Is NEAR AI Losing the AI Distribution Race? A Request for Metrics, OpenRouter Integration, Leadership Accountability, and vGPU Disclosure

Near Community and nodes
1

Is NEAR AI Losing the AI Distribution Race? A Request for Metrics, OpenRouter Integration, Leadership Accountability, and vGPU Disclosure

I am writing this as a NEAR holder who wants NEAR’s AI strategy to succeed.

NEAR AI has a credible technical thesis: confidential and verifiable AI infrastructure, private inference, hardware-backed attestations, an OpenAI-compatible API, secure agents, and a confidential GPU marketplace.

The concern is not that nothing is being built. The concern is whether this technology is translating into:

  • meaningful developer distribution;
  • sustained paid demand;
  • competitive unit economics;
  • measurable value accrual to NEAR;
  • transparent leadership accountability.

A separate and more urgent issue also requires clarification: the relationship, if any, between NEAR Foundation, NEAR AI and vgpu.network.

This post is not an accusation of misconduct against any individual. It is a request for verifiable information that should be available to the community and NEAR token holders.

1. NEAR AI Has Built Real Technology

It is important to acknowledge the progress already made.

NEAR AI has launched or announced:

The current NEAR AI Cloud website lists 51 models, including 15 models running confidentially inside TEEs.

This is meaningful engineering work. Privacy, confidential computing and verifiable execution could become an important AI infrastructure category.

However, strong infrastructure alone does not guarantee adoption.

2. The Main Concern Is Distribution

AI infrastructure is a distribution business as much as a technology business.

Developers generally choose services that are already available through the tools and platforms they use. One of the largest AI inference distribution platforms is OpenRouter.

OpenRouter states that it provides access to more than 80 inference providers and serves more than 10 million developers. It offers:

  • an OpenAI-compatible API;
  • automated billing and provider payments;
  • performance-based routing;
  • public latency, throughput and uptime statistics;
  • immediate exposure to an established developer base.

NEAR AI already operates an OpenAI-compatible API, so there appears to be a reasonable technical foundation for integration.

As of 29 August 2026, NEAR AI is not present in the public OpenRouter provider directory. Phala, which has previously collaborated with NEAR AI, is already represented there.

OpenRouter does disclose that it has a large application backlog and currently prioritizes providers with proprietary models. Therefore, absence from OpenRouter does not automatically mean NEAR AI has failed to apply or failed a technical review.

That is why the community should receive direct answers:

  1. Has NEAR AI applied to become an OpenRouter provider?
  2. If so, when was the application submitted?
  3. What is its current status?
  4. Are there technical, commercial or legal blockers?
  5. If NEAR AI has deliberately decided not to integrate, what is the alternative distribution strategy?
  6. Which platforms are expected to generate comparable developer reach and paid inference demand?

Without answers, it is difficult to determine whether OpenRouter is being actively pursued, delayed or simply overlooked.

3. Announced Partnerships Are Not the Same as Commercial Traction

NEAR AI publicly states that its infrastructure serves more than 100 million users across platforms such as Brave Nightly, OpenMind and Phala.

This statement requires clarification.

“Serving platforms with more than 100 million users” can mean several very different things:

  • access to a partner’s total potential user base;
  • availability to a subset of users;
  • actual monthly active users of NEAR AI-powered features;
  • completed inference requests;
  • paying customers;
  • subsidized or free usage.

These categories should not be combined into a single headline number.

The community needs operating metrics that distinguish potential reach from real usage and revenue.

At minimum, NEAR AI should publish quarterly:

Category Requested metrics
Demand Monthly paid inference tokens and total inference requests
Customers Number of paying organizations and paying developers
Retention Customer and API-key retention after 30, 90 and 180 days
Revenue Quarterly AI Cloud, agent hosting and enterprise revenue
Unit economics Infrastructure cost, gross margin and average revenue per million tokens
Infrastructure Installed GPU capacity and average utilization
Reliability Uptime, latency, throughput and request failure rate
Distribution Traffic and revenue by integration or distribution channel
Subsidies Portion of usage paid by customers versus NEAR Foundation or ecosystem funding
Partnerships Production integrations versus pilots, MOUs and announcements

Sensitive customer information does not need to be disclosed. Aggregate metrics would be sufficient.

4. What Does NEAR AI Mean for the NEAR Token?

The launch of staking-based payments is an important step toward connecting AI usage with NEAR.

Under the current model, users can stake NEAR and receive credits for confidential inference or agent hosting. The underlying NEAR remains owned by the user, while the staking rewards are routed to NEAR AI as payment for services.

This creates demand to stake NEAR, but it is not equivalent to:

  • burning NEAR;
  • permanently removing NEAR from circulation;
  • generating protocol revenue for all token holders;
  • distributing NEAR AI profits to validators or token holders;
  • creating equity ownership in NEAR AI.

According to the NEAR AI Terms of Service, staking rewards associated with these subscriptions accrue to NEAR AI as payment for services.

The community should therefore receive clear reporting on:

  1. Total NEAR staked through NEAR AI products.
  2. Total staking rewards routed to NEAR AI.
  3. Total AI credits issued and consumed.
  4. Percentage of credits that expire unused.
  5. Revenue generated from direct payments versus staking rewards.
  6. Whether Jasnah Inc. or NEAR AI pays protocol fees, revenue share or any other consideration back to the NEAR ecosystem.
  7. Whether NEAR AI operations create structural NEAR demand beyond temporarily delegated stake.
  8. What happens to NEAR AI revenue and whether any portion is reinvested into protocol development.

The current model may create useful demand for NEAR, but its economic effect should be measured rather than assumed.

5. Leadership Accountability Is Unclear

The official NEAR AI team page lists:

  • Illia Polosukhin — Founder;
  • George Zeng — CPO and GM;
  • Sergey Astretsov — Head of Product;
  • Cameron Dennis — Director of AI;
  • Pierre Le Guen — Head of Engineering;
  • Pranav Raja — ML Researcher.

This suggests that technical, product and commercial responsibilities are distributed across several leaders.

However, the exact mandate of the Director of AI is unclear.

In his governance delegate proposal, Cameron Dennis disclosed that he began contracting for NEAR AI through Banyan Collective in October 2024. He identified one of his main priorities as making NEAR AI the best way to monetize model training, fine-tuning and inference.

More recently, a NEAR AI partner described him as Head of Partnerships at NEAR AI.

There is nothing inherently wrong with a non-engineering executive leading partnerships, integrations, investments or commercialization. These functions are essential.

The issue is accountability.

If the Director of AI is responsible for commercial adoption and distribution, the community should be able to evaluate that role using measurable KPIs:

  • paid inference growth;
  • customer acquisition;
  • customer retention;
  • revenue;
  • strategic integrations;
  • OpenRouter and other distribution channels;
  • conversion of partnerships into production workloads;
  • GPU utilization.

If these responsibilities belong to another executive, NEAR AI should explain who owns them.

Questions for the team:

  1. What is the exact mandate of the Director of AI?
  2. Is the role primarily technical, strategic, commercial or partnership-focused?
  3. What 2025 and 2026 KPIs were assigned to the role?
  4. What were the targets and actual results?
  5. Who is directly accountable for developer distribution and paid inference growth?
  6. How is the performance of NEAR AI leadership reviewed?
  7. Does NEAR Foundation receive regular operating and financial reports from NEAR AI?

This is not about attacking one person. It is about avoiding a structure where broad titles make it impossible to identify who owns specific outcomes.

6. Urgent Disclosure Request Regarding vGPU.network

A separate website, vgpu.network, is currently using the names and photographs of several NEAR and NEAR AI leaders.

The website lists or displays:

  • Illia Polosukhin;
  • David Norries;
  • Abhi Vaidyanathan;
  • Cameron Dennis;
  • George Zeng;
  • Bowen Wang;
  • other NEAR ecosystem participants.

It describes Cameron Dennis as “Director of AI, NEAR Foundation” and George Zeng as “CPO at NEAR and GM of NEAR AI.”

The website claims that:

  • NEAR AI guarantees full GPU utilization from day one;
  • GPU infrastructure processes workloads for NEAR Foundation;
  • NEAR Foundation continuously purchases generated reward units;
  • NEAR Foundation is the guaranteed purchaser of 100% of deployed GPU compute output.

The platform offers GPU-related NFTs priced from approximately $300 to $50,000 and advertises referral rewards.

These are material financial representations.

However, the platform’s Terms of Service identify the operating company only as:

“Legal Entity Name, a company organized and existing under the laws of Ukraine.”

Its Privacy Policy similarly fails to provide a completed legal entity name and registered address.

The Terms also state that:

  • an NFT does not provide equity, debt, revenue-sharing rights or physical possession of the underlying GPU;
  • NFT transfers are restricted for 500 days;
  • vGPU token rewards are locked for 550 days;
  • the Token Generation Event is determined by the company at its sole discretion;
  • claiming tokens depends on a manual approval mechanism that the company can enable or disable;
  • the company retains full operational control over the hardware;
  • the company can modify the reward schedule;
  • NFT purchases are made using USDT or USDC.

The website claims that its smart contracts are audited, but I have not found publicly identified contract addresses, an audit firm or an independent audit report on its public pages.

Meanwhile, the official NEAR AI Terms identify the service operator as Jasnah Inc., a Delaware corporation doing business as NEAR AI. The official NEAR AI Confidential GPU Marketplace is presented as infrastructure connecting GPU providers with enterprise and government workloads—not as a retail NFT investment and referral program.

I have not found an announcement on near.ai or near.org confirming that vgpu.network is an authorized NEAR Foundation, NEAR AI or Jasnah Inc. initiative.

This does not prove that the platform is fraudulent or that the people displayed on the website are involved. Their names and photographs could have been used without authorization.

However, because the website solicits financial participation and makes claims about guaranteed demand from NEAR Foundation, an immediate official clarification is necessary.

Questions for NEAR Foundation and NEAR AI:

  1. Is vgpu.network formally affiliated with NEAR Foundation, NEAR AI or Jasnah Inc.?
  2. Have the individuals displayed on the website authorized the use of their names, photographs and job titles?
  3. Does NEAR Foundation have a legally binding obligation to purchase 100% of the platform’s compute output?
  4. Has NEAR Foundation authorized the website to advertise such a guarantee?
  5. What legal entity operates the platform?
  6. Why is the company identified only as “Legal Entity Name” in its legal documents?
  7. What entity receives NFT purchase proceeds?
  8. Where are the smart-contract addresses and independent audit reports?
  9. Where are the GPU serial numbers, procurement records, data-center locations and utilization reports?
  10. Is there a signed compute-purchase agreement with NEAR Foundation or Jasnah Inc.?
  11. Who approved the referral program and its financial representations?
  12. If the website is unauthorized, will NEAR Foundation publish a warning and request the removal of its names, photographs and branding?

Until these questions are answered through an official NEAR-controlled channel, community members should exercise extreme caution and should not connect wallets or transfer funds based solely on the representations made by this website.

7. Requested Actions

I respectfully request that NEAR Foundation and NEAR AI take the following actions.

Within 7 days

  • Confirm or deny any relationship with vgpu.network.
  • Confirm whether the listed team members authorized the use of their identities.
  • Confirm whether any compute-purchase guarantee exists.
  • Publish a warning if the website is unauthorized.

Within 30 days

  • Confirm the status of the OpenRouter provider application or explain the alternative distribution strategy.
  • Publish the mandate and KPIs for the main NEAR AI leadership roles.
  • Distinguish production customers, active users, pilots and announced partnerships.
  • Publish initial commercial and infrastructure metrics.

Quarterly

Publish a NEAR AI transparency report covering:

  • paid inference volume;
  • revenue and aggregate expenses;
  • paying customers;
  • retention;
  • GPU capacity and utilization;
  • uptime and performance;
  • customer-funded versus subsidized usage;
  • NEAR staked through AI products;
  • staking rewards routed to NEAR AI;
  • material related-party relationships;
  • funding received from NEAR Foundation or ecosystem entities;
  • progress against previously announced targets.

Conclusion

NEAR AI may have a genuine technical advantage in confidential and verifiable inference. The work on TEEs, attestation, private agents and wallet-native payments is real and potentially valuable.

But the market will not reward technology that lacks distribution, measurable demand and clear accountability.

The community should not be asked to evaluate NEAR AI through announcements, potential partner reach or broad claims alone. It should be able to evaluate:

  • how many customers are paying;
  • how much inference is being consumed;
  • whether GPU infrastructure is utilized;
  • whether operations are economically sustainable;
  • how NEAR benefits;
  • who is accountable for each outcome.

The vgpu.network matter is particularly urgent because it combines the identities of NEAR leaders with financial promises, guaranteed-demand claims and incomplete legal documentation.

A clear and evidence-based response would protect the community, improve confidence in NEAR AI and demonstrate the transparency expected from an ecosystem building user-owned infrastructure.

1 Like