Some Ideas for What Comes Next, May 2026

Key Takeaways

• Open-weight models lack agentic performance comparable to Claude Code.
• Gemini 3.5 Flash still trails Claude Code and Codex in productivity.
• U.S. labs dominate compute share, limiting Chinese open model scaling.
• Gemma 4’s permissive license drives rapid adoption among developers.
• Regulatory and institutional actions intensify AI’s social and political challenges.

Pulse Analysis

The widening chasm between open‑weight and closed‑source AI models is reshaping enterprise adoption strategies. Closed systems such as Claude Code and Codex have demonstrated agentic capabilities that translate directly into productivity gains, prompting firms to favor subscription‑based access over cheaper, open alternatives. This dynamic pressures open‑model developers to accelerate performance breakthroughs while navigating tighter budgets, creating a market where cost‑effectiveness competes with functional superiority.

Compute allocation remains a decisive factor in the global AI race. Recent analyses show U.S. labs controlling roughly 47% of the world’s AI compute, dwarfing Chinese counterparts whose resource constraints limit the scaling of open‑weight models like Kimi and Qwen. In contrast, American releases such as Gemma 4, backed by an Apache 2.0 license, have sparked a surge in developer uptake, illustrating how permissive licensing can quickly translate benchmark parity into real‑world deployment. This trend underscores the strategic advantage of aligning open‑model ecosystems with accessible tooling and community support.

Regulatory and institutional pressures are escalating, turning AI development into a geopolitical and social battleground. From the Vatican’s extensive doctrinal statement on AI ethics to U.S. designations of certain models as supply‑chain risks, policymakers are signaling a willingness to intervene in the technology’s trajectory. These actions amplify public concerns about data‑center expansion and job displacement, potentially curbing investment flows and prompting firms to prioritize compliance and transparency. Companies that can navigate both technical innovation and evolving regulatory landscapes will be best positioned to sustain growth in an increasingly contested AI environment.