The intersection of cryptocurrencies and artificial intelligence (AI) has entered a quieter, more selective phase, two prominent venture capitalists say.
Canonical Crypto’s Anand Iyer and Spartan Group’s Kelvin Koh describe the current environment as a post-hype moment for decentralized AI protocols, with capital and talent shifting toward more focused, utility-driven applications during Hong Kong Consensus 2026.
“I think we’re at the bottom right now,” said Iyer, whose San Francisco company supports early-stage infrastructure and applications based on decentralized networks. “We’ve been through a bubble period. Now it’s about figuring out where the real power is.”
Both Iyer and Koh have criticized what they see as overinvestment in the GPU market and attempts to build decentralized alternatives for large-scale artificial intelligence models such as OpenAI or Anthropic. Koh noted that the funds required are “day and night” compared to what is available in cryptocurrencies.
Instead, they see the potential in purpose-built, full-stack solutions and tools that start with a specific problem and build all the way to the model, compute, and data layers.
Iyer noted that startups are skipping expensive SaaS tools and using artificial intelligence to build custom in-house systems in days. “Speculation will no longer drive product,” he said. “We have to think about the user first.”
Both investors highlighted the importance of proprietary data, regulatory advantages or go-to-market advantages as moats for new forms of competition.
For founders looking to raise capital, Koh has blunt advice: “Twelve months ago, having a wrapper on ChatGPT was enough. But that’s no longer the case.”