AI rules fragment markets
The EU’s AI Act is set to impose strict compliance obligations this year — including fines up to €35 million or 7% of global turnover — raising real enforcement risk for companies using AI in standard software. At the same time Russia is moving to ban foreign tools like ChatGPT/Gemini and regional bodies are urging telcos to localize governance, a split that’s already fueling a surge in demand for AI ethics and governance platforms (projected from $2B in 2025 to $45B by 2035). (xpert.digital) (firstpost.com) (openpr.com)
The EU AI Act’s staggered timetable makes enforcement imminent: the regulation entered into force in 2024 and the Act’s general date of application for most obligations is 2 August 2026. (europarl.europa.eu) Key earlier milestones are already binding — certain prohibited practices and transparency duties took effect on 2 February 2025, while obligations for general‑purpose AI (GPAI) models began applying on 2 August 2025. (artificialintelligenceact.eu) Enforcement will be handled by a hybrid architecture that places the European AI Office alongside national market‑surveillance authorities, and the Commission published a draft implementing regulation on 12 March 2026 that opened a public consultation running to 9 April 2026. (digital-strategy.ec.europa.eu) (europesays.com) Providers of high‑risk systems must complete formal conformity assessments, install a certified quality‑management system, keep technical documentation and logs, issue an EU declaration of conformity and affix CE marking before market placement. (ai-act-service-desk.ec.europa.eu) (glocertinternational.com) Investor interest has already shifted into governance tooling: pure‑play AI‑governance vendors raised roughly $144.05 million across 15 deals in 2025 and the sector recorded about $691 million in equity funding between 2022–2025, according to an industry tracker. (newmarketpitch.com 1) (newmarketpitch.com 2) Market consolidation and enterprise uptake are visible in vendor reports naming Microsoft Purview, OneTrust, IBM Watsonx.governance, Credo AI and Atlan among leading platforms, and the IAPP published an AI Governance Vendor Report in January 2026 to map competing offerings. (gartner.com) (iapp.org) Russia’s Ministry for Digital Development published a draft law on sovereign AI that the ministry says would come into force on 1 September 2027 and create categories of “sovereign,” “national” and “trusted” models. (www1.ru) The draft and related proposals would allow Moscow to restrict or ban foreign models such as ChatGPT, Claude and Gemini and would require platforms above specified user thresholds to localize Russian user data for multi‑year retention. (money.usnews.com) (theoutpost.ai)
