Yesterday, the Linux Foundation announced the Agentic AI Foundation (AAIF) with the kind of excitement that usually surrounds a moon landing. OpenAI and Anthropic appeared to join hands which is great move. Anthropic kindly donated the whole MCP project and OpenAI also donated AGENTS.md to the Linux Foundation. On top of that Google, Microsoft, AWS which are investing heavily in Agentic AI and have some great innovation already proven in the space. They all appeared as platinum founding members. The press release described it as “a neutral home for the future of agentic AI.” And for a moment, I wanted to believe it. I genuinely did and wanted to celebrate the movement AAIF as start of something big as big players joining hands.
Anthropic, OpenAI and Google Joining Hands
The three top model providers joining hands and working together is a great sign. Historically, Google has donated many projects to the Linux Foundation, such as Kubernetes, and recently they also donated the A2A protocol. In theory Google is the dominating player who has proven the track of donation to open foundations like the Linux Foundation. For once, the three biggest closed-model labs were not arguing about secrecy or benchmarks. They were agreeing on something that actually matters: shared infrastructure. Open protocols like MCP, AGENTS.md, and tools that could finally prevent fragmentation in agent ecosystems before it becomes permanent. That part is real progress. The industry truly needs interoperable foundations.
But the longer I looked at the launch, the more the optimism started to fade. The excitement slowly turned into a familiar concern, longer I looked who else are there, the worse it smelled. Here is the quiet truth that almost no one in the announcement thread is saying out loud:
The seats at the table were not earned.
Platinum membership, which is the only tier that provides real voting power and a board seat, costs 350,000 dollars, matching other Linux Foundation directed funds. I totally get the point that OpenAI, Anthropic and Google driving this as platinum members. I can also accept the presence of Microsoft and AWS as they provided infrastructure to drive this innovation but not sure what the role of Block, Bloomberg, Cloudflare there? Why they are platinum members and what criteria applied to give them the platinum seat and drivers seat to claim the ownership of the Agentic AI innovations. There is no public and transparent criteria for technical merit. No requirements related to contributions, open governance experience, or shipped agent systems.
What's the contribution of Goose, Bloomberg and Cloudflare in Agentic AI?
Block, primarily a payments company, is suddenly a cornerstone steward of agentic AI governance because they open-sourced Goose, a tool most developers had never seen before this week. It seems like the early MCP client that has been used only internally at Block and very few others outside, I barely heard about Goose apart from logo in some sponsored conferences. Bloomberg, Cloudflare, Shopify, and Oracle, all great companies in their respective industries, are now positioned as foundational voices in agent governance even though they have almost no visible history in Agentic AI evolution I genuinely don't see any contribution from these companies in the Agentic AI space apart from some articles here and there. They might have something internally with no visibility publicly.
Where are the Real Builder of Agentic AI
If you scan the Agentic AI Foundation membership right now, you will be surprised who are these early companies became members of the Agentic AI Foundations who lack contribution compared to the agent-native startups going fast. Most members seem like incumbents who slapped Agentic AI on top of old systems or startups that never did meaningful contributions to Agentic AI. The only genuine members, I can see there are Pydantic, Salesforce, LanceDB and HuggingFace but they are also not on the top seat.
Meanwhile, the teams and startups who actually built the field are absent. This includes the agent building companies like LangChain, CrewAI, LlamaIndex who have more developers adopting and using these tools. The coding agents like Cursor, Windsurf, Warp, Factory, OpenHand. Most importantly, xAI and the Grok models, close competitors to Anthropic and OpenAI, are missing in this list. They are either not considered at all or completely excluded from the initial foundation list. They might join later but it seems like the foundation did not consider them initially even though they are real players and innovators in the Agentic AI space.
The player like Grok who has money could be in the Platinum list but it seems like the door has been closed already and only a handful of companies will take the driver seat, and some of them genuinely do not have a considerable contribution to be there. Some of the Agentic AI and agent-native startups might join later and some will stay out due to the opaque membership structure.
Models Stay Closed, Interfaces Open
The current state of the Agentic AI Foundation feels like this is not openness, this is a country club wearing an open-source hoodie. It is looking like a conference sponsorship or trade show where anybody can pay money and become a member. The three big labs, Anthropic, OpenAI and Google, have become platinum members and Anthropic donated the whole MCP project, OpenAI donated AGENTS.md, and Google donated A2A. This is wonderful but the reality remains that all of them maintain closed-source models.
I am not opposed to foundations, standards, or coordination by big labs. I am also not opposed to any members either. Their contributions may be reasonable tools. Goose is an early MCP client, but it is not one of the most important agentic projects in the world today. It is simply an MCP client like thousands of others. Yet it has been elevated into the top tier of the AAIF because Block could pay the platinum fee and other, more influential projects could not. This reveals the real governance model: money determines merit, even when the technical contributions tell a different story. On another note, players with both innovation and money have seen the door closed for platinum membership already. That is a structure I cannot endorse as a founder or as someone who believes deeply in the future of autonomous systems and Agentic AI. It sounds like open Linux Foundation is just face.
Superagentic AI will never join the AAIF
Superagentic AI, the name itself reflects what we do. We have the biggest open Agentic AI community in the UK, London Agentic AI meetup, and we are coming to the USA soon. Superagentic AI already delivered a research-level Agent Optimization platform and Agent Framework SuperOptiX. We already have SuperRadar, a global Agentic AI landscape map that is unbiased and covers every single Agentic AI tool in the world. We have dozens of open source projects on GitHub. We have 5 long-term pillars of Agentic AI. What we lack is VC backing and fake GitHub stars.
Superagentic AI is an early-stage startup and cannot afford gold or platinum AAIF memberships, but we can easily afford the $10,000 silver membership. However we are staying out because the membership structure feels like a conference sponsorship and the merit criteria is unclear. There are great members who are actively participating in the Agentic AI movement, but we do not want to be surrounded by those taking advantage of the label to surf the wave.
Future Consideration
I believe the future of autonomous agents should be decided by the people who actually build them every day and real Agentic AI companies, not by whoever has the largest corporate treasury and the best PR team. There is nothing against foundations or standards. As founder of Superagentic AI, I will keep using and potentially contributing to MCP and AGENTS.md because they are useful and permissively licensed and now they are even more open. But I will not pay hundreds of thousands of dollars to sit at a table where innovation is optional. If the AAIF evolves, if they create real, low-barrier paths and entires based on the merit, if they tie voting power to measurable technical impact instead of dues tier, I will happily reconsider.
Until then, we will keep building in public, shipping fast, and staying independent. Good luck Agentic AI Foundation and I hope you introduce some merit-based gates to entry before it gets too late!