Artificial intelligence is reshaping software development, with AI coding tools now generating 80% of all code written. The shift comes as OpenAI reportedly stands 70-80% of the way toward achieving Artificial General Intelligence, while sovereign AI initiatives are rewriting the rules of data ownership worldwide.
How AI is taking over coding
AI-powered assistants like GitHub Copilot and others have become standard in developer workflows. According to industry reports, these tools now produce four out of every five lines of code. The productivity gains are significant — developers can prototype features in hours that once took days.
Still, human oversight remains essential. The tools excel at boilerplate and repetitive tasks but still struggle with nuanced logic, security edge cases, and novel architectures. Teams report that the final 20% of code — the part that requires architectural decisions and domain-specific reasoning — still comes from human engineers.
OpenAI's march toward AGI
OpenAI is reportedly between 70% and 80% complete on its journey to Artificial General Intelligence — a system that can perform any intellectual task a human can. The company has not publicly confirmed the figure, but internal assessments suggest the remaining hurdles involve reasoning, long-term planning, and robust generalization across unrelated tasks.
Reaching AGI could supercharge coding tools further. A fully general intelligence would not just write code but design entire software systems, debug them, and adapt to new requirements without retraining. That would upend the current model where humans still direct and review the AI's output.
Sovereign AI and data ownership
Governments and large enterprises are increasingly building their own AI systems — a trend called sovereign AI. These initiatives aim to keep data within national borders and under local control, rather than relying on foreign cloud providers.
The shift is redefining who owns the data used to train and operate AI models. Countries from India to Germany are investing in domestic computing infrastructure and data centers. They're also drafting new regulations that require AI models trained on citizen data to be hosted and governed locally.
For companies like OpenAI, sovereign AI creates both a challenge and an opportunity. Their models may need to be deployable in isolated, air-gapped environments. Data from one country cannot simply flow into a central training pool without risking regulatory clashes.
These three trends — AI-written code, progress toward AGI, and sovereign data ownership — are intertwined. As machines write more code, the line between human and machine intelligence blurs. As AGI approaches, the question of who controls the data becomes existential. And as nations assert ownership over their digital assets, the global AI landscape fractures into distinct, locally governed ecosystems.
The next milestone will be whether OpenAI can bridge the remaining 20-30% to AGI while keeping its models compliant with a patchwork of national data laws. That tension will define the next phase of the AI industry.
