Last week, a headline crossed my screen that pulled me away from my audit of an L2's sequencer upgrade: TCS, the $150 billion IT services behemoth, is hiring 8,900 ‘AI deployment engineers’ and actively scouting for acquisitions. My first reaction was not to envy their hiring budget, but to check the ledger of assumptions baked into this move. As a smart contract auditor who spent 2017 parsing the ethical edge cases of ERC-20 token standards, I learned that scale often masks centralization. TCS is not building a blockchain; they are building a fortress. But this fortress has implications for every decentralized technology that relies on trustless execution, from DeFi protocols to decentralized AI networks. The question we must ask: does the TCS model represent the future of AI infrastructure, or is it a centralized antithesis to the values crypto espouses?
Tracing the moral code behind every token.
To understand why this matters in a blockchain context, we first need to decode what TCS is actually doing. The hiring of 8,900 deployment engineers signals a shift from AI model development to AI model deployment at industrial scale. TCS, along with cousins like Infosys and Accenture, has long profited from being the ‘last mile’ integrator for enterprise technologies. Now they are applying that playbook to AI. The acquisitions target small players with vertical AI solutions—likely in banking, insurance, or retail. The goal: create a walled garden where TCS controls the pipeline from model selection to production monitoring. The data they gather from these deployments will feed their own proprietary models, creating a flywheel that strengthens their hold. For those of us who have built open educational platforms like ‘The Open Ledger’ in Nairobi, this feels like watching someone build a library and then charge admission to read the books.
Building libraries where others build empires.
Now, the core analysis: how does this centralized AI deployment strategy clash with the decentralized ethos of blockchain? I will draw on three technical observations from my 27 years in the industry—15 of them specifically in crypto auditing and education.
Observation One: The Oracle Problem Redux. In DeFi, we know that oracle feed latency is the Achilles' heel. Chainlink tries to solve decentralization of data with a network of nodes, but the nodes themselves often depend on centralized infrastructure. TCS's deployment engineers will rely on centralized cloud APIs—AWS, Azure, GCP—to serve AI inferences to enterprise customers. This creates a single point of failure not just in data freshness, but in governance. If TCS decides to change a pricing model or an algorithm, every client is forced to comply. I saw this same dynamic during the ZEIP-20 audit: a proposed ERC-20 upgrade that gave more control to the default admin wallet. Just because a centralized entity can scale quickly doesn't mean the system is robust. For blockchain-based AI models (like those on Bittensor or Render Network), the promise is that inference is verified on-chain and not dependent on a single corporate roadmap. TCS's hiring spree is a bet that enterprises prefer the convenience of a single throat to choke over the resilience of a decentralized network. In a bull market, convenience often wins; but during a crash, resilience matters.
Observation Two: The Data Flywheel vs. Data Sovereignty. TCS's model captures immense amounts of enterprise data. With 8,900 engineers facilitating deployments, they will see the raw inputs and outputs of AI models across thousands of clients. This data is gold for fine-tuning vertical models. But it also means that the enterprise loses sovereignty over its data. Compare this to a blockchain-based AI deployment where the model runs on user-controlled hardware and only aggregated results are shared publicly. In my work with the Savanna Voices NFT collective, we structured a DAO where artists retained ownership of their royalties. The same principle should apply to data: users should not hand over their business logic to a centralized integrator. TCS's strategy accelerates the opposite trend—further entrenching data oligopolies. For blockchain advocates, this is a call to build better privacy-preserving AI infrastructure.
Observation Three: The ‘Last Mile’ Decentralization Gap. Over 90% of current AI workloads run on centralized cloud infrastructure. TCS is doubling down on that reality. But the crypto ecosystem has been building alternatives: decentralized compute platforms like Akash Network, Filecoin's compute layer, and zero-knowledge proof verifiers. The problem is that these platforms are still clunky for enterprise deployment; they lack the service-level agreements and handholding that TCS offers. The 8,900 engineers are essentially a human bridge between the AI model and the corporate client. In crypto, we often assume that code replaces trust, but for enterprise adoption, trust still requires humans—auditors, lawyers, deployment engineers. TCS is capturing that trust premium. The contrarian angle? Perhaps this is a necessary phase. Just as centralized exchanges like Coinbase paved the way for institutional crypto adoption, centralized AI deployment by TCS can educate enterprises about the need for later decentralization. But I am skeptical of that narrative—it echoes the ‘regulate now, decentralize later’ fallacy that has plagued DeFi.
Community over capital, always.
Let me step back. I've spent the last year co-authoring the African AI-Blockchain Ethics Charter, a 50-page framework that balances innovation with social protection. One of our core principles is that technology must serve human dignity—not just capital efficiency. TCS's approach is capital-efficient: it centralizes control to maximize profit and lock-in contracts. But it is not dignity-preserving, because it denies the end-user any say in the governance of the AI systems they depend on. In crypto, we face a similar tension. Many DeFi protocols are governed by a handful of multi-sig signers. The ‘code is law’ maxim fails when those signers collude or are coerced. TCS's centralized deployment is the traditional IT version of multi-sig governance: comfortable for the few, brittle for the many.
Now, the contrarian angle that might surprise: TCS's hiring binge could actually be bullish for decentralized AI networks—if they fail to deliver trust. History shows that centralized solutions often become bloated, expensive, and resistant to change. The 2008 financial crisis birthed Bitcoin because centralized banks failed. Similarly, if TCS struggles to manage 8,900 new hires—leading to integration delays, security breaches, or model bias scandals—enterprises will seek alternatives. That alternative could be blockchain-based AI deployment, which offers immutability, auditability, and user control. I've seen this pattern in NFTs: OpenSea's royalty surrender killed the creator economy, but it also spawned a wave of decentralized marketplaces that respect creator terms. TCS's centralized empire might inadvertently create the same push for decentralized AI tools.
Walking away from the hype to find the soul.
Yet I must temper that optimism with a dose of reality. The crypto industry has a habit of over-promising: we said decentralized storage would replace AWS by 2020, we said DeFi would replace banks by 2022. Those shifts are happening, but slower than the hype cycle predicted. TCS is a 50-year-old company with deep enterprise relationships. Their 8,900 engineers will be deployed to clients who already trust them with core banking systems. That trust will not be eroded by a whitepaper alone. It will require crypto-native AI deployment platforms to deliver the same level of handholding—perhaps through decentralized DAOs that provide support via reputation tokens or on-chain service SLAs. This is a huge opportunity for projects like Pocket Network or Chainlink, which are already bridging centralized and decentralized infrastructure.
As a final takeaway, let me reflect on my survival of the 2022 bear market. When my educational platform lost 60% of its donations, I chose to downsize and focus on open-source curriculum rather than chasing VC funding. That taught me that authenticity is maintained by consistency, not by scale. TCS is building at scale, and scale can be impressive. But in blockchain, we are not building empires; we are building libraries—open, accessible, and resilient. The real question is whether the crypto community will invest in the deployment layer with the same fervor we have invested in the consensus layer. If we don't, we risk ceding the most valuable part of the AI stack to centralized giants. The moral code behind every token should also guide how we deploy the intelligence that tokens govern.