The Visoko Darvo Monthly: April 2026

The structural shifts of April and what they require of European decision-makers.

On what we chose and what we did not

The Monthly selects stories that reshape the architecture European decision-makers operate within. It does not select for the stories that dominated the news cycle. The cost in human life of the fighting in the Middle East, the earthquake on Japan's northeast coast, the asylum seekers lost in the Mediterranean, and the wider humanitarian toll of the month are real, and they are covered with greater authority and immediacy by institutions whose mandate is exactly that. Our mandate is different. Our task is the structural reading: what changed this month in the conditions under which European institutions, enterprises, and policymakers will decide? Where the news cycle and the structural shift overlap, as with the renegotiation of Western security architecture under conditions of war, we cover the structural shift. Where they do not, we step aside, and we trust the reader to read elsewhere.

Opening

April was the month three architectures came into view at once.

The first is the technology economy reverting to physical industry at the frontier. For fifteen years, the dominant logic of the sector was zero marginal cost in distribution. Software scaled without proportionate physical demand. AI inference does not. Every reasoning step performed by a frontier model carries a measurable cost in electricity, in cooling, in silicon, and in capital. Power, fabrication, and integrated hardware have become the layers on which competitive advantage is decided.

The second is the open renegotiation of Western security architecture. The Iran war, the closure and reopening of the Strait of Hormuz, the deployment of additional United States forces to the region, and the public threat by the United States President to withdraw from NATO and to pull forces from Germany and Spain have together moved European strategic autonomy from a discussion topic to an operating necessity. The European defense economy has, in the same month, moved from strategy to industry. The two shifts are connected.

The third is Bulgaria's own. The country is in its fourth month of eurozone membership. The Bulgarian National Bank Governor votes on the European Central Bank Governing Council. The currency board is retired, the monetary insulation that protected the country for three decades has been replaced with full integration into the single currency, and the structural conditions under which Bulgarian institutions and enterprises operate are no longer those of last year.

These three architectures are forming faster than the institutions designed to govern them. The work of the Monthly is to read the signals through this fact.

United States threatens NATO withdrawal during Iran war

Geostrategy & Economic Intelligence

The Iran war shaped the month in ways that extend well beyond the Middle East. The Strait of Hormuz was closed, then reopened under United States pressure, then partially restricted again. Energy markets absorbed the shock unevenly. Developing economies in Africa and South Asia faced compounding fuel and fertilizer costs. The Pentagon deployed additional forces to the region. And the President of the United States, in response to the level of European support he considered insufficient, openly threatened withdrawal from NATO and the removal of United States forces from Germany and Spain.

The threat itself may or may not be executed. The fact that it was made publicly, in the middle of an active war, with no operational caveat, has shifted the planning environment. The transatlantic security guarantee, which has been the unstated foundation of European defense planning since the post-war period, is now an explicitly conditional instrument. European capitals can no longer plan around it as a constant. They must plan around it as a variable.

In parallel, the political map within the Union shifted. The change of government in Hungary, where a long-incumbent leadership lost office to a center-right successor, removed the principal internal block on the disbursement of Union funds for Ukraine, on the implementation of the Defense Readiness Roadmap, and on the cohesion of the Eastern Flank position. The Visegrad dynamic that has shaped Central European policy for over a decade has been reset.

For Bulgaria, the implications are direct. Strategic autonomy is the operational requirement of every member-state government, including those that have spent years arguing against it. The capitals that update their defense, industrial, and energy planning to the new frame within the next two quarters will define the next decade. The cost of European defense is European. The institutions required to bear that cost must be built within this decade rather than the next generation.

The Defense Readiness Roadmap moves from announcement to delivery

defense & Security Systems

The most consequential European policy development of the month received the least international coverage. The Defense Readiness Roadmap 2030 has moved from announcement into funded delivery. Capability coalitions in nine priority areas have led to nations and operational mandates. The European Drone Defense Initiative and the Eastern Flank Watch are scaling toward initial capacity. The European Air Shield and Space Shield programs are entering activation. The Commission has proposed the AGILE program, a one hundred and fifteen million euro pilot instrument designed to fund disruptive defense innovation at speed and bring emerging technologies from laboratory to field within months rather than years, with operational launch expected in early 2027. And the European Defense Commissioner has openly called for a defense union grounded in a new intergovernmental treaty, alongside Norway, the United Kingdom, and Ukraine. No senior commission figure would have made that call twelve months ago.

European defense has stopped being a coordination problem and started being an industrial one. Joint procurement targets, defense-industrial-base requirements, reskilling commitments measured in hundreds of thousands of workers, and ramp-up overviews issued at the commission level are not the language of strategy. They are the language of execution. The Union has accepted that deterrence now depends on factories rather than on declarations and that those factories must operate within the Union's jurisdiction or within the jurisdictions of trusted partners.

Amazon's agreement to acquire the satellite operator Globalstar, valued near twelve billion United States dollars and intended to expand Amazon Leo's low-earth-orbit network, sits within the same frame. Sovereign communications infrastructure is now considered too strategic to leave to commercial market dynamics alone. The same logic that animates the European Drone Defense Initiative animates this transaction.

For Bulgaria, this is the most underestimated structural opportunity of the year. The country is an Eastern Flank member state. It carries an existing defense-industrial base in conventional munitions and small-arms manufacturing, accumulated over decades. It now operates under eurozone monetary protection, which lowers the cost of capital for industrial expansion. The Eastern Flank Watch is, by design, a distributed industrial architecture that requires production capacity within member states geographically positioned to support it.

Bulgaria's defense-industrial position is a strategic asset that the country has not yet articulated as one. The capacity exists. The geography is correct. The financing architecture is in place. What is missing is institutional alignment between the defense-industrial base, the export-control framework, the human-capital pipeline, and the Union-level capability coalitions. The states that integrate these four within the next eighteen months will define the Eastern Flank industrial map for the decade. Those that wait will be on it as consumers rather than as producers.

Beijing and Washington tighten enforcement against decoupling

Geostrategy & Economic Intelligence

Beijing passed a series of anti-decoupling measures designed to penalize foreign firms that relocate supply chains away from Chinese manufacturing at speed. The legislation reaches across capital repatriation, technology transfer during exit, and discretionary review of departing firms' continuing operations. In parallel, the United States Congress advanced the MATCH Act, which closes the gaps between United States and allied export controls and imposes country-wide prohibitions on the sale and servicing of the most advanced chipmaking tools, including ASML deep ultraviolet lithography systems, to China. The reports of Chinese electronic warfare components reaching Iranian forces, taken alongside the public Chinese position on the Strait of Hormuz, situate this commercial hardening within an explicitly geopolitical frame.

For most of the previous decade, decoupling was a declarative posture. The institutions, the legislation, and the enforcement apparatus required to give it effect were largely absent. That gap is now closing from both sides. Beijing is positioned to make exit costly. Washington is positioned to make circumvention difficult. The middle position, which has been the de facto European posture for several years, has become the most expensive of the three options.

For European multinationals operating through China and for the Bulgarian and regional firms integrated into their supply chains, this is a direct change in the operating environment. The cost of exit has risen. The cost of staying has not fallen. The cost of indecision has risen fastest of the three.

The Union's response under the Economic Security Strategy will determine whether European firms operate within a coherent decoupling framework or are left to navigate the United States and Chinese architectures independently. The first round of the Raw Materials Mechanism, which connects Union buyers with diversified suppliers of strategic raw materials, is the operational signal that Brussels intends to build that framework rather than improvise it. Regional decision-makers should plan for tightening, not loosening, of these constraints. The legislation passed this month does not reverse easily. It compounds.

IEA marks the beginning of the Age of Electricity

Energy & Grid Intelligence

The International Energy Agency has described the world as entering an Age of Electricity, marked by accelerating electricity demand, rapid solar deployment, and renewed nuclear investment. Electricity is now the binding constraint on technological capacity, replacing chips in that role. The bottleneck has moved to transformers, substations, interconnections, and the political authority to expedite their construction.

In response, both the Union and the United States have begun reclassifying AI data centers as critical national infrastructure, granting them fast-track energy permitting and sovereign-tier grid priority. This is a profound reordering of how grid capacity is allocated. The introduction of strategic compute infrastructure as a separate priority class alters the calculus for every other connected user.

The capital markets are pricing this shift. The financing round of three hundred and eighty million United States dollars secured by a developer of grid-scale modular nuclear reactors, intended for production in shipyards rather than in conventional construction sites, is one of several signals that modular nuclear has crossed the threshold from concept to industrial proposition. The energy supply curve for the next decade is being decided by capital allocations made now, not by capacity that comes online in fifteen years.

For Bulgaria, this reordering arrives under unusually favorable structural conditions. The country holds nuclear baseload at Kozloduy. Solar capacity is expanding. The position within the regional interconnection corridor is strategically central. Eurozone membership lowers the cost of long-duration capital required for transmission upgrades and data center construction. Several other member states combine the same physical assets, including France, Finland, and the Czech Republic. What distinguishes a credible Bulgarian proposition from theirs is not the assets. It is the speed of decision.

Bulgaria is in a position to articulate a compute-infrastructure proposition to the Union that rests on three commitments. Statutory time limits on grid-connection decisions at the transmission system operator, where the current process operates without binding deadlines and without a priority class for strategic infrastructure. Grid-priority designation for compute infrastructure embedded in national legislation rather than in ministerial discretion. Locational decisions are taken through a single national coordination function rather than across competing ministries. The countries that capture compute infrastructure in the next thirty-six months will be defined by the speed of their decision cycles, not the cost of their electricity. The structural advantage is current. The window is not.

Frontier technology reverts to physical economics

Innovation & Competitiveness

The most important shift of the month was conceptual. For fifteen years, the dominant explanatory framework for technology economics was the assumption of zero marginal cost. Distributing software was free. Scaling a platform required no proportional increase in physical resources. Three signals make clear that this framework no longer describes the frontier of the industry, even as it continues to describe the distribution layer.

The first is the announced succession at Apple, where the chief executive transition will pass to the head of hardware engineering. The selection of a hardware engineer signals where Apple believes its next decade of advantage will be located. Not in services revenue, but in custom silicon capable of running advanced models locally, on-device, without cloud dependency. The second is the SpaceX and Cursor transaction, in which a launch and satellite-network firm secured an option to acquire an AI coding platform, reflecting the strategic value of automating the software overhead of physical operations at scale. The vertical integration of physical and software industries is no longer theoretical. It is being capitalized through transaction structures valued as high as sixty billion United States dollars. The third is the launch of OpenAI's GPT-5.5 with native computer-use capability. Every multi-step agentic task carries a measurable energy and compute charge.

Together, these signals describe the frontier of the sector reverting to the economics of physical industry. Software distribution still scales at near-zero marginal cost. AI inference does not. The firms positioned to lead are those with vertical control over the layers below software: power, silicon, and integration. The firms exposed are those whose advantage rested on distribution alone.

For Union policy, the reversion clarifies the agenda. The continent's relative weakness in foundation models is in fact several distinct gaps: a foundation-model gap, a compute gap, a fabrication gap, and a power gap. They cannot be addressed in isolation. The Strategic Technologies for Europe Platform and the proposed Industrial Accelerator Act are the instruments of an industrial-strategy response. The question is whether they are governed coherently with energy strategy and digital strategy or whether they remain three frameworks operating in three directions.

For Bulgaria and the regional economies tied to the European industrial base, the operational implication is direct. Industrial policy, energy policy, and digital policy must now be the same document. Treating them as separate planning streams, as they have been for two decades, produces gaps that the new economics of the frontier expose immediately.

Humanoid robotics and agentic software cross thresholds

Robotics & Autonomous Systems

Two developments demonstrated that autonomous systems have crossed performance thresholds previously held by humans in both physical and cognitive domains. Sony AI's Project Ace, published in Nature, presented a table-tennis robotics platform capable of competing with professional players and defeating some high-level human opponents in a domain selected precisely because it requires extremely fast perception, prediction, and actuation. What the demonstration reveals is the latency frontier: robotic perception and response are approaching the time scales required to compete with elite humans in a contested physical environment.

In the same month, humanoid robots in Beijing completed a half-marathon distance in under two hours. The achievement is significant for a different reason than table tennis. Endurance robotics, as distinct from precision robotics, requires actuator durability, thermal management, energy density, and gait stability over extended periods. The combined demonstration of high-speed precision and long-duration endurance is the operational profile required for industrial deployment at scale. Manufacturing inspection, logistics handling, surgical assistance, and defense applications all depend on it.

In parallel, OpenAI's GPT-5.5 introduced native computer-use capability, strengthening the ability of software agents to perform multi-step office tasks autonomously. Google's Android 17 second beta continued the operating-system shift toward more privacy-preserving, context-aware, and developer-facing capabilities, though the move toward proactive agentic interfaces remains more visible as a broader platform direction than as a single beta feature. Software agents are moving from assistive to autonomous. The operating system is being reorganized around them.

For European institutions, the question is no longer whether to deploy autonomous systems. It is which systems, governed how, with what accountability. The governance gap is significant. Existing frameworks address data protection and content provenance, both important and both insufficient. The autonomy of an agent that takes multi-step actions across an institution's data stack raises a different category of question. Who is responsible when the agent acts in error? On what authority does it operate? What records does it leave? How is its behavior audited?

Bulgarian and regional enterprises adopting these systems in the next twelve months will be doing so largely in advance of clear governance standards. The institutional discipline required is internal. Defined authority limits for agentic systems. Mandatory logging. Human-in-the-loop checkpoints for consequential actions. Explicit contractual allocation of liability with vendors. The firms that codify these now will not need to retrofit them later. Retrofitting governance is more expensive than building it in.

TSMC confirms the durability of the AI silicon cycle

defense & Security Systems

TSMC's most recent earnings call confirmed the durability of the AI-driven silicon cycle. The foundry raised its full-year revenue guidance to above thirty percent growth, with high-performance computing now sixty-one percent of revenue, and identified the shift from generative to agentic AI as a structural multiplier of token consumption and computation. Within that aggregate strength, the center of gravity in AI hardware is moving from generalized capability toward specialized integration, with custom on-device silicon and workload-specific accelerators absorbing an increasing share of the value. The same logic drove the Apple succession decision and the SpaceX and Cursor transaction. Vertical integration is the form competition is now taking at every layer of the technology stack.

The reframing reaches European policy directly. The European Chips Act, in its current configuration, is oriented largely toward fabrication capacity at established process nodes. If the next decade's competitive advantage accrues to specialized, integrated silicon designed for particular software workloads, the policy framework requires extension. Fabrication alone is necessary but insufficient. Design capability, advanced packaging, and the integration between hardware and the models running on it are now part of the same strategic problem.

The MATCH Act's tightening of semiconductor manufacturing equipment controls runs through European supply chains as well as American ones. ASML's position as the principal supplier of the most advanced lithography equipment makes the Union an unavoidable participant in the enforcement architecture, regardless of formal alignment. European firms in the equipment supply chain will be required to demonstrate compliance with controls whose ultimate authority is American. This asymmetry is structural. It is unlikely to be renegotiated in the current political configuration.

For Bulgaria and the regional economies, the practical implication is sequencing. Sovereign capability in compute, in fabrication, and in the equipment that produces both is increasingly indivisible. Member states without fabrication capacity are not therefore excluded from strategic relevance. They are positioned at the design, packaging, and integration layers, where the next decade's value increasingly resides. IMEC's position in Leuven demonstrates what a union-relevant design and research anchor looks like at scale. The Strategic Technologies for Europe Platform now permits direct industrial-strategy financing of comparable institutions in the member states best placed to host them. Bulgaria carries an existing engineering-services and embedded-systems base, accumulated through two decades of work for the European automotive and industrial-electronics sectors. The strategic question is whether that base is treated as a peripheral export industry, as it has been, or as the foundation of a national design and integration proposition aligned with the Union's industrial accelerator framework.

WIPO documents diffusion compression

Education & Human Capital

The World Intellectual Property Organization's most recent report documented a pattern with implications well beyond the technology sector. The telegraph required around fifty years to achieve global reach. Generative AI reached users in nearly every country within days of becoming available online. WIPO described the phenomenon as diffusion compression. Its consequences are most acute in jurisdictions whose legislative cycles are longer than the diffusion curves of the technologies they are required to govern. For most of the world's economies, this is now a structural condition rather than a transitional difficulty.

The implication is institutional. The capacity of mid-sized economies to govern technologies is no longer primarily a function of their legal frameworks. It is a function of the speed at which their institutions can learn, adapt, and revise those frameworks. The standard tool of the policy cycle, analysis followed by consultation followed by legislation followed by implementation, operates on a timescale that the underlying technology has already left behind by the time the legislation enters force.

For Bulgaria, the response is not to legislate faster. Faster legislation produces brittle regulation. The response is to build standing institutional capacity for continuous adjustment. Regulatory bodies with permanent technical staff. Sandbox frameworks that allow controlled deployment ahead of full regulation. Structured channels through which front-line implementation experience updates the regulatory framework in real time. This is the work of decades, but the components are well understood.

The forecast on which this letter is willing to be judged is the following. Within twelve months, at least two of the four signals below will be present, or the structural advantage described in this issue will have been forfeited. First, a Union-significant compute infrastructure project sited in Bulgaria, with permitting velocity rather than electricity cost as the determining factor. Second, the inclusion of a Bulgarian defense-industrial firm in a Union-level capability coalition procurement under the Defense Readiness Roadmap. Third, the public designation of a Bulgarian institution by the Commission as a sandbox or pilot site for Article 57 implementation. Fourth, a measurable increase in private capital commitment to Bulgarian energy infrastructure over the prior twelve-month baseline, attributable to permitting reform rather than to monetary conditions. The Monthly will revisit these four signals at the close of the publication year.

What this means for the region

This assessment rests on practice accumulated across institutional design, the activation of intellectual capital, and the governance of strategic technology in jurisdictions including the Nordic states, Central Asia, the Western Hemisphere, and the European Union itself, and on the Center for Intellectual Capital, Innovation, and Competitiveness's. The pattern this issue describes, of structural assets squandered through institutional fragmentation, the Center has observed in each of those settings. The pattern of structural assets activated through deliberate alignment, the Center has also observed. The difference between the two outcomes has consistently been one variable: the presence or absence of a coordination function with named accountability and political authority sufficient to compel cross-ministerial action.

The structural shifts of April do not arrive separately. They are a single phenomenon, visible eight different ways. Each of the shifts described above is the collapse of an abstraction that allowed European institutions to defer decisions in its domain. The transatlantic security guarantee abstracted over American political will, now conditional. Software economics abstracted over physical resources, now costed. Global supply chains abstracted over geopolitical alignment, now enforced. Grid capacity abstracted over compute demand, now binding. Legislative cycles abstracted over diffusion speed, now compressed. The decisions European institutions deferred under each of those abstractions are now decisions with deadlines. That is the architecture.

For Bulgaria, the position is more advantageous than is generally recognized. The country sits within the Union's regulatory perimeter. It holds nuclear baseload capacity. It occupies a strategic position in regional energy interconnection. It carries an existing defense-industrial base. It operates under the protection of eurozone monetary architecture in its first year as a full member. These are structural assets. Their value depends on whether they are activated.

The decision the Monthly recommends to its readers for the next quarter is singular. The Council of Ministers should establish a Strategic Coordination Office under the direct authority of the Prime Minister, headed by a named coordinator with deputy-prime-minister rank, with a formal cross-ministerial mandate spanning the Ministry of Energy, the Ministry of Defense, the Ministry of Innovation and Growth, and the Ministry of Education and Science. Its operational scope is the alignment of four planning streams that today operate separately. Compute infrastructure planning, including grid priority and data center permitting. Defense-industrial integration with the Union-level capability coalitions. AI governance and agentic-system standards for the public and enterprise sectors. The development of the human capital required to operate at the boundary between technology, policy, and execution at the speed the diffusion curve now demands. They constitute one architecture, currently divided across ministries that do not coordinate. The decision is not to solve them. It is to make them one.

The objection to this recommendation is structural and fair. Bulgaria has tried coordination functions before. They have foundered on government turnover, on ministerial autonomy, and on the absence of statutory authority to compel decisions across ministries that resist them. The proposal here is different in three respects. The mandate is statutory rather than discretionary and survives changes of government. The authority is convening rather than executive, which makes ministerial cooperation easier to compel and harder to refuse. The reporting line is to the Prime Minister and to the National Assembly's relevant committee jointly, which raises the political cost of dismantling it. Whether these design choices are sufficient is an empirical question. The forecast above will answer it.

The structural shifts of April do not reverse. They compound. Institutions and enterprises that align with them now will compound with them. Those that delay will compound against them. The architecture is forming. The question for the region is whether to be among its designers or among its inhabitants.

References

Scroll to Top