Meta is weak below its 200-day moving average because Q2 2026 earnings growth, revenue growth, and net margins are slowing, while CAPEX fears keep pressure on META, GOOGL, and TSLA. The core crypto view is that AGI can build a better Bitcoin but cannot replace Bitcoin’s Lindy, trust, and Schelling-point properties, so BTC stays the monetary insurance asset while SOL looks like the dominant L1 by fees and transactions and Base has not overtaken it. The portfolio message is allocation-first: don’t chase narratives like AI-agent transactions, don’t sell BTC at the bottom, and size higher-conviction names like SpaceX and MU around the upcoming window where summer weakness and September seasonality create entries.
META is below its 200-day moving average, stuck in a compressed 4-zone chopping area, with weakness tied to slowing earnings growth and revenue that is flat to down depending on the quarter.
META net margins are down hard in Q2 2026 and actual earnings are down, with the chart said to mirror the fundamentals directly.
CAPEX fear is the cited driver behind weakness in META, TSLA, and GOOGL; GOOGL is tagged for $341B of spend by 2028 and $302B in 2027, while META is tagged for $217B in 2028, $196B in 2027, and $139B this year.
GOOGL is framed as able to earn a return on AI spend, while AAPL is described as not playing the CAPEX game and as being "cooked" if it does not build.
META’s 12-month target is 753, about 27–28% upside from the current price; the analyst high is $1,000, the low is $580, and the buy zone is 525–530.
Buying META near 525–530 is expected to allow an easy rebound to 660–680 if earnings growth improves and CAPEX can earn a return; the metaverse is cited as the warning case, after roughly $100B spent generated nothing.
AGI can build a better BTC — faster, cheaper, more energy efficient — but cannot rewrite BTC’s Schelling point, Lindy effect, or trust layer, so BTC’s product is trust rather than code elegance.
BTC is held as insurance against fiat going to zero: it is scarce, the network is protected by energy cost and years of rule agreement, and AI agents needing finality and fast payments are not expected to use BTC for payments because it is too slow.
99.8% of cryptos are called garbage, and there has been no widely deployed consensus change since Taproot in 2021.
SOL is still the favored L1 if one chain wins: the notes say not to buy tokens just because AI agents will need transactions; instead, watch real-world assets, tokenized equities, stablecoin flow, and DEX activity, where SOL is said to dominate.
Base has not taken over: in the cited X402 for AI agents chart it beat SOL in transaction volume, but the broader read from transaction count, daily active users, and market-cap-per-transaction metrics is that Base is not winning.
The raw activity numbers cited show SOL at 184.7M transactions yesterday, Base at 9.5M, and ETH at 3.2M; SOL is said to handle 92% of all transactions, while fees on the 12th were 708,000 for SOL, 264 for ETH, and 51K for Base.
SOL’s thesis is unchanged and the allocation rule is to "don’t chase replace"; the warning is not to rotate into a token on AI-agent hype alone.
BTC is framed as near the bottom of the bear at 59K–60K, with maybe 45–50 days left, and the 2-year path is put around $120K before the next halving, about 87–89% upside; the instruction is to stack BTC, not sell it to chase a move.
SpaceX is pitched as the faster horse versus BTC for the next year: ARR is said to be nearing $100B in September, the upside is estimated at 65% over 1 year, and a move to $600 is described as a possible 4x over a short window.
MicroStrategy is in a "weird box": Saylor is trying to get STRC back to 100 by raising fiat cash to support 3 years of dividends, but the criticism is that selling BTC at the bottom is wrong; a return to 100 is said to depend on BTC ripping through 75K, 85K, 95K, and back above 100K.
Druckenmiller is cited as buying a large amount of TSLA calls after a 30% return on capital for 30 years without a negative year, while MU is called an easy double from 800 to 1600 because memory demand is still rising.