On established benchmarks (GPQA, MMLU, HumanEval), the ceiling was reached in late 2024. New models appeared throughout 2025 but no longer pushed the frontier; they matched it with smaller, more efficient architectures. Google's Gemma 3 N4B, designed to run offline on a smartphone, scored as high on HumanEval as GPT-4O had eighteen months earlier. Intelligence continued to increase on newer, harder benchmarks, with IQ-test-style evaluations showing gains of 20 to 30 points, but for everyday tasks, the difference was imperceptible. All frontier closed-source models (GPT-5, Claude Opus, Gemini) clustered at the same performance point; open-source models trailed by roughly six months.
Two days after our previous webinar in January 2025, DeepSeek released its model. US big tech spent over 100 billion dollars per quarter on AI infrastructure, betting that more compute leads to better models. China released frontier models as open weights, published the research behind them, and iterated at two to three state-of-the-art releases per month by year-end. The cost of interacting with a model at a given intelligence level dropped from $32 per million input tokens in January to $0.10 by December (a factor of 300). Chinese labs never produced the absolute top-scoring model, but consistently came within a few IQ points at a fraction of the cost.
With model intelligence converging, the differentiator shifted to the platform layer. OpenAI's major 2025 announcements were platform features: native UI components in ChatGPT and agentic workflow builders. Google's advantage became structural. Gemini integrates with Gmail, Calendar, Drive, and YouTube natively. Any platform could technically connect to any model; the bundling is a business decision, not a technical constraint.
Image generation became indistinguishable from real photography, even for arbitrary individuals given a single reference image. Video generation crossed a similar threshold. In software engineering, the lead developer of Claude Code reported that 100% of his contributions over the past 30 days were written by the tool itself. In our team at bbv, none of us has written a complete function by hand in months. The job has moved from writing code to managing coding agents: formulating requirements, reviewing output, maintaining documentation, and iterating on permanent instructions to correct recurring mistakes.
In 2025, 51% of internet traffic came from bots, 15% of Reddit posts were demonstrably AI-generated, and 18% of website content was AI-produced. On GitHub, 300 million new repositories appeared in a single year, and at least 41% of all code on the platform is now AI-generated. AI models cite AI-generated web pages as sources, users trust the citations without checking, and fabricated information acquires an appearance of legitimacy. During the Venezuela crisis, AI-generated images of events that had not yet occurred were shared by US senators as fact.
Software engineering has crossed the productivity threshold. Other industries follow as AI now covers text, image, video, and code generation at production quality.
AI platforms work well in their native ecosystems (Gemini with Google Workspace, Copilot with Microsoft 365), but most enterprises use custom ERPs, on-premise databases, and proprietary tools these platforms cannot reach. The Model Context Protocol, originally developed by Anthropic and now governed by the Linux Foundation, standardizes how AI platforms communicate with business software. Unlike REST APIs, MCP servers describe their capabilities in natural language, allowing language models to autonomously decide which tools to invoke. OpenAI, Anthropic, and Google all support it. Thousands of public servers already exist.
With models interchangeable and MCP standardizing tool integration, the remaining competitive moat is lock-in. Passive lock-in accumulates through chat history, memory systems, and audit logs. Active lock-in comes from enterprise investments: vectorized knowledge bases, custom agent workflows, and compliance configurations. None of these transfer between platforms.
US tech dominance, export controls on Chinese chips, and rising subscription costs are driving enterprises toward digital sovereignty. Open-source alternatives exist at every layer: open-weight models hosted by Swiss providers like Infomaniac or Exoscale, open-source platforms deployable on-premise, and MCP servers for Swiss business integrations like Bexio.
Model convergence describes the 2025 trend where all frontier language models proprietary and open-source reached approximately the same intelligence level on standard benchmarks. Closed-source models from OpenAI, Anthropic, and Google cluster at the same performance point, and open-source models trail by only about six months, closing rapidly.
The US invested over 100 billion dollars per quarter in AI infrastructure, betting on compute scale. China pursued an open-weight strategy, releasing state-of-the-art models freely and publishing the research behind them. China dominated open-source model releases throughout 2025, and its innovations drove the cost of intelligence per dollar down by a factor of 300.
MCP is a protocol originally developed by Anthropic and now governed by the Linux Foundation. It standardizes the communication between AI platforms (MCP clients) and business applications (MCP servers). Unlike REST APIs, MCP servers describe their capabilities in natural language so that a language model can autonomously decide when and how to use them.
AI slop refers to content generated by AI without meaningful human review or thought. In 2025, 51 percent of internet traffic came from bots, 15 percent of Reddit posts were demonstrably AI-generated, and 18 percent of website content was AI-produced. This creates dangerous feedback loops where AI models cite AI-generated content as sources.
AI platform lock-in occurs when switching costs accumulate through chat history, memory systems, audit logs, vectorized company data, and custom agent workflows built on a specific platform. The largest lock-in comes from a company's own investments in data curation, agent configuration, and workflow modeling costs that cannot be transferred to another platform.
2026 will be defined by five trends: (1) AI entering the daily work of non-technical industries, (2) business integration through MCP becoming mainstream, (3) platform lock-in awareness growing, (4) digital and AI sovereignty becoming strategic priorities, and (5) enterprises deliberately choosing open-source alternatives across all layers of the AI stack.
.
Copyright 2026 - Joel P. Barmettler