legitimacy for US policy). High power but low interest sits with the Venezuelan post-Maduro regime — a critical oil supplier whose attention is on its own consolidation. Low power but high interest belongs to the Cuban public and urban protesters, political prisoners, the healthcare system, children and the elderly, and — on contested ground — Rodríguez Castro himself. Low power and low interest describes foreign companies that have already exited, the Cuban private sector awaiting unimplemented reforms, future generations, and regional neighbors whose stake remains latent.

The article’s frame carries systematic absences that matter for any complete assessment. Rural populations outside Havana — dependent on electricity for water pumps, hospital refrigeration, and agricultural cold chains — are unreported. The Chinese government and firms appear only as a logistics node; Beijing’s policy position is not reported. The Russian Federation, a historical energy patron, has no current engagement reported; a pattern of quiet re-engagement is not confirmed. The European Union has no current position. Afro-Cuban communities are not named despite carrying a disproportionate economic burden. Women and children are not named despite disproportionate water and heat vulnerability. Independent Cuban media are not referenced. The healthcare system — whose grid-dependence covers vaccine cold-chain storage, dialysis, surgical lighting, and neonatal care — is implied but not centred; each nationwide collapse carries patient fatalities that will never appear in a grid-failure statistic. Future Cuban generations have no representation in current negotiations. These gaps are structural to the source, not oversights in mapping: a boundary-critique step would ask whose perspective has been assumed in the reporting frame, and the answer is that it centres Havana, official actors, and visible protest over the invisible populations who bear the heaviest burden.

What remains uncertain

Four questions test whether the identified root conditions are the right ones. Whether the Rodríguez Castro negotiation channel is active or closed tests the institutional fragmentation root — if the channel is already dead, the fragmentation may be even more severe than the analysis assumes. Whether the 176 announced reforms gain implementation timelines and accountability mechanisms tests the state-monopoly procurement root — continued non-implementation would confirm that the announcement function is a policy signal, not a programme. Whether the reinforcing feedback loop can be broken without a settlement that addresses both fuel access and institutional credibility tests whether the political mechanism is truly independent of fuel supply — if the loop persists after fuel restoration, the regime’s political fragility is more fundamental than its energy dependency. Whether the pre-existing economic deterioration — pandemic-era hyperinflation, rising crime, withdrawing police — constitutes an independent variable that no fuel delivery or political settlement can reverse tests whether restoring power alone resolves the humanitarian crisis. The evidence does not yet answer these questions; it sets their terms.

Analytical techniques used in this piece

This analysis applies the methods below. Each links to a short, plain-English explainer you can read and reuse.

Relationship Mapping
Extracts the network of ties among people, institutions, and entities.
Root-Cause Analysis
Traces a symptom back along its causal chain to the conditions that actually generated it.
Stakeholder Mapping
Charts the parties to a situation — their interests, power, and alignments.