Why Crunch Lab's $5M Raise Could Transform How Enterprises Build AI Models

How Crunch Lab Plans to Replace $100M Hiring Budgets with 10,000 Crowdsourced Engineers

On October 7, 2025, Crunch Lab announced it secured $5 million in funding co-led by Galaxy Ventures and Road Capital, with participation from VanEck and Multicoin. The round, which closed in June, brings the company's total funding to $10 million after a $3.5 million seed round in 2024. The company builds what it calls an "intelligence layer" for decentralized AI, connecting enterprises with a network of more than 10,000 machine learning engineers and 1,200 PhDs through its platform, CrunchDAO.

\ The concept challenges how organizations approach AI development. Instead of spending years recruiting specialists or building internal teams, enterprises submit problems as encrypted modeling challenges. Contributors compete to build solutions, and rewards flow to whoever delivers the best results. The Abu Dhabi Investment Authority (ADIA) Research Lab reported a 17% improvement in cross-sectional asset pricing predictions using this method. The Broad Institute of MIT and Harvard used the network for cancer gene research with computer vision. A global investment bank now runs Mid+One, a crowdsourced pricing engine for FX OTC markets, in production.

\ The funding reflects growing interest in decentralized approaches to AI infrastructure. Galaxy Ventures, Road Capital, VanEck, and Multicoin have all made bets that collective intelligence models can compete with centralized alternatives. The company was also selected for the Solana Incubator's second cohort earlier in 2025, and the platform operates on Solana's blockchain.

\

The Economics of Talent Scarcity

Hiring specialists in machine learning costs money. A senior ML engineer in the United States commands an average salary between $150,000 and $250,000 per year, according to industry data. Building a team of 10 to 20 engineers to tackle predictive modeling can easily exceed $2 million annually before factoring in infrastructure, benefits, or retention costs. Enterprises in finance, healthcare, and technology compete for the same limited pool of candidates.

\ Crunch Lab's model offers a different path. Instead of hiring full-time staff, companies post modeling challenges. Thousands of engineers participate, and organizations pay only for results. Jean Herelle, CEO of Crunch Lab and CrunchDAO, said, "AI today is constrained by hiring bottlenecks, siloed teams and an inability to scale effectively. We've flipped that model. Instead of competing for scarce talent, we give enterprises secure access to all of it through a decentralized network."

\ The approach mirrors crowdsourcing platforms like Kaggle, but with a focus on production deployment rather than competition alone. Companies submit real problems, not academic exercises. Contributors work with encrypted data, so proprietary information stays protected. The network then aggregates predictions from multiple models to generate final outputs. Herelle added, "This isn't theoretical hype, it's proven. When thousands of practitioners compete, you uncover solutions even the best internal teams miss."

\

Proven Performance in High-Stakes Environments

Results matter more than promises in AI. Crunch Lab points to three deployments as evidence its method works. The first involves ADIA Lab, the research division of the Abu Dhabi Investment Authority. ADIA manages more than $900 billion in assets, making it one of the largest sovereign wealth funds in the world. The lab used CrunchDAO's network to improve predictions for asset pricing across different securities. The 17% accuracy gain translates to better portfolio decisions and risk management at scale.

\ The second case involves the Broad Institute of MIT and Harvard, a research organization focused on genomics and biomedical science. The institute used the network for cancer gene research, applying computer vision techniques to analyze data. The press release describes the results as a "breakthrough," though no specific metrics were disclosed. Cancer research involves identifying patterns in vast datasets, and machine learning models help researchers detect relationships that might otherwise go unnoticed.

\ The third deployment is Mid+One, a pricing engine for foreign exchange over-the-counter (OTC) markets. OTC trades happen directly between parties, without exchanges, and pricing depends on real-time supply and demand. A global investment bank, which Crunch Lab did not name, now uses Mid+One in live trading. The engine relies on crowdsourced models to calculate mid-market prices, the midpoint between buy and sell quotes. Accurate pricing reduces transaction costs and improves execution for large trades.

\

Why Investors See an Infrastructure Play

Venture capital in AI has tilted toward infrastructure in recent years. Investors pour money into platforms that enable other companies to build applications, rather than applications themselves. Crunch Lab fits this category. Will Nuelle, General Partner at Galaxy Ventures, said, "Crunch Lab is building an intelligence layer for global enterprises. Whether predicting asset prices, optimizing energy demand, or advancing healthcare diagnostics, CrunchDAO's crowdsourced models unlock smarter, faster decision-making."

\ Thomas Bailey of Road Capital echoed the sentiment: "

\

The framing positions CrunchDAO as infrastructure, not a vertical solution. The platform could serve finance, healthcare, logistics, energy, or any field where predictive modeling drives value.

\ The choice of investors also signals strategy. Galaxy Ventures operates as the venture arm of Galaxy Digital, a firm focused on digital assets and blockchain technology. Road Capital and Multicoin both invest in decentralized networks and crypto infrastructure. VanEck, known for exchange-traded funds, has expanded into digital assets. The investor group suggests Crunch Lab sees itself as part of the decentralized web, not traditional software-as-a-service.

\

Risks and Questions About Decentralized AI

The concept raises questions. First, how does the platform ensure quality control when thousands of anonymous contributors submit models? CrunchDAO uses a performance-based reward system, so contributors earn tokens based on accuracy. But one bad model in a production environment can cause damage. Enterprises need guarantees, not probabilities.

\ Second, data security remains a concern. Crunch Lab encrypts data before sharing it with contributors, but encryption methods vary in strength. Homomorphic encryption, which allows computation on encrypted data, is still developing and can be slow. Zero-knowledge proofs offer another path, but they add complexity. The company has not disclosed which methods it uses or how it audits security.

\ Third, the platform depends on network effects. With 10,000 contributors, CrunchDAO has enough participants to generate useful results. But maintaining engagement over time is difficult. Contributors need incentives to keep participating, and rewards must feel worthwhile. If the network shrinks, the quality of predictions could decline.

\ Fourth, regulatory uncertainty looms over decentralized platforms. Enterprises in finance and healthcare operate under strict compliance rules. Using a decentralized network with anonymous contributors might conflict with Know Your Customer (KYC) or Anti-Money Laundering (AML) requirements. Crunch Lab will need to navigate these rules as it scales.

\

The Path Forward for Collective Intelligence

Crunch Lab plans to expand beyond finance and biomedical research. The company did not specify which industries it will target next, but predictive modeling applies broadly. Energy companies forecast demand. Logistics firms optimize routes. Retailers predict inventory needs. Any organization that relies on forecasting could benefit from better models.

\ The Solana Incubator selection adds credibility. Solana is one of the faster blockchains, processing thousands of transactions per second with low fees. Building on Solana allows CrunchDAO to handle high transaction volumes without cost prohibitions. The incubator provides technical support and connections to the Solana ecosystem, which could help with partnerships and integrations.

\ The $5 million round will fund platform development and team expansion. Crunch Lab did not disclose current headcount or hiring plans, but scaling a decentralized network requires engineering resources. The company must build tools for enterprises to submit challenges, manage data encryption, coordinate contributors, aggregate predictions, and monitor performance. Each step involves technical complexity.

\ Competitors exist in this space. Numerai, founded in 2015, operates a similar model for hedge fund predictions. Ocean Protocol builds data marketplaces with blockchain infrastructure. Fetch.ai focuses on autonomous agents for decentralized systems. Crunch Lab differentiates by emphasizing production deployments and measurable results, not just research or speculation.

\

Final Thoughts

Crunch Lab's model offers a practical test of whether decentralized networks can deliver results that matter. The 17% accuracy improvement for ADIA Lab is measurable. The deployment of Mid+One in live trading is real. The Broad Institute breakthrough, while less quantified, comes from a respected research organization. These are not thought experiments or white papers. They are production systems handling consequential problems.

\ The question is whether this approach scales across industries and use cases. Financial modeling and biomedical research are natural fits for crowdsourced AI. Both fields have large datasets, clear performance metrics, and tolerance for experimentation. But will the model work in manufacturing, where failures cause downtime? Or in autonomous vehicles, where mistakes endanger lives? Or in legal analysis, where accountability matters?

\ The $10 million in funding gives Crunch Lab time to find answers. The investor group brings credibility and connections in decentralized infrastructure. The network of 10,000 contributors provides a foundation to build on. But the company must prove it can maintain quality, protect data, navigate regulations, and keep contributors engaged over time.

\ If Crunch Lab succeeds, it could change how enterprises think about AI development. Instead of hoarding talent, organizations could tap into global networks. Instead of building models from scratch, they could access collective intelligence. The shift would redistribute value from centralized teams to decentralized contributors, and from proprietary systems to open protocols.

\ That outcome depends on execution. Crunch Lab has results to point to, funding to deploy, and a network to leverage. Now it must scale without breaking what works.

Don’t forget to like and share the story!

:::tip This author is an independent contributor publishing via our business blogging program. HackerNoon has reviewed the report for quality, but the claims herein belong to the author. #DYO

:::

\