CZ has warned that artificial intelligence will keep growing exponentially, but most companies building in the sector may still fail.
The Binance founder compared the setup to earlier technology cycles, where a fast-growing industry creates major winners, brutal failures, and constant rotation between old leaders and new entrants. His point was not that AI is fading. It was that too many companies are chasing the same opportunity at the same time.
That distinction matters as AI investment keeps accelerating. Anthropic just raised one of the largest private AI rounds on record, while OpenAI, xAI, Google, Meta, Microsoft, and dozens of smaller startups continue competing for users, compute, talent, and enterprise budgets.
The Crypto Comparison Fits The Market Cycle
CZ’s warning lands because crypto went through the same pattern. Bitcoin and Ethereum survived multiple shakeouts, but thousands of exchanges, tokens, mining companies, DeFi apps, NFT platforms, and infrastructure startups disappeared after raising capital into crowded markets.
AI could follow a similar path. The technology may become more important while many companies built around it still fail. In that kind of cycle, the winners can become much larger, but weak products, copycat apps, expensive infrastructure bets, and businesses without real demand get exposed quickly.
That is already visible in crypto’s AI niche. AI agents, smart wallets, automated trading tools, and DeFAI apps are expanding, but the category also carries serious execution risk. Products such as Binance’s keyless Agentic Wallet show where useful AI automation may go, while the wider market is still full of projects using AI branding without durable utility.
Failure Data Backs The Caution
The broader enterprise data supports a more sober view of the AI boom. RAND research has highlighted estimates that more than 80% of AI projects fail, often because companies misunderstand the business problem, lack usable data, or cannot integrate systems into real workflows.
The MIT GenAI Divide report found that 95% of organizations saw no measurable return from generative AI pilots, while only a small group of integrated deployments created meaningful financial impact. The problem is not simply model quality. Many companies still struggle to turn AI tools into products that users keep using and businesses keep paying for.
Early-stage funding is also getting harder for generic AI startups. Some investors now warn that 2026 could punish companies building small variations of the same chatbot, workflow assistant, or automation layer without a clear moat.
AI Winners May Still Get Much Bigger
CZ’s warning is bearish on crowded AI companies, not on AI itself. The same cycle can produce huge survivors. In crypto, market crashes did not kill the sector; they concentrated liquidity, users, developers, and institutional attention around stronger networks and platforms.
AI may see the same split. Companies with real distribution, proprietary data, strong infrastructure access, useful agents, enterprise retention, or deep product integration may keep growing. The weaker layer of copycat apps and capital-heavy startups could disappear as funding tightens and customers demand measurable value.
That makes the next phase less about AI hype and more about survivability. The market is not asking whether AI matters. It is starting to ask which companies can still matter after the shakeout.
Be the first to comment