AI Trends 2026: Market Concentration, DeepSeek, and European Digital Sovereignty

Executive Summary

The FAZ Digital Economy Podcast analyzes the most important technology trends for 2026. The focus is on valuation bubbles in the AI sector despite impressive investments, surprising competition from DeepSeek from China, and the question of European digital sovereignty. The moderators discuss extreme market concentration among tech giants, social media regulation for teenagers, and the need for European AI infrastructure.

People

Topics

Detailed Summary

AI Market: Between Boom and Bubble

The FAZ Digital Economy Podcast begins 2026 with a critical analysis of AI development. The discussion shows 2025 as marked by massive investments: the Stargate announcement with 500 billion US dollars for AI infrastructure, followed by regular billion-dollar investments. At the same time, market sentiment has shifted – initial euphoria is increasingly giving way to skepticism about potential valuation bubbles.

Particularly remarkable is extreme market concentration: towards the end of the year, over 40 percent of the S&P 500's market value was concentrated in the ten leading companies – significantly higher than in previous years (20-25 percent) and even during the dotcom bubble. Seven to eight of these top-ten companies depend heavily on AI success. Nvidia, with a market value of about 4.5 trillion dollars, is in a unique position as the leading chip supplier for high-performance AI applications, which raises questions about the disconnect between stock price and fundamental business facts.

The circular deal structures – such as Nvidia's stake acquisitions in OpenAI and Anthropic, or OpenAI's 300-billion-dollar investment commitment to Oracle – show a tightly interwoven system vulnerable to chain reactions. The Oracle episode illustrates this clearly: the stock jumped dramatically after the announcement but fell again when questions about OpenAI's financing ability emerged.

DeepSeek Shocks the Market

A turning point in the year was the release of DeepSeek's R1 model – a Chinese company that was barely known until then. The model achieved surprisingly good results with a fraction of the resources that American providers invest. This triggered market panic: in one day, tech stocks lost up to a billion euros in market value. The fundamental question "Why invest so massively if it apparently works with significantly less?" persistently unsettled investors.

Meanwhile, the narrative has calmed down. DeepSeek is available as open-source and has become particularly popular in the open-source community, where German technology consultancies like TNG develop variants based on DeepSeek that reduce Chinese censorship aspects. The race between China and the US in AI development is definitely in full swing.

Europe's Role and Digital Sovereignty

While China and the US dominate, Europe's position is rather weak. Germany has not yet produced any significant foundational models. However, there are positive developments in the application sector: Swedish Lovable, German N8N, Valor, and Langdog – companies creating AI applications, not developing models – are growing successfully. Interestingly, Berlin is experiencing a comeback as a startup hotspot for AI applications.

The question of European sovereignty is being addressed by the Schwarzgruppe: with its Stackit cloud platform and an 11-billion-euro investment project for a data center in Lübbenau, the discounter company is positioning itself as a European AI and cloud player. The company is deliberately building a German-European ecosystem – investing in Aleph Alpha, other AI companies, and cybersecurity solutions.

The multicloud strategy of large European companies shows pragmatic thinking: not complete migration away from Microsoft Azure, but diversification with European providers for sensitive data, while non-critical data remains with American cloud providers. This is technically feasible and value-creation oriented.

Google and OpenAI: Competition Instead of Monopoly

Contrary to early concerns following ChatGPT, genuine competition has established itself. Google, initially surprised, responded impressively with Gemini 3. The company had the technological foundations (DeepMind, AI expertise) for a long time but long struggled with market penetration. An important difference: established companies like Google, Facebook, or Microsoft could not publicly release an error-prone model like the original ChatGPT – the reputational damage would have been too great. OpenAI as a startup could take that path.

Also fascinating is the high acceptance speed: millions of people intensively use ChatGPT and other AI tools to increase personal productivity. This even surprised OpenAI itself.

Search Engines and AI Integration

The future of search will be hybrid: not complete replacement of classic search engines by AI search, but integration of AI-generated summaries alongside classic search results, as Google is implementing. However, younger generations show changed behavior – instead of Google, 17- and 18-year-olds prefer YouTube, TikTok, or AI search engines. This trend will aggregate into market share changes.

Social Media and Youth Protection

A critical topic for 2026 is the regulation of social media for teenagers. The negative psychological effects are scientifically proven. Countries like Australia are introducing complete bans under 16 years. The EU and other countries are considering age restrictions.

Realistic mid-term solution: Effective age verification rather than complete bans, to preserve digital participation for teenagers. Currently, age verification is a joke – a 12-year-old can easily claim to be 14. Tech companies delay implementation for cost reasons and because fewer users mean less advertising revenue. However, political pressure will increase.

The neurological consequences are significant: neuroscientist Martin Korte points out that intensive social media use not only shortens attention spans but changes the physical brain structure in developing people. Importantly: social media also offers real benefits (spontaneous meetups, quick communication), but the business model – endless short videos, algorithmic optimization for dwell time – is harmful.

TikTok Deal and Geopolitical Implications

The long saga over TikTok and possible bans/takeovers is nearing its end. A deal with Oracle, Silver Lake, and MGX (Abu Dhabi sovereign wealth fund) is taking shape. Interesting: the plan to protect TikTok from Chinese control leads to involvement of a sovereign wealth fund from Abu Dhabi.

Geopolitically relevant: Trump benefited in 2024 from TikTok-mobilized voters, which explains his erratic course. The US wants to maintain control over critical communication platforms – not only because of espionage concerns but also to secure spheres of influence. Europe could soon come under pressure to use the "better" American version of TikTok.

Meta has held its ground well – advertising business is flourishing, antitrust proceedings have been won, Threads (Twitter clone) is performing very well in the US. Individual missteps (Metaverse) cost nothing in profits.

Key Takeaways

• Market concentration is extraordinarily high: 40+ percent of S&P 500 market value lies with ten companies, 7-8 of which depend massively on AI success – higher than in the dotcom bubble.

• Valuations are disconnected: Stock increases at Nvidia (4.5 trillion dollar market value) and others still need to be justified by genuine productivity gains or business growth – this is questionable.

• DeepSeek proved efficiency: A Chinese model showed that AI breakthroughs are possible with significantly fewer resources – intensified market skepticism, but technology competition between the US and China intensifies.

• Europe must take a different approach: Not competitive in foundational models, but growing ecosystem in AI applications (Lovable, N8N, Langdog) and with Schwarzgruppe/Stackit in European cloud infrastructure.

• Digital sovereignty needs multicloud strategies: Complete independence is unrealistic, but diversification with European providers for sensitive data is feasible and necessary.

• Google remains competitive: Not the monopoly fate predicted after ChatGPT – Gemini 3 shows a strong counterattack, genuine competition is established.

• Search behavior changes among youth: 17-18-year-olds use classic Google search less, more YouTube/TikTok/AI search – the trend will show in market shares.

• Youth protection will be regulated: Effective age verification instead of complete bans will become the norm – political pressure grows due to scientifically proven psychological damage.

• Brain development is at risk: Intensive social media use changes physical brain structure in developing people, not just attention.

• TikTok deal is nearing its end: US takeover likely – geopolitical implication: US secures control over critical communication platform.

• Meta maintains dominance: Advertising business flourishes, antitrust proceedings won, Threads successful – individual missteps (Metaverse) play no role in profits.

• Three priorities for 2026: (1) Monitor hybrid threats/disinformation campaigns, (2) observe whether large industrial companies scale back AI investments (market correction signal), (3) continue educating yourself and understand as parents what children do online.

Metadata