Our proposal was straightforward. Use Blockchain technology to create transparency, through smart contracts that would enable fair pricing, and to bridge the gap between rural farms and premium organic buyers in Europe. Looking back from 2026, we now live in a world where we see a pattern repeating across agricultural traceability projects worldwide. The technology works in pilot projects, but scaling to real-world adoption remains frustratingly elusive.

Blockchain Success Stories: What Works

In a world of mistrust especially regarding food, Walmart partnered with IBM to trace mangoes. They have reported that they managed to reduce the trace back time from 7 days to 2.2 seconds. The solution was built on Hyperledger Fabric, an open-source digital ledger technology running on IBM Cloud. Started off as an answer to ensure food safety issues, Walmart piloted the project as an answer to reduce trace time in case of a food poisoning outbreak might occur. As of 2026, the company is further augmenting this blockchain foundation with Agentic AI and automated supply chain centers to predict demand and reduce waste at scale.

In another part of the world, TraceX in India works with thousands of farmers, partnering with major brands like MTR Foods and Blue Tokai Coffee. They were able to expand the pilot project to 4000 farmers in Araku Valley. TraceX has expanded the same blockchain architecture in to other industries as well; poultry, milk and seed traceability.

These success stories have a few thing in common

1. Strong anchor buyers - Walmart funded its own initiative driven by food safety concerns, while TraceX secured funding from major brand partnerships

2. Strong consumer demands - Walmart consumers and the other luxury brand consumers.

3. Strong participation throughout the supply chain - Every stakeholder in the assignment understood the incentive on having a clear, transparent, immutable record system.

4. Robust technical support - TraceX raised their own capital to develop the technology and IBM came through with a strong system

Can it fail? Yes. But replicating these conditions is exactly where most projects struggled.

Challenges all the way

There are three main challenges that we faced during our implementation. And same can be said for the many blockchain integration in to supply chain.

1. The Participation Paradox

Very early on we encountered an issue. The wholesale buyers of smallholder farmers, who have a stake in the organic agriculture value chain did not want to participate in the shared database. Although simulations were run to counterpoint and show the security of their data, they strongly disagreed to be part of the project. From their point of view, in a red ocean this is not paranoia, this is experience. This review article in ScienceDirect also mention lack of standardization and participation barriers are the primary challenges in implementing digital traceability.

Blockchain relies on network participation. This cannot be achieved if there is competitor mistrust.

2. Missing links and data errors

In our project, digital literacy was not an issue, however there was reluctance in entering data. They didn’t see the incentive of putting extra effort. We prepared to delegate the work to agriculture extension officers. This led to discovery of another critical vulnerability; blockchain is only as trustworthy as the person entering data at the source.

In many cases not so thrilled small holder farmers, suppliers, to technically un-savvy stakeholders in a value-chain could be costly (onboarding and training) and cumbersome. Despite their success, Walmart faced the same challenge.

3. Disconnection in sharing the value

Price premium was our hook to get the stakeholders involved. Traceability complemented the organic certification process that we had begun several years back. Therefore we were fortunate. But this could easily be a challenge for many other projects. Premiums existed only in urban niche markets, requiring long-term investment beyond certain project timelines.

This pattern extends beyond organic agriculture. According to a ScienceDirect Article of Aquaculture and Blockchain
The majority of the blockchain use cases are for traceability and storytelling, and to a lesser extent for payments or incentives

There is still work to do

For blockchain technology this is just the start. We have success stories and we have lessons that we learnt. I remain convinced blockchain offers the best path forward for food supply chain transparency and equitable value distribution. But technology alone isn't the answer.

How national policy can help?

Our project revealed a critical policy gap: the absence of a national organic standard meant blockchain could verify data integrity but couldn't establish certification legitimacy. Throughout our programme period there was no centralised national organic standard. We saw although Blockchain could verify data integrity but couldn't establish what organic certification actually meant or who had authority to grant it.

What could the policy environment could look like?

1. Clear certification frameworks: National standards defining what blockchain verifies

2. Incentive alignment: Price premiums, market access, or reduced costs rewarding participation

3. Digital infrastructure: Basic connectivity before demanding real-time IoT data

4. Capacity building: Ongoing support for extension officers bridging the digital divide

The question for policymakers in emerging markets isn't whether to adopt blockchain. The focus should be on foundational policies, infrastructure, and incentive structures exist to make adoption viable beyond pilots. Because without those foundations, we're just creating more expensive ways to confirm what we already knew: the technology works beautifully, but the ecosystem isn't ready.

As for a way forward, pilot projects serve their purpose: they teach us, we iterate, we improve. But what all these pilots have taught us is one thing: for blockchain technology to truly achieve its potential in agriculture, we need strong policy foundations that make scalability possible, not just more proof-of-concepts.

Glossary

Agentic AI - autonomous systems that go beyond generating content to proactively set goals, plan, and execute complex, multi-step tasks with minimal human oversight

Hyperledger Fabric - permissioned distributed ledger technology (DLT) platform designed for enterprise use

Sources

• Walmart bets big on AI - Talk Business, February 2026

• TraceX Case Studies

• Blockchain in Agriculture - Farmonaut, June 2025

• Walmart blockchain challenges - PYMNTS, December 2022

• Aquaculture blockchain use cases - ScienceDirect

• Blockchain traceability barriers - ScienceDirect, November 2024

And here we are now, in 2026 with over 19 million AI models, over 8 AI model types, and over 378 million users in 2025. To prevent harms that AI can bring, we need to address advanced cyberattacks (like phishing and malware), AI-powered misinformation and deepfakes, algorithmic bias, and data privacy violations. And on the other hand we need to encourage growth, healthy competition, right for information and innovations. There for more than a decade countries are coming up with AI policies, regulations and act.

For this piece since I am living in the EU, and since US is where almost all the prominent AI companies are developed, I will discuss my perspective on EU AI strategy and US approach to AI.

They both are trying to control. One takes the view of cautious, and careful adoption to AI, where human rights should be valued. The other gives a prominence to innovate and create very powerful AI projects. One values caution, the other innovation.

I've been reading : the EU's AI strategy (2018), and now the simplified version on the EU’s Digital Omnibus(2025) , and the various US regulatory approaches, Biden executive order to foster the AI environment to Trump administration opening up to sell AI hardware to international buyers. I am trying to understand what's actually happening. And I keep noticing that the good, the bad, the ugly, and most importantly the USEFUL to unearth some complicated underneath.

What are They Actually Trying to Do?

What is important is to understand the logic behind what they are saying.

The EU's framework is built on a specific threat model. A concentrated power in AI systems can make high stack decisions about people's lives. Such as manipulation and deception, exploitation of vulnerabilities, and social scoring The EU looked at various scenarios and decided to establish the rules before deployment. This is what they call "ex-ante enforcement."

The US approach reflects a different bet: AI is new and growing really fast. Predictions are uncertain and the perspective rules can be outdated quickly. Therefore let’s foster theenvironment and hold companies accountable after deployment through litigation, Federal Trade Commission enforcement, liability. This is "ex-post enforcement."

Here is an analogy. Let’s say you are driving a car, WITHOUT a seatbelt. In the EU approach law enforcement stops you and asked you to wear the seat belt and fines you in case of accident you are unharmed. That is Ex-ante enforcement. In the US approach you never ran in to the law enforcement. You hit a tree and you get injured. The law enforcement will charge you for reckless driving, and you end up in a hospital with a hospital bill. That is ex-post enforcement.

So when people say "Europe is restrictive," what they actually mean is: "Europe prioritizes preventing harm before it happens." And when they say "the US is permissive," they mean: "The US relies on catching and punishing harms after they happen."

These aren't just different policies. They're fundamentally different BETS about what we can prevent versus what we can control through enforcement.

Are they really that different?

Here's where it gets interesting—and messy.

The US doesn't actually have "light-touch" regulation, even if it looks that way.

Yes, there's no comprehensive federal AI law like the EU's. But the regulatory intensity is there. It is distributed. And it is actively watched by man regulatory guidians. The FTC has become aggressively litigious on AI—investigating algorithmic discrimination, deceptive practices, all of it. And the there is the State of California. The California Privacy Rights Act (CPRA) has granted that effective from July 1, 2023, the users have the right to opt-out of automated profiling and, in some cases, the right to access information about the logic behind AI. On top of all that executive orders on AI safety, procurement standards, algorithmic accountability, the list adds up. The regulatory machinery exists but it's just operating through different channels.

Meanwhile the EU's AI Act, while comprehensive, isn't uniformly strict. It has safe harbours. It has exemptions for research. It has a simpler version for SMEs and SMCs. Has a innovation package for SMEs and Startups. It recognises that a recommender algorithm isn't the same risk as a hiring system.

The timeline is revealing: I asked Claude to map out when each started thinking about AI policy. The US began about 10 years ago. The EU, 8 years ago. Yet the EU moved to comprehensive regulation faster.

Are there any Real Trade-offs ?

And interestingly, while the EU has been tightening rules, the US has been loosening them. During the Biden administration and also with Trump administration, we see deals between OpenAI, Anthropic along with other companies,(Safety and Testing Agreements (Aug 29, 2024),Department of Defense (DoD) Contracts (June/July 2025), Anthropic Partnership (Nov 2024),The General Services Administration (GSA) agreement (August 2025). This signals a government fostering the AI ecosystem. Further sell AI hardware internationally is not allowed. Meanwhile, the EU raided Grok's headquarters, considered banning social media platforms, and simultaneously pushed for developing its own competitive AI platforms.

What are the actual trade-offs?

Speed of Innovation: The US approach lets companies move faster. OpenAI, Anthropic, and others have bloomed under this environment. They're everywhere, and they deliver products quickly. In 2024 Mario Draghi through his report on EU competitiveness pointed how EU has a competitiveness deficit due to its administrative burdens and regulatory inconsistencies. 2024-2029 mandate the Commission pledged to simplify.

Prevention vs punishment: EU invest heavily on prevention. Risk assessments, guidelines, code of practice, that all should follow. It is expensive and requires heavy institutional involvement. US choose punishment. Punishment requires accountability mechanisms that actually work. Involvement of courts, legal systems, and companies solvent enough to pay damages. Both have gaps and comes with tax.

Who pays for all this? EU compliance hits the smaller innovators hardest. So does in US, small innovators will have to defend themselves. Either way small guy has to pay, the giants will have teams and capital to work for them.

I have worked in the SME sector development as a project implementor and my past and current research focus a lot on the BoP, SME and Startups. So the answer for the last question is something I am passionate about.

And to expand more, what about the emerging markets, where 80% of the SMEs in the world are operating? These economies can't easily adopt the EU model where they lack institutional capacity to enforce complex rules and regulations. They can't fully adopt the US model either. They lack litigation infrastructure and capital abundance. (Currently, I am doing a whole policy paper on this for an emerging market with some tweaks).

To conclude, in the outset the two approaches look very different, but for some questions they have the same approach. In the whole debate the real questions are taking a back step.

Few articles to get some critique on the different approaches

Europe loosens reins on AI – and US takes them off

Europe Tech Regulation: Yes to Simplification, No to Deregulation

he Global Race to Regulate AI: Biden’s Executive Order Spillover Effects on the EU AI Act

Direct links to acts

EU's AI Act

Federal Trade Commission - AI

Biden’s Executive Order

Trump's Executive Order