Jesus Martinez: TAO Is Waking Up… Here’s What I’m Watching

In this video, Jesus Martinez walks through his current TAO trading position, his broader outlook for the AI crypto cycle, and a few notable developments across the Bittensor subnet ecosystem.

Trading position & exit plan

• He’s currently up on a leveraged TAO long opened about a week ago, and plans to ladder out rather than sell all at once (e.g., 20% at $360, another 20% at $400, more at $600, and so on).
• The full leverage position will be closed before the end of May, regardless. He doesn’t like holding leverage into the summer.

Market outlook

• He expects a big rally in May, followed by a steep summer correction into sideways action, with all-time highs likely in September–October. This is based on historical volume patterns where summer months tend to taper off.
• On the spot side, he’s planning to hold through the entire summer, even through a potential 50% correction, because he believes the AI crypto niche has bottomed and the risk/reward is strongly skewed up over the next 6–12 months.

Profit-taking philosophy

• He’s mostly allocated in spot and proud to be “a bit of a boomer” about it.
• Key principle: take profit on the way up, even small amounts. If TAO hits $5,000, more than half of his bag gets sold. The point is to secure life-changing money rather than diamond-hand into the next cycle’s drawdown.
• He recently de-risked his Zcash position for the same reason. Still bullish, but it’s run a lot and isn’t his #1 narrative.

Subnet news worth flagging

• Targon (Subnet 4) launched its Supply Portal, letting compute suppliers easily onboard idle hardware and start monetizing it, aiming for 100% utilization for data centers and compute providers.
• Venice AI‘s Uncensored 1.2 model was built using compute from Targon. Venice is a $700M token, while Targon sits at just $81M. A striking valuation gap given Targon is powering some of Venice’s core models.
• Arbus on Subnet 97 (Distill) released a 4B parameter model scoring higher on human evaluation, accessible via Open WebUI. Rewards on Subnet 97 go to top performers based on model quality, knowledge sharing, challenge wins, and active participation.

Portfolio structure

• He’s now mostly allocated in root (plain TAO) rather than individual subnets (roughly 80–90% TAO, 10% subnets). The logic: TAO is the safer long-term hold, while subnets are still AI startups finding their footing, and concentrated subnet risk doesn’t let him sleep well at night.
• He notes the subnet market has been quiet since the Subnet 3 rally, but volatility is returning, which makes them tradable again. He plans to keep adding selectively.

If you want the full breakdown with charts and commentary, the video is here: