Ideas
Bitcoin still looks weak, needs time.
Crypto looks pretty terrible, and Bitcoin still needs time rather than just more downside. He could still see a move toward 48k and thinks a price bottom is more of a time issue. He is not taking new longs, risk, or leverage in crypto and equities broadly.
Buy Neoclouds on cheaper AI compute demand.
He has been adding high-beta AI exposure through Neoclouds such as Nebius, CoreWeave, and IREN. Data centers are likely to remain scarce, and Neoclouds have an easier business to understand and benefit regardless of which model wins because the cost of intelligence is falling and demand for compute is rising. CoreWeave is a levered bet on data centers and AI demand with risk, including roughly 65% of revenue tied to OpenAI, so it should not be oversized, but he is bullish and has started buying.
Hold Mag 7 for multi-year AI upside.
He remains mostly in boring Mag 7 exposure, which has been relatively unbothered by the AI selloff. He is holding for a few years and believes the AI trade has much higher to go as it climbs the wall of worry.
Neoclouds benefit from accelerating compute demand.
He has become more bullish on the Neocloud bet because AI compute demand should accelerate, open-source models and falling cost per unit of intelligence should expand use cases, and the long tail of users is still early. He wants exposure to the compute buildout via Neoclouds.
Buy US tech dip, own Mag 7.
He has been long US tech for years and remains overtly bullish. Forced selling or deleveraging with an intact thesis is a golden buy-the-dip opportunity, and he thinks the market overreacted to Kimi 3. AI demand potential is much larger than today, and cheaper intelligence should open new use cases, so he prefers owning a basket of Mag 7 and AI names rather than being concentrated.
Small higher-beta memory sentiment plays.
He owns small positions in Micron and SanDisk as higher-beta AI sentiment plays. They are not his strongest convictions, but he does not want to be on the sidelines and prefers owning the basket.
Avoid altcoins; AI offers better volatility.
Altcoins have lost their trading appeal because AI stocks like Micron and DRAM offer far more volatility and daily movement. Most crypto speculators have moved to AI names, leaving little activity in altcoins beyond the top names.
Prefer AI infrastructure over applications.
AI infrastructure has been the easier way to get paid in the current upswing: application revenue outside OpenAI and Anthropic is tiny while infrastructure revenue is hundreds of billions, and serving models at speed and cost is still hard enough to earn margin. The application layer is harder to underwrite because serving incremental users has real marginal costs, so he prefers infrastructure over applications.
Buy HYPE, add lower on RWA perps.
He is a buyer of HYPE and would buy more if it drops another 10-15% or below 50. He sees the trend toward HIP-3 markets and trading tokenized stocks onchain with perps as clear, with potential KYC-gated US access via allowlists and market makers bridging permissioned and permissionless liquidity. He views HYPE as one of the few compelling secular crypto assets to own.
MetaDAO ICO model aligns incentives.
He is bullish on the MetaDAO ecosystem because its ICO model incentivizes founders to raise at low valuations and aligns founders with investor performance; this offers exposure to a variety of early projects at attractive valuations, and MetaDAO has good momentum.
QQQ trend has rolled over.
Equity indices are rolling over on his Guppy moving-average trend filter, especially the Q's, which look really bad. He prefers not to be long when short-term moving averages cross below long-term ones.
ETH showing relative strength, not buying.
ETH has been looking the best among major coins on his charts, though he is still not buying it. It is a relative-strength watch rather than a clean long.
Dollar looks strong on trend filter.
The dollar index looks great on his Guppy moving-average trend filter, one of the few things that looks strong while crypto and equity indices roll over.
Watch Zcash Ironwood security upgrade.
Zcash launched the Ironwood pool, a formally verified version of the patched Orchard technology. Orchard is now withdraw-only, and as supply migrates to Ironwood, the amount vulnerable to any potential Orchard exploit falls. The migration is proceeding slowly, which he views as a good sign, and the upcoming Tachion pool is more scalable and formally verified. The market seems not to expect an exploit, though ZEC has been weak since the upgrade.
Avoid long-term Treasuries as yields rise.
Global government bond yields are breaking out as countries run sizable fiscal deficits and need more capital for defense, war, and reshoring. Debt becomes the answer and pressures the long end. He is glad he is not sitting in long-term Treasuries.
Speculative FWA gacha token gamble.
He owns FWA and has been playing the gacha game aggressively, but calls it straight-up gambling. The token can only be acquired by playing the game while external buys are off; heavy daily token incentives are running out and external buys or buybacks may turn on soon. The platform is expanding beyond NFTs into ERC20s and collectibles, and if it gains real markets and momentum it could become much bigger, though the FWA bag could go to zero.
This Delphi Digital video, published July 30, 2026,
features Jason Calacanis, Ceteris, Kevin Kelly, Jose
discussing BTC, Neoclouds, NEBIUS, CoreWeave, IREN, MAGS, US Tech, MU, SNDK, ALTCOINS, AIQ, HYPE, METADAO, QQQ, ETH, DXY, ZEC, TLT, FWA.
16 trade ideas extracted by AI with direction and confidence scoring.
Speakers:
Jason Calacanis,
Ceteris,
Kevin Kelly,
Jose
· Tickers:
BTC,
Neoclouds,
NEBIUS,
CoreWeave,
IREN,
MAGS,
US Tech,
MU,
SNDK,
ALTCOINS,
AIQ,
HYPE,
METADAO,
QQQ,
ETH,
DXY,
ZEC,
TLT,
FWA
