Intelligence Cubed - d/acc-Driven Open Source Modelverse

Barriers in today's AI Model Ecosystem

Current AI Model Market: Big Tech Dominates, Smaller Creators Struggle

Closed‑source LLMs from tech giants hog the spotlight and revenue, crowding out niche models. Their locked code blocks fine‑tuning and remixing, throttling innovation and denying creators true ownership or income.

AI Market Dominance

Expensive Subscriptions vs. Pay-Per-Use: End Users Seek Diverse AI Models

For media creators, engineers, and everyday users, the current AI model pricing landscape often feels restrictive. Expensive subscription packages lock users into high recurring costs even if they only need occasional access, while pay-per-use options can become unpredictable and costly at scale.

Total Cost: $430/month

Opportunities: Low-Cost Access & Open-Source Growth

Low-Cost AI Model Boom: Flexible Ownership & investing on the Rise

The rise of low-cost AI models such as DeepSeek is reshaping the landscape by lowering entry barriers and enabling more flexible ownership.

Open-Source Beats Closed-Source for LLMs

Open-source large language models are rapidly outpacing closed-source alternatives by offering transparency, adaptability, and community-driven innovation.

Category	Open Source LLMs: Fine-tuning + Prompting	Closed Source LLMs: Prompt tuning ONLY
Task Specificity	Superior for niche tasks (customizable to data).	Limited by model's generality; struggles with edge cases.
Data Privacy	Full control; sensitive data stays in-house.	Risk: Data sent to third-party APIs.
Control	Full: Own model weights, data, and updates.	None: Dependent on provider's API/features.
Inference Cost	Fixed costs after setup; cheaper long-term for large workloads.	API costs grow with usage; may become prohibitive.
Latency	Low: Hosted locally; no API delays.	High: Dependent on external API response times.
Renewable Options	Possible: Choose green energy providers.	Indirect: Relies on provider's sustainability efforts.
Community Support	Strong: Open-source ecosystems enable collaboration.	Limited: Restricted to provider's roadmap.
Future-Proofing	Flexible: Adapt to new trends (e.g., new architectures).	Risky: API changes/pricing shifts may disrupt workflows.

Open-Source Co-creation with Revenue Infrastructure

We provide collaborative open-source models with ownership and revenue-sharing infrastructure for both Web2 and Web3 users.

Model as a Service (MaaS)

Access to 200+ ready-to-use models, available instantly on the platform
Connect multiple models into custom workflows for flexible automation
Auto-router functionality that intelligently selects the best model for your task

Tokenized Ownership & Usage-Based Pricing

Creators list models via IMO
Pay-per-use inference API
Built-in crowdfunding to open-source
51% sold ⇒ open-source release

Model as Liquid Asset

Seamless trading & instant transfer of ownership shares.

Community Co-Creation

Easily remix and chain models with low-code tools—turn multiple models into powerful AI workflows with loyalty accumulation.

fluent & flexible Workflows

Easily remix and chain models with low-code tools—turn multiple models into powerful AI workflows with loyalty accumulation.

Dynamic Pricing

Model value adjusts with real usage. Good models rise, weak ones fade—ensuring survival of the fittest.

Democratic AI Benchmark

AI Model Nasdaq: performance and adoption create live indices, replacing static expert scores

🔒

Other key features coming soon!

Create to earn. Co‑create to prosper.

Create & Release

Upload models to earn

→

Trade & Use

Pay for inference, buy ownership

→

Earn & Grow

Continuous royalties from usage

→

Co-creation

Combine into new assets

Any builders can mint their models as on‑chain assets, earn from automated usage royalties and ownership swapping, and instantly fork or combine any open model—every remix is tracked and revenue is split back to all contributors.

Intelligence³
Ecosystem

Creator

Consumer

Investor

How Intelligence³ Stands Apart

The smarter path to intelligence isn't building one massive LLM—it's creating a swap of vertical models. Each specialized for specific domains, collectively more powerful, and democratically owned by their creators and users.

Dimension

Intelligence³

Traditional Closed-source LLM

Path to 'Smarter' AI

✓ Ecosystem of Specialists- Fosters smaller, expert models trained on high-quality, vertical data, "smarter" for specific tasks with less processing time.

✗ Focuses on creating massive, general-purpose models. The immense data requirements can limit agility and make it difficult to achieve deep expertise in niche areas.

Investment

✓ Community funded

✗ Funded by large corporations.

Control

✓ Shared ownership. You own the model weights and can fine-tune them on your private data.

✗ API-level access, no Control.

Cost Model

✓ Transparent & Efficient, inference costs are fixed and predictable.

✗ Unpredictable, usage-based fees.

Data Privacy

✓ Sensitive Data stays in-house, max privacy.

✗ Users pay recurring subscription or API fees.

Community

✓ Community as owners & governors.

✗ Community as customers.

Dimension

Intelligence³

Pond

Sentient

Number of models

✓ 200+

✗ 39

✗ 9

Quanlity of models

✓ High

✗ Unknown

✗ Low

Create to earn

✓ Fully rewarded

✗ Partially

✗ No reward

Web2 user attraction

✓ Strong

✗ Limited

✗ Weak

AI empowered community growth

✓ Encouraged

✗ Limited

✗ Not supported

N-creation watermark

✓ Enforced

✗ Not enforced

Crypto for AI benchmark

✓ Yes

✗ No

See Intelligence Cubed in Action

Traditional AI Development

// Closed Source Approach
class ProprietaryAI {
    private blackBoxModel;
    private restrictedAccess;
    
    moveFastBreakThings() {
        // Rapid deployment
        // Unknown safety implications
        // Limited oversight
    }
}

d/acc Intelligence Cubed Approach

// Open Source d/acc Approach
class IntelligenceCubed {
    public transparentModel: DeepSeek;
    public openAccess: Community;
    
    defensiveAcceleration() {
        // Safety-first development
        // Democratic governance
        // Transparent processes
        return this.buildResponsibly();
    }
}

Open Source Modelverse Ecosystem

🌐 Interconnected AI models and frameworks
🔗 Seamless integration across platforms
📚 Open documentation and transparent development
🤝 Community-driven innovation and collaboration

DeepSeek Investment Strategy

📊 Strategic focus on next-generation AI models
🔭 Due diligence with safety and transparency criteria
🎨 Portfolio diversification across AI infrastructure
🚀 Long-term value creation through open-source innovation

About Us

Our Mission

Democratizing AI through defensive acceleration—building safe, transparent systems that empower creators and accelerate innovation responsibly.

Our Team

The I³ team is a collective of world-class builders, researchers, and strategists from elite institutions like Carnegie Mellon, Stanford, Berkeley, Duke, and Cornell. Our deep expertise in AI, blockchain, venture capital, and product design is backed by experience from Y Combinator, top tech firms like TikTok, and leading research labs. United in purpose, we are building the future of decentralized AI infrastructure.

Fellows

Our innovation is guided by an elite group of fellows including researchers from premier institutions. This powerful blend of academic rigor and industry savvy pushes the boundaries of what's possible.

This esteemed group includes Xuandong Zhao (Postdoc, UC Berkeley RDI, supervised by Prof. Dawn Song), Yuejiang Liu (Postdoc, Stanford AI Lab), Yaqi Xie (Postdoc, CMU Robotics Institute), Jason Dou (Postdoc, Harvard), and Peter Wang (Postdoc, Caltech), Shiyi Du (Ph.D., CMU, Computational Biology), Jiayuan Liu (Ph.D., CMU, Computer Science), Yitong Li (Ph.D., Stanford, Computational Science), and Chengfeng Mao (Ph.D., MIT, Computer Science).

Advisor

Prof. Tai-Sing Lee

Professor of Computer Science and Neuroscience, Carnegie Mellon University

Dual PhDs (Harvard & MIT HST) • Director, Lee Lab for Biological & Machine Intelligence • AI Mentor to Andrew Ng (CMU '97) • ICCV Helmholtz Prize, NSF CAREER, McDonnell–Pew Award • Pioneer in Bayesian Cortical Models & Neural Vision • Trained Leaders at DeepMind, OpenAI, Google, Berkeley

Investors & Partners

200+

AI Models

90 K

Citations

500k

Compute Credits

100K+

Followers

So inspired by what you are doing now and encourage you guys to explore further

— Vitalik's comment to I³, during the communication between Fernando & Vitalik at d/acc Day at Berkeley during ETH SF

Remodeling the AI World

We're creating the NASDAQ for AI—a transparent marketplace where model valuations are discovered through real market activity, democratizing AI investment for creators everywhere.

Explore Platform

Barriers in today's AI Model Ecosystem

Current AI Model Market: Big Tech Dominates, Smaller Creators Struggle​

Expensive Subscriptions vs. Pay-Per-Use: End Users Seek Diverse AI Models​