Overview
This Future Insights report reviews five technology signals from Europe that have emerged since the previous Europe report in this series. The strongest pattern is that Europe is trying to turn digital ambition into institutional capability: it is simplifying parts of the AI Act without abandoning risk-based regulation, packaging sovereignty around chips, cloud, AI, open source and energy, discovering the hard infrastructure politics of AI data centres, using public finance to keep deep-tech companies inside the innovation system, and seeing renewed private capital for AI while still facing the gravitational pull of larger U.S. markets. For South Africa, these signals are useful not because Europe is a model to copy, but because Europe’s constraints reveal the practical questions that any serious technology strategy must answer: who can implement rules, who owns infrastructure, who funds long-horizon capability, who captures value, and how technological ambition becomes public legitimacy rather than policy theatre.
Five-signal overview
- The European Parliament approved AI Act simplification measures, extending high-risk AI deadlines while adding a targeted ban on non-consensual intimate-image and child-abuse-generating AI systems.
- The European Commission put forward a technology sovereignty package that links semiconductors, cloud, AI, open source and energy-system digitalisation into a single strategic autonomy agenda.
- Europe’s AI gigafactory plan is encountering delays and funding uncertainty, showing that sovereign compute depends on procurement timing, subsidy design, power, demand aggregation and private-sector confidence.
- The European Innovation Council selected 38 start-ups and SMEs for accelerator support, signalling continued European use of blended public finance to move deep-tech firms through the valley between invention and scale.
- European AI venture funding is rising sharply, but frontier-model capital, compute access and global expansion still reveal a scale gap against the United States.
Signal 1: Europe is slowing parts of AI Act implementation while hardening specific harm protections
What happened
On 16 June 2026, the European Parliament gave final approval to amendments to parts of the EU AI Act as part of the digital omnibus package. The changes postpone obligations for stand-alone high-risk AI systems to 2 December 2027 and for AI systems embedded as safety components in regulated products to 2 August 2028. They also delay watermarking obligations for AI-generated content to 2 December 2026, clarify overlap with machinery rules, extend some exemptions to small mid-cap enterprises, centralise certain general-purpose AI enforcement in the AI Office, and ban AI systems that generate child sexual abuse material or non-consensual intimate sexual images, video or audio, subject to formal Council adoption before entry into force (European Parliament, 2026).
Why it matters
The signal is not simply deregulation. It is a more complex institutional moment in which Europe is trying to preserve the architecture of risk-based AI governance while admitting that implementation capacity, standards, business readiness and enforcement machinery must mature before the law can work as intended. The politically important move is that Europe is distinguishing between compliance sequencing and substantive protection: some high-risk obligations are delayed, but a specific class of intimate-image and child-abuse harms is prohibited faster. That distinction matters for South Africa because AI governance debates often polarise between maximal regulation and permissive experimentation. Europe’s adjustment suggests a third problem: the credibility of governance depends on matching legal ambition to administrative timing, technical standards, institutional authority and the harms that society is least willing to defer.
What it could mean
For South Africa, the European move points toward staged AI governance that is neither a copy of the EU framework nor a vague set of ethical aspirations. A credible South African approach would need to separate urgent prohibitions, public-sector deployment rules, sectoral obligations, procurement standards, model documentation expectations and longer-term assurance capacity. The key lesson is that the timing of obligations can be a design feature rather than a sign of weakness, provided the state uses the time to build enforceable standards, institutional competence and predictable guidance. If South Africa waits for perfect capacity before regulating, public harms may accumulate. If it legislates beyond capacity, rules may become symbolic and favour larger firms that can absorb compliance costs.
Possible futures
Possible future A: South Africa builds a sequenced AI governance regime with early harm prohibitions
South Africa could use Europe’s recalibration to design a staged AI governance framework that starts with harms where constitutional, gender-based violence, child protection and public-law concerns are already clear. Non-consensual intimate-image generation, impersonation for fraud, manipulative systems aimed at children, discriminatory public-service scoring and opaque automated decisions in high-stakes domains could be addressed before the country attempts a comprehensive AI Act equivalent. The strategic advantage would be institutional credibility: regulators would show that AI governance is not abstract futurism, but a targeted response to harms citizens can recognise. The constraint is that early prohibitions must still be technically enforceable. Without investigative capacity, platform cooperation, forensic capability and clear liability rules, a ban can function mainly as symbolism. The deeper opportunity is to connect AI policy to existing constitutional rights and sectoral regulators, allowing governance to grow through practical enforcement rather than one large statute that outruns the state.
Possible future B: South Africa imports European language but reproduces implementation fragility
Another trajectory is that South Africa borrows the vocabulary of risk-based AI, transparency, human oversight and high-risk systems without building the standards bodies, technical audit capability, procurement expertise and administrative routines that make such language operational. In this future, legislation or policy would appear sophisticated, but firms and public agencies would face uncertainty about what compliance actually requires. The likely result would be uneven enforcement, legal ambiguity and a compliance market dominated by large consultants and multinational vendors. Smaller South African firms could be discouraged from experimenting because they cannot interpret the rules, while powerful actors continue to deploy systems with limited scrutiny. The second-order risk is legitimacy erosion: citizens may learn that AI governance promises rights protection but fails in concrete cases. Europe’s delay would then have been misread. The lesson is not that complexity should be copied slowly; it is that governance architecture must be built around the administrative capacities that actually exist and can be improved.
Possible future C: South Africa positions itself as a capacity-aware AI governance convenor for emerging markets
A more ambitious future is that South Africa turns Europe’s implementation difficulties into a diplomatic and policy opening. It could convene African regulators, standards bodies, universities, civil-society organisations and industry to develop AI governance tools that are rights-based, development-aware and administratively realistic. Such a model would not reject European principles, but would adapt them to contexts where public-sector digitisation, language diversity, informal economies, data scarcity and enforcement constraints differ sharply from those of the EU. The strategic payoff could be regional influence: South Africa might help shape procurement templates, risk-assessment methods, public-sector AI registers, incident-reporting norms and protections against synthetic media abuse across the continent. The trade-off is that leadership without domestic proof would be thin. To be credible, South Africa would need demonstrable reforms at home, especially in public procurement, data stewardship and sectoral supervision. If achieved, it could make the country a governance translator rather than a passive rule-taker.
Signal 2: Europe is consolidating technology sovereignty into a wider industrial policy frame
What happened
On 3 June 2026, the European Commission announced a European technological sovereignty package focused on semiconductors, artificial intelligence, cloud, open source and digitalised energy systems. The package includes a proposed Chips Act 2.0 to support advanced semiconductor capacity, a Cloud and AI Development Act to encourage sustainable cloud technologies and streamline data-centre deployment, an open-source strategy to scale alternatives in priority areas and public administrations, and a strategic roadmap for digitalisation and AI in the energy sector so that data centres and AI systems are integrated into Europe’s energy system (European Commission, 2026).
Why it matters
This is important because Europe is treating technology sovereignty less as a slogan and more as a coordination problem across industrial policy, infrastructure, public procurement, energy systems, software ecosystems and geopolitical dependency. The package implicitly recognises that AI capability cannot be separated from chips, cloud, energy, standards, public-sector demand and developer ecosystems. It also signals that open source is no longer only a software culture issue; it is being framed as a sovereignty instrument. For South Africa, the lesson is that digital autonomy is not achieved by declaring independence from foreign platforms. It requires a disciplined inventory of dependencies, a realistic view of which layers can be influenced domestically or regionally, and procurement choices that build capability instead of simply buying services.
What it could mean
South Africa cannot replicate Europe’s fiscal scale or institutional density, but it can learn from the architecture of the sovereignty package. The practical question is where South Africa needs minimum sovereign capability: public digital identity, sensitive government data, cybersecurity, payments and financial infrastructure, language technologies, geospatial data, energy-system analytics, health and education platforms, and procurement rules that avoid avoidable lock-in. Europe’s package suggests that sovereignty should be selective, not totalising. South Africa’s strongest path may be to define a small number of strategic digital capabilities that support state capacity, economic competitiveness and democratic accountability, while openly using global platforms where dependence is manageable and bargaining power can be preserved.
Possible futures
Possible future A: South Africa adopts selective digital sovereignty anchored in procurement
In this future, South Africa develops a practical digital sovereignty doctrine that is enforced through procurement rather than only strategy documents. Public agencies would classify systems by sensitivity, require interoperability and exit rights in major technology contracts, prefer open standards, assess cloud jurisdictional risks, and build shared platforms for identity, payments, data exchange and cybersecurity. The goal would not be domestic self-sufficiency. It would be to prevent critical public functions from becoming technically opaque, contractually trapped or geopolitically exposed. The second-order benefit could be market creation: if procurement clearly rewards open interfaces, local support capacity and secure hosting options, South African firms would have more predictable opportunities to build around public needs. The constraint is political economy. Vendor lock-in often persists because it serves bureaucratic convenience, private interests or short-term risk avoidance. A sovereignty doctrine would only matter if Treasury, regulators and departments translate it into contract clauses, budget choices and accountability mechanisms.
Possible future B: Sovereignty language becomes a cover for inefficient localisation
There is also a risk that South Africa adopts the rhetoric of digital sovereignty in a way that produces expensive symbolic localisation rather than genuine capability. Government could favour nominally local vendors without requiring security, interoperability, performance or skills transfer. Cloud localisation could become a box-checking exercise, open source could be invoked without maintenance capacity, and “local AI” could mean reselling foreign models through domestic intermediaries. This future would be damaging because it would discredit sovereignty as protectionism while leaving real dependencies intact. Citizens and businesses would pay for weaker systems, and public institutions would still lack the internal knowledge needed to govern technology. Europe’s package is useful precisely because it links sovereignty to concrete layers: semiconductors, cloud, software, energy and AI infrastructure. South Africa would need the same discipline. The test is whether sovereignty reduces strategic vulnerability and grows domestic competence, not whether it produces politically attractive procurement announcements.
Possible future C: South Africa uses European sovereignty policy to negotiate better partnerships
A more constructive future is that South Africa treats Europe’s sovereignty agenda as a basis for negotiated partnership rather than imitation. European institutions seeking trusted partners in digital infrastructure, cybersecurity, critical minerals, energy-system analytics, open-source public tools and AI governance may have reason to work with South Africa. South Africa, in turn, could use such partnerships to demand local skills formation, shared intellectual property where appropriate, research collaboration, regional deployment rights and transparent public-interest safeguards. The strategic logic is that Europe’s search for autonomy from dominant non-European providers may open space for countries that can offer credible institutions, regional access and domain knowledge. The risk is asymmetry: partnerships can reproduce dependency if technology, decision rights and value capture remain offshore. The difference between partnership and outsourcing would depend on whether South African universities, firms and public agencies become co-producers of capability rather than implementation sites for externally designed systems.
Signal 3: Europe’s AI gigafactory ambitions are colliding with infrastructure and funding realities
What happened
A Bloomberg report carried by the Financial Post on 12 June 2026 said the European Union’s €20 billion plan for large AI data centres, or gigafactories, was facing delays and funding uncertainty. The report said the bidding process, previously expected in May, was now expected in July, that funding would be phased around 2028 and 2030, and that limited near-term funding could support only part of the envisioned plan. It also noted that Europe’s public effort is occurring alongside much larger private data-centre commitments, including SoftBank’s announced plan to invest as much as €75 billion in French projects (Financial Post, 2026).
Why it matters
This signal shows that sovereign compute is an infrastructure programme, not merely an innovation slogan. AI gigafactories require subsidy design, demand aggregation, GPU access, power contracts, grid connections, cooling, land, permitting, private co-investment and credible timelines. Europe’s difficulty is therefore analytically useful: even a wealthy bloc with strong institutions struggles to convert political urgency into buildable compute infrastructure. For South Africa, this matters because AI infrastructure debates can easily skip over the physical and financial constraints. If a region with Europe’s resources faces delays, South Africa must be especially careful about how it frames AI data centres, local compute, renewable energy, municipal capacity, water use and public benefit.
What it could mean
South Africa’s AI infrastructure choices will probably be selective. The country may not need or be able to fund frontier-scale compute, but it will need reliable cloud access, some local or regional high-performance computing capacity, secure public-sector hosting, energy-aware data-centre policy and research infrastructure. Europe’s gigafactory delays suggest that South Africa should not announce compute ambitions without a bankable model for demand, power, governance and skills. The country’s comparative advantage could lie in linking renewable energy, regional cloud demand, research computing, financial-services workloads and public-sector data needs in a smaller but more coherent infrastructure strategy.
Possible futures
Possible future A: South Africa builds an energy-aware compute strategy at regional scale
South Africa could respond to the European signal by designing a compute strategy that is deliberately smaller, energy-aware and regionally oriented. Instead of attempting a prestige AI megaproject, it could align renewable energy zones, fibre routes, university high-performance computing, secure public cloud requirements and regional enterprise demand. The goal would be to create a credible platform for applied AI, research, public-sector analytics and African-language technologies without pretending to compete with hyperscale frontier-model clusters. The advantage is strategic realism: South Africa could build enough capability to reduce dependency in sensitive domains while avoiding ruinous capital expenditure. The trade-off is that such a strategy requires coordination among energy regulators, Eskom, municipalities, cloud providers, universities, Treasury and industrial-policy actors. If those institutions can define stable demand and transparent access rules, South Africa could become a serious regional compute node. If coordination fails, the plan would dissolve into isolated facilities with little cumulative capability.
Possible future B: Private data centres grow faster than public governance
Another plausible future is that data-centre investment continues in South Africa primarily through private deals, embedded generation, wheeling arrangements and hyperscaler expansion, while public governance remains reactive. This could improve capacity for firms that can pay, but it may also deepen uneven access to reliable digital infrastructure. High-value customers would receive secure cloud and AI services, while public agencies, smaller firms, universities and municipalities remain dependent on fragmented systems and foreign-hosted tools. The second-order issue is legitimacy. If energy-intensive digital infrastructure is perceived as serving elite markets while the broader electricity system remains constrained, data centres may become politically vulnerable. Water use, land allocation, grid access and tax incentives could provoke contestation. Europe’s experience shows that compute infrastructure needs public confidence as much as capital. For South Africa, that means data-centre policy should require visible public benefits, such as added clean generation, skills pipelines, local procurement, research access or support for critical public services.
Possible future C: South Africa stays a compute consumer and loses strategic leverage
A more passive future is that South Africa does not build meaningful local or regional compute capability and remains largely a consumer of AI services hosted elsewhere. This would not stop AI adoption; South African banks, retailers, mines, law firms, media organisations and public agencies would continue to buy access to global platforms. But the terms of participation would be weaker. Data governance would be more dependent on foreign legal jurisdictions, latency-sensitive applications would be constrained, researchers and start-ups would face higher costs, and national bargaining power with cloud providers would remain limited. The hidden cost is institutional learning: countries that host and manage infrastructure develop expertise in operations, security, procurement and regulation, while countries that only consume services learn less about the underlying stack. Europe’s gigafactory struggle illustrates that compute sovereignty is hard, but difficulty does not make the issue irrelevant. For South Africa, the danger is mistaking affordability today for strategic resilience tomorrow.
Signal 4: Europe is using public finance to move deep-tech firms from invention toward scale
What happened
On 15 June 2026, the European Innovation Council announced that 38 start-ups and SMEs had been selected for support in the latest EIC Accelerator round. The selected firms, drawn from 16 EU and associated countries, will receive a proposed €90 million in grant funding, with 84% eligible for blended finance combining grants and equity. The EIC highlighted examples including automated crop genetics, one-minute solid-state battery charging, cooperative hybrid-kinematic robots, ultra-high-speed wireless connectivity and enhanced-rock-weathering carbon capture, with equity investments to follow through the EIC Fund depending on company strategy (European Innovation Council, 2026).
Why it matters
The signal is that Europe continues to use public finance as an instrument for deep-tech market formation. Deep-tech firms often face long development cycles, technical risk, uncertain customers, capital intensity and weak early revenue. Ordinary venture finance alone may be insufficient, especially outside the United States. The EIC model uses grants, equity and signalling effects to help firms cross the difficult zone between laboratory proof, prototype, early customers and private capital. For South Africa, this matters because the country has research strengths, industrial problems and entrepreneurial talent, but often lacks patient capital and procurement pathways that turn science into scalable firms. Public innovation funding is not automatically effective, but without it many strategic technologies remain under-commercialised.
What it could mean
South Africa could read the EIC signal as a prompt to improve the connection between research funding, industrial policy, public procurement and scale capital. The country does not need to match Europe’s funding volumes, but it needs better instruments for technologies relevant to local constraints: grid analytics, mining automation, agricultural resilience, water management, health diagnostics, cybersecurity, language AI, advanced materials and climate adaptation. The most important question is institutional design. Funding must select for technical seriousness and commercial potential without becoming patronage. It must also be connected to customers, regulatory pathways and follow-on capital, or grants will produce isolated prototypes rather than companies.
Possible futures
Possible future A: South Africa creates mission-linked deep-tech finance with real customers
South Africa could build a more coherent deep-tech finance instrument around national missions rather than generic start-up support. For example, public funding could target grid stability, mine safety, water security, agricultural productivity, public-health diagnostics, language access and cybersecurity, with departments, state-owned enterprises, municipalities and large firms acting as reference customers. Grants would finance technical risk, while matched equity or revenue contracts would test market discipline. The strategic gain would be that innovation support becomes anchored in problems South Africa actually needs to solve. This could reduce the common gap between university research and commercial adoption. The constraint is governance: mission-linked finance can be captured by incumbents, politically connected firms or fashionable themes unless selection is rigorous and transparent. The institutional implication is that South Africa needs not only funding, but evaluators who understand technology, markets and public value. If built well, such a mechanism could keep more deep-tech value inside the country.
Possible future B: Innovation funding remains fragmented and prototype-heavy
In a weaker trajectory, South Africa continues to fund innovation through fragmented grants, competitions and pilot programmes that reward proposal-writing more than scale. Universities produce research outputs, start-ups win small awards, incubators report activity, and agencies publish success stories, but relatively few firms reach procurement, export markets or follow-on capital. This future is plausible because it avoids hard choices about concentration, failure and accountability. Deep-tech finance requires accepting that some projects will fail and that others need larger, longer support. Fragmentation feels fairer politically but can starve serious companies of the resources they need. Europe’s EIC model is imperfect, yet it recognises that deep-tech firms often require blended instruments and institutional endorsement. South Africa’s risk is that it spreads limited money too thinly, leaving promising technologies trapped between proof-of-concept and industrial use. The result would be persistent dependence on imported technologies even where domestic research capability exists.
Possible future C: South African firms use European deep-tech finance as a scaling bridge
A third future is that South African researchers and founders increasingly connect to European accelerator, Horizon, EIC-adjacent or corporate innovation channels to scale technologies that have African relevance but need larger capital pools. This could be beneficial if it gives South African companies access to customers, regulatory validation, technical partners and patient finance. Technologies in climate adaptation, agriculture, energy, mining safety and health diagnostics could travel well because they address problems that are globally salient but locally acute. The risk is that the centre of gravity moves offshore: intellectual property, headquarters, senior talent and manufacturing may relocate to access funding. The strategic challenge is to design partnerships that keep meaningful capability in South Africa while using European capital intelligently. That might require local co-investment, university IP policies that enable domestic growth, regional customer pipelines and government support for firms entering European programmes. Done well, Europe becomes a bridge; done poorly, it becomes an exit route.
Signal 5: European AI capital is growing, but scale still pulls toward larger global markets
What happened
On 12 June 2026, Crunchbase reported that roughly half of European venture funding in 2026 to date had gone to AI-related companies, including frontier model companies, data-centre firms, semiconductor ventures, robotics, aerospace, defence, biotech and AI applications across sectors. It said European start-up funding exceeded $17 billion in each of the two most recent quarters, while three new frontier labs, Recursive Superintelligence, Ineffable Intelligence and Advanced Machine Intelligence, had collectively raised $2.6 billion in 2026. The report also noted that European foundation labs have raised more than $8 billion since 2021, far less than leading U.S. frontier-model companies (Crunchbase News, 2026).
Why it matters
This signal is analytically useful because it is both positive and sobering. Europe has talent, capital, founders, research heritage and renewed AI momentum, yet the scale of U.S. frontier funding and compute remains difficult to match. The implication is that AI ecosystems are shaped by cumulative advantages: capital depth, infrastructure, anchor customers, talent density, regulatory confidence, acquisition markets and global ambition. For South Africa, the lesson is not that AI entrepreneurship is futile. It is that ecosystem strategy must be honest about where the country can compete. Building frontier general-purpose models may be less plausible than building AI firms around domain depth, African data, multilingual services, energy constraints, financial inclusion, mining operations, education, health and public administration.
What it could mean
South Africa’s AI opportunity is likely to be application-rich rather than frontier-model-led. The country can build firms that use global models, fine-tune domain systems, create local datasets, design governance tools and solve problems that imported platforms do not understand well. But doing so requires more than talent. It requires patient capital, procurement demand, research-industry bridges, compute access, trusted data arrangements and routes into African and global markets. Europe’s AI funding momentum suggests that even wealthy ecosystems must work hard to retain ambition. South Africa therefore needs deliberate mechanisms to keep founders from treating the local market only as a training ground before moving intellectual property and decision-making elsewhere.
Possible futures
Possible future A: South Africa builds AI companies around difficult local domains
South Africa could develop a distinctive AI ecosystem by concentrating on domains where local operational complexity is a source of advantage rather than embarrassment. Electricity forecasting, municipal maintenance, mine safety, logistics under infrastructure constraints, multilingual customer service, credit risk, fraud detection, legal access, public-health triage and education support are not peripheral use cases; they are demanding environments where generic tools often fail. Firms that solve these problems in South Africa may be able to scale across other emerging markets facing similar constraints. The strategic logic is that domain depth can partly substitute for frontier compute. The trade-off is that such firms need patient customers willing to co-develop products, not only venture investors seeking quick software margins. Public procurement and large corporate demand would therefore be central. If South Africa can turn its hardest institutional and infrastructure problems into reference markets, AI entrepreneurship could become developmental rather than merely extractive.
Possible future B: South African AI talent externalises faster than institutions can absorb it
A less favourable path is that South Africa produces capable AI engineers, researchers and founders, but the most ambitious projects move offshore because local capital, compute and customers are insufficient. Individuals would still succeed: they may work for global labs, join foreign start-ups or build companies registered elsewhere. The national ecosystem, however, would lose compounding benefits. Intellectual property, managerial experience, investor networks and advanced product knowledge would accumulate outside South African institutions. Europe’s anxiety about the Bay Area’s pull becomes even sharper in this context. If European founders with access to large markets still look abroad for scale, South African founders will face stronger incentives to leave unless local and regional opportunities are credible. The policy implication is not to trap talent, which would be impossible and undesirable, but to create enough local demand, research funding, data access and co-investment that staying connected to South Africa remains strategically rational.
Possible future C: South Africa becomes a bridge between European AI capital and African problem markets
An optimistic but demanding future is that South Africa positions itself as a bridge between European AI investors and African markets. European capital and companies may seek differentiated growth areas beyond saturated U.S. competition, while South Africa offers financial infrastructure, universities, legal sophistication, corporate customers and access to regional problems in energy, logistics, agriculture, health, mining and public services. South African AI firms could partner with European investors without relocating their entire centre of gravity, using Europe for capital, standards validation and market access while retaining African domain ownership. The risk is that partnerships become asymmetric: European firms capture intellectual property and South African teams become implementation subsidiaries. Avoiding that requires strong founder bargaining power, local co-investment, clear data-rights arrangements and regional customer pipelines. If those conditions are met, Europe’s AI capital momentum could widen South Africa’s opportunity space rather than intensify talent extraction.
Conclusion
Europe’s latest technology signals show a region trying to turn policy ambition into practical capability under pressure. The AI Act amendments reveal the tension between legal architecture and implementation capacity. The sovereignty package shows how chips, cloud, AI, open source and energy are being pulled into one strategic frame. The gigafactory delays expose the infrastructural reality beneath AI ambition. The EIC funding round demonstrates the continuing role of public finance in deep-tech scale-up. The AI venture funding signal shows that momentum can coexist with structural dependence on larger capital and compute ecosystems.
For South Africa, the central lesson is strategic selectivity. The country should not imitate Europe’s scale, but it should study Europe’s institutional questions. Which AI harms require early prohibition? Which digital systems must remain governable by the state? Which infrastructure dependencies are tolerable, and which create long-term vulnerability? Which deep-tech firms need patient finance because markets alone will underinvest? Which AI opportunities fit South Africa’s domain realities rather than imported hype? The answer is not a single grand technology plan. It is a disciplined portfolio of governance, procurement, finance, infrastructure and partnership choices that allow technology to become capability. Europe’s current experience is valuable because it makes the hard parts visible: sequencing, coordination, legitimacy, funding and the translation of ambition into institutions that work.
References
Crunchbase News. (2026, June 12). European AI funding is growing. Will that boost the region’s startup scene? Crunchbase.
European Commission. (2026, June 3). Strengthening Europe’s tech sovereignty.
European Innovation Council. (2026, June 15). 38 start-ups and SMEs secure EIC support in latest round of the EIC Accelerator.
European Parliament. (2026, June 16). AI Act: EP approves simplification measures and “nudifier” app ban.
Financial Post. (2026, June 12). EU’s AI data center plans stumble due to delays, funding.
Publication links (website version)
European Parliament: AI Act simplification measures and nudifier app ban
European Commission: Strengthening Europe’s tech sovereignty
Financial Post: EU’s AI data center plans stumble due to delays, funding
European Innovation Council: 38 start-ups and SMEs secure EIC support
Crunchbase News: European AI funding is growing