The New Enterprise Battle Over Who Owns the Model

2026-07-16 · Source: The AI Daily Brief: Artificial Intelligence News and Analysis · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, extended

Summary

Thinking Machines Lab (TML) has released Inkling, a new open-weight, mixture-of-experts model featuring 975 billion total parameters and 41 billion active parameters. This model supports a million-token context window and reasons across text, images, and audio. TML pre-trained Inkling from scratch, using a small bootstrap from KimiK 2.5. While its benchmarks are competitive but not state-of-the-art against frontier models, Inkling's significance lies in its design as an American-developed, non-distilled open model optimized for customization and fine-tuning via TML's Tinker platform. This release highlights an emerging enterprise focus on data sovereignty and cost-effective, custom model solutions, challenging the reliance on closed-source alternatives.

Key takeaway

For Directors of AI/ML evaluating their organization's model strategy, the emergence of open-weight, non-distilled models like Inkling signals a critical market shift. You should actively experiment with and assess these solutions, particularly for use cases requiring data sovereignty or cost optimization. Ignoring this increasingly diverse ecosystem risks missing opportunities for tailored, secure, and potentially more efficient AI deployments.

Key insights

Open-weight models like Inkling are driving an enterprise shift towards data sovereignty and custom fine-tuning solutions.

Principles

• Enterprise AI demands data sovereignty.
• Fine-tuning requires continuous effort.
• Open-weight models foster ecosystem diversity.

In practice

• Explore open-weight models for customization.
• Evaluate fine-tuning for specific workflows.
• Consider data sovereignty in model selection.

Topics

• Open-weight Models
• Enterprise AI Strategy
• Model Fine-tuning
• Data Sovereignty
• Thinking Machines Lab Inkling
• AI Infrastructure

Best for: CTO, AI Architect, AI Engineer, Director of AI/ML, VP of Engineering/Data, Consultant

Related on AIssential

• Thinking Machines amps up its bet against one-size-fits-all AI with its first open model, Inkling — Adaptable, open-weight AI models, fine-tuned by organizations, can outperform and be more cost-effective than general-purpose proprietary systems.
• The Token Apocalypse — Government restrictions and high costs are driving a rapid global shift from closed-source to open-weight AI models.
• Up the Stack: How AI’s Escape From the Commodity Trap Risks Enterprise Lock-in — AI labs are moving up the stack to avoid commoditization of model inference, creating potential customer lock-in.
• The Sequence Opinion #896: Spark, Compute, and the Two Metas — Meta is vertically integrating its AI stack, shifting from open-weights to proprietary models and infrastructure.
• The Next Wave of Enterprise AI — Enterprise AI is shifting from subsidized exploration to cost-optimized, agentic adoption, driving new tools and regulatory considerations.
• The real AI race may no longer be at the frontier — The AI industry is shifting towards open-weight models for production due to cost, ownership, and control benefits.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The AI Daily Brief: Artificial Intelligence News and Analysis.