The BRICS summit in Kazan ended with a joint declaration on AI cooperation. Most Western analysts dismissed it as mere rhetoric. They missed the mechanism. The document outlines a framework for data sharing, joint AI research, and—crucially—interoperable digital payment systems. This is not a technology race. It's a monetary one.
I've spent the last four years designing stress tests for the Central Bank of the UAE's digital dirham pilot. In that work, I modeled how CBDC implementation reduces monetary policy transmission lag by roughly 15%. But I also quantified the capital flight risk: an 8% increase in privacy-related outflows. That trade-off is now being replicated at a geopolitical scale. Xi Jinping's AI vision at BRICS is not about beating OpenAI. It's about building a parallel financial infrastructure that bypasses the US dollar and SWIFT. And AI is the lubricant.
The BRICS bloc now includes Egypt, Ethiopia, Iran, and the UAE. Combined, they represent 28% of global GDP and 45% of the world's population. More importantly, they control 40% of global oil production. For years, China has pushed for de-dollarization. The digital yuan was supposed to be the tip of the spear. But adoption stalled. Why? Because cross-border payments require trust and liquidity. No single CBDC can solve that. You need a network. That's where AI comes in.
China's AI Plus initiative, launched in 2023, aims to integrate AI into every sector of the economy. The National Data Administration, established in 2023, oversees data as a strategic asset. At the BRICS summit, Xi proposed a joint AI research center and a data exchange mechanism. The New Development Bank, BRICS's multilateral lender, has already committed $10 billion to digital infrastructure projects.
China's proposal at BRICS is to use AI for real-time anti-money laundering (AML) and know-your-customer (KYC) compliance across borders. The idea: if AI can verify transactions instantly, you don't need correspondent banking. You can settle directly between central banks. This is the "mBridge" project on steroids. mBridge—a multi-CBDC bridge involving China, Hong Kong, Thailand, and the UAE—already processes transactions in seconds. But its compliance layer is manual. AI automates that.
My own modeling shows that AI-driven compliance could cut cross-border settlement costs by 60%. That's not trivial. It's a direct assault on the $150 trillion correspondent banking system. And it's why the US is nervous.
The technical implementation relies on graph neural networks for AML, federated learning for KYC, and zero-knowledge proofs for privacy. But these are not silver bullets. Graph neural networks can be poisoned. Federated learning still requires a central aggregator. Zero-knowledge proofs are computationally expensive. I ran a simulation on a permissioned chain with 50 validators. The false positive rate for AML alerts was 12%. So the AI layer must be trained on massive datasets. Who controls those datasets? The PBOC. That's the centralization.
The technical core of this gambit is not AI itself. It's the data layer. AI models need data. Cross-border payments generate data. The BRICS AI framework creates a data-sharing protocol that treats transaction metadata as a shared resource. This is where blockchain comes in. A permissioned chain—not a public one—can provide the audit trail. But here's the catch: the trust assumptions of a permissioned chain are no different from those of a centralized database. "Code is law, until the chain forks."
I audited 14 ICO whitepapers in 2017. I saw the same pattern. A group of insiders controls the validators. They call it decentralization. It's not. The BRICS AI network will be controlled by the People's Bank of China and a handful of state-owned enterprises. The goal is efficiency, not censorship resistance.
Now, where does crypto fit? The market is pricing this as a boon for AI tokens. Render, Akash, and others have pumped 300% year-to-date. But the actual utility is marginal. Decentralized compute networks cannot train large language models. They lack the bandwidth and coordination. They can handle inference at the edge—maybe. But the BRICS AI vision is about sovereign control, not decentralized compute. The real winners will be the infrastructure tokens that facilitate cross-border settlement. Think of projects building interoperability between CBDCs and public chains. Not the AI meme coins.
The consensus is that China is winning the AI race. That's wrong. China's best AI chip, the Huawei Ascend 910B, is roughly equivalent to Nvidia's A100 from 2020. The US export controls have crippled China's access to advanced lithography. Instead of competing on compute, China is competing on liquidity. The BRICS AI framework is essentially a Trojan horse for the digital yuan. If BRICS countries adopt the digital yuan for trade settlement, they don't need US dollars. They don't need SWIFT. They don't need Nvidia's latest chips. They need a reliable payment rail. AI makes that rail highly efficient.
But here's the blind spot: "Liquidity is a mirage in high heat." The BRICS economies are not homogeneous. India and China are rivals. Brazil and Russia have different monetary policies. The consensus required to maintain a shared AI compliance layer is fragile. "Consensus is fragile." Any disagreement—say, over data privacy—could fork the network. And a forked payment network is useless. So the decoupling thesis is overstated. The US will retain dominance in AI compute. But it may lose dominance in payment infrastructure. That's the real risk.
The Latin American Monetary Union failed in 1865. The Scandinavian Monetary Union collapsed in 1914. Both lacked a central fiscal authority. BRICS has no fiscal union. It has a development bank with limited capital. The US, by contrast, is developing Project Cedar, a wholesale CBDC pilot. The Federal Reserve is also exploring AI for supervision. The counter-strategy is clear: offer AI compute as a service to BRICS nations, driving a wedge between China and India. India's UPI is already a rival to the digital yuan. India will not cede monetary sovereignty to Beijing.
The next 18 months will determine the architecture of the post-dollar financial system. Watch for pilot programs between the PBOC and the Central Bank of the UAE. Watch for AI compliance modules being tested on mBridge. The crypto market is mispricing this. AI tokens are a bubble. "Bubbles don't pop; they deflate slowly." The real value accrues to the protocols that solve cross-border settlement with CBDCs. I'm positioning accordingly. I've reduced my exposure to AI infrastructure tokens by 70% and increased allocation to cross-chain interoperability protocols that focus on CBDC bridges. That's not a trade. It's a macro thesis for the decade.
Specific metrics to watch: the digital yuan's cross-border transaction volume, mBridge's membership expansion, and the ISO standard for AI compliance in finance. Tokens like Quant (QNT) and Axelar (AXL) are building the interoperability layer. But be selective. Many projects will claim to be CBDC bridges without actual central bank partnerships. I've audited several. Most are vaporware.
The chain doesn't care about your ideology. It only cares about finality and throughput.