Barry Silbert, founder of Digital Currency Group (DCG), is making a major shift in his investment strategy by backing the Bittensor AI token ($TAO) as the next major cryptocurrency play. According to reports from U.Today, Silbert, whose companies have been investing in Bittensor since 2021, openly calls it the most exciting project since Bitcoin. In his view, Bitcoin has already moved into the category of mature defensive assets and is no longer capable of delivering 100x–500x returns. Silbert is reproducing his early Bitcoin and Zcash strategies, building commercial infrastructure around Bittensor with the goal of monetizing distributed intelligence.
Cardano founder Charles Hoskinson predicts that blockchain technology will eventually absorb artificial intelligence, following a pattern he witnessed in his own career. Speaking on the Deeptech Insights podcast, Hoskinson drew a direct parallel to his experience with cryptography, noting that cryptographers once resisted any association with cryptocurrency until crypto's money became powerful enough to hire away the field's best talent. According to reports from the podcast, Hoskinson expects AI to follow a similar arc within five to ten years.
Hoskinson identified a fundamental economic challenge driving his prediction: data center spending keeps growing tenfold year after year, yet the electricity grid simply cannot scale at that pace. As reported by the podcast, he noted that labs like OpenAI and Anthropic still need to turn a profit eventually, and the enormous cost of pretraining new models keeps narrowing the path to profitability. The Cardano founder expects this economic pressure to force AI companies to seek more efficient solutions.
Hoskinson outlined specific areas where blockchain technology could address AI's limitations, including payments, alignment, and data provenance. According to the podcast, his alignment argument centers on governance, noting that individual AI companies currently set their own rules on acceptable behavior and free speech, whereas a blockchain-based system could establish shared standards among participants. The same infrastructure could also track content origin and automate royalty payments when AI systems draw on someone else's work.
Rather than building more data centers, Hoskinson floated an alternative approach: pooling everyday phones and GPUs into a distributed training network. As reported by the podcast, he compared today's AI buildout to the late-1990s fiber optic boom, when roughly 90% of newly laid cable sat idle for nearly a decade. Hoskinson suggested that similar patterns could follow with data centers, shifting toward smaller models running locally on devices like Apple's M5 Mac Studio, with cryptocurrency serving as the coordination layer.
Hoskinson also predicted that the CLARITY Act will not clear Congress until 2029, citing three specific missteps by the Trump administration. According to the podcast, he blamed the administration for tying crypto's public image too closely to Trump-branded tokens and predicted that political pressure would force a heavily unfavorable bill including ethics provisions. His prediction came just a day after the Senate failed to advance the CLARITY Act on September 15, falling short of the 60 votes required.
