I’m speaking in hindsight, but hindsight is a great way to learn lessons for the future.
Just remember: when gold was running, every single analyst on Wall Street was telling you $6,000, $7,000, even $8,000 gold. When crude oil was at $55-$60, every single one of them was telling you it’s going below $50.
Nobody knows anything.
All predictions are essentially worthless and probably cause more damage than benefit. The reason people love predictions is because they make us feel better. We wake up in the morning, hear someone tell us what will happen next, and get the comforting illusion that everything is under control.
It rarely is.
Marimaca Copper: Pampa Medina Sulphide Drilling Intersects 16m of 5.70% Cu and 62.6 g/t Ag including 4m of 11.32% Cu and 144.0 g/t Ag in Significant Extension of UHG Bornite Zone $MARI.TO https://t.co/hDuqFmXFIr
Some conversations stay with you long after you've had them. That's exactly how it felt interviewing @michaelxpettis for @CurrenPower with @Nicolas_Colin.
A sample of what we covered in our 2h conversation.
What stops a Chinese household from consuming more and what stops an American household from saving more? How sustainable is China's over-investment? Why infrastructure is a cost of growth rather than a source of it? And more on deindustrialisation, the dollar, stablecoins, and who bears the cost when the imbalances unwind.
A fascinating conversation, which hopefully will leave you as it left us: coming back to it, and further thinking about it.
https://t.co/QdpR9P3qhF
Nice plug from Capstone here #mintwit!
If $CS Santo Domingo's location is good then the MOD's location represents a near-unrivaled infrastructural advantage in the region. https://t.co/w76YLWr4Rb Marimaca #copper.
Nice to be the benchmark by which others are messured!
El Nino update. Godzilla is coming.
Most of the world will see top quartile temperatures over Jul/Aug/Sep, which is peak vegetative growth season globally for grains, half the world oilseeds and also for sugar and rice in Asia. That means more evapotranspiration and stress on crops.
India and South-East Asia will also likely see their weakeast monsoon rains in decades.
This is happening while there's a shortage of fertilizers.
"If you exclude semiconductors and AI servers, Taiwanese exports have actually fallen by 40% since 2022. In South Korea, non-AI exports have stagnated and Japan’s industry is in decline." 1/5
https://t.co/v7R4m3lKKL
1/5
Very good Caixin article on the struggle to manage local-government debt: "With financing squeezed, local governments are turning to a new strategy: revitalizing state-owned resources, assets and funds. Championed by Hubei and Hunan, the idea is to...
https://t.co/Qwb607wqcm
LunR Royalties Completes Acquisition of the Fruta Del Norte Silver Stream and Receives Conditional Approval to Graduate to the TSX.
https://t.co/6wSFHMnWxL
$LUNR.V
#reachforthemoon
In the last 30 days alone:
– Microsoft cancelled most of its Claude Code licenses, citing cost
– Uber burned through its entire 2026 AI budget in 4 months
– Uber's COO publicly said AI costs are "harder to justify"
– A Fortune 20 CEO ordered token spending to be "dramatically slashed"
– One company spent $500M in a single month on Claude with no usage limits
– H200 rental prices crashed from $7/hr to $4/hr in three weeks
The stocks: still at all-time highs.
How long can the market keep ignoring the receipts?
DeepSeek just popped the American AI bubble.
Not by killing AI.
By killing the fantasy of unlimited AI pricing power.
DeepSeek V4 Pro:
Input: $0.435 per 1M tokens
Output: $0.87 per 1M tokens
OpenAI GPT-5.5:
Input: $5.00
Output: $30.00
Claude Opus 4.7:
Input: $5.00
Output: $25.00
Claude Sonnet 4.6:
Input: $3.00
Output: $15.00
DeepSeek is roughly:
11.5x cheaper than GPT-5.5 on input
34.5x cheaper than GPT-5.5 on output
28.7x cheaper than Claude Opus on output
17.2x cheaper than Claude Sonnet on output
If a model is “good enough” at 1/20th or 1/30th the cost, margins will compress faster than Wall Street expects.
AI is not dead.
But the AI bubble just lost its pricing power.
$NVDA $MU $AMD $SMCI $AVGO $ARM $MSFT $GOOGL $META $AMZN $ORCL $PLTR $CRWV $NBIS
#AI #DeepSeek #AIBubble #ArtificialIntelligence #Semiconductors #HBM #DRAM #Nvidia #Micron #AMD #TechStocks #Nasdaq #StockMarket #Investing #WallStreet #YapayZeka #Borsa
Chinese AI models are 90% to 95% cheaper than the American AI bubble models.
And people still don’t understand what is happening.
Look at DeepSeek, Kimi, Qwen, MiniMax and the rest of China’s AI stack.
They are cheap, fast and useful enough for most real-world use cases.
The market is pricing American AI as if pricing power will last forever.
It won’t.
The AI bubble is not being killed by lack of demand.
It is being killed by collapsing cost curves.
$NVDA $AMD $AVGO $TSM $ASML $MU $MSFT $GOOGL $META $AMZN $BABA $BIDU