Skip to content

All articles

Open AI standards will shape business software for five years

Anthropic, OpenAI, Alibaba, and Hugging Face contributed their AI agent tools to the Linux Foundation. Here is how open standards shape your software.

On 22 October 2026, the Linux Foundation (opens in a new tab) announced that the Agentic AI Foundation expanded its open-governance portfolio to 10 hosted projects. The foundation launched in December 2025 with three projects. Within its first year, it has gathered contributions from the largest developers of artificial intelligence. Anthropic, OpenAI, Hugging Face, and Alibaba each contributed core agent technologies to the neutral foundation.

This matters because the companies building the most advanced artificial intelligence are agreeing on how these systems should connect. For a business owner planning their technology spending, this signals a clear direction for the next five years. The software industry is building a shared, open foundation for AI agents. You will not have to lock your company into a single vendor to use automated workflows.

The shared plumbing of artificial intelligence

The four new projects address different layers of how AI agents operate. Anthropic contributed Agent Skills. This project provides a standardized format for packaging reusable agent instructions, scripts, and domain knowledge. When an agent needs to learn a specific task, it can load these skills on demand. This makes specialized knowledge portable across different agent products and platforms.

OpenAI contributed Open Responses. This is an open specification for building multi-provider interfaces, based on the company's own application programming interface. It defines a shared schema and consistent streaming events. Developers can write their requests and outputs once, and then run them across different language models with minimal translation.

Hugging Face contributed OpenEnv. This framework provides execution environments for model evaluation and reinforcement learning. It uses a containerized service to expose a familiar agent interface, making the environments interoperable and scalable. Alibaba contributed OpenSandbox. This platform provides secure sandbox runtime execution for AI applications. It includes software development kits for multiple programming languages and a protocol that defines execution lifecycles.

These contributions represent a major shift in the technology market. The largest players are choosing to share the underlying infrastructure of artificial intelligence. They are giving these tools to a foundation that operates under open, vendor-neutral governance. Manik Surtani, CTO of the Agentic AI Foundation, stated the goal is "putting the community first" by building shared standards.

What it means for your business

For a small or mid-sized business, open standards dictate how much you pay for software and how easily you can change providers. The expansion of the Agentic AI Foundation provides concrete benefits for companies of all sizes.

If you run a 50-person logistics firm, Open Responses changes how you buy technology. You might build an automated routing system using one provider's language model today. If a competitor releases a cheaper or faster model next year, the open specification means you can switch. You will not have to rewrite your entire software stack to change providers. This keeps your technology costs competitive and prevents vendor lock-in.

For a specialized consulting firm, Agent Skills creates new opportunities. You can package your firm's specific domain knowledge into a standardized format. Because the format is portable, you can deploy these instructions across different agent products and platforms. You write the specialized capability once and use it wherever your clients need it. This turns your internal processes into reusable assets.

For the finance lead at a mid-sized manufacturer, these standards mean predictable integration costs. When software speaks a common language, you do not need to hire expensive specialists to connect different systems. The plumbing works immediately. You can allocate your budget to solving business problems instead of paying for custom software integration.

A secure foundation for the next five years

Over the next five years, artificial intelligence will move deeper into business operations. Models will transition from answering questions to executing complex tasks across your company data. This transition requires secure and reliable environments.

Alibaba's OpenSandbox addresses the security side of this transition. When an AI agent writes and executes code, it needs a safe place to do so. A secure sandbox ensures that an automated process cannot accidentally access sensitive payroll data or delete customer records. Standardizing this security layer means smaller companies can trust agent workflows without building custom security infrastructure.

Hugging Face's OpenEnv addresses how these systems learn and improve. By standardizing the environments where models are evaluated, the industry can accurately compare different agents. A business will be able to look at standardized benchmarks to see which agent performs best at a specific task. You will make purchasing decisions based on verifiable data rather than marketing claims.

This shared infrastructure mirrors how cloud computing developed. Early cloud adoption was fragmented and expensive. Open standards eventually made cloud infrastructure a cheap, reliable utility. The Agentic AI Foundation is accelerating that same process for artificial intelligence.

What to watch

Watch for enterprise software vendors updating their 2027 product roadmaps. The true test of these standards is how quickly commercial software adopts them. Look for announcements from your current software providers stating they support Agent Skills or Open Responses.

Monitor the pricing of AI integration services over the next 18 months. As these open standards take hold, the cost to connect different AI tools should drop significantly. If your technology agency quotes high fees for basic model integration next year, you should question their methods.

Pay attention to how smaller model providers position themselves. Open standards benefit challengers. Watch to see if regional AI companies adopt Open Responses to capture market share from the established global players.

Open standards historically lower costs and increase choices for small and mid-sized companies. The agreement between major vendors on 22 October provides a clear map of how business technology will operate through the end of the decade. Businesses that plan their technology around these open protocols will maintain flexibility and control over their software budgets.