Elon Musk just annihilated the memory bears.
Elon Musk on SpaceX earnings call: “Limiting factor currently is memory. Memory output is increasing by around 20% per year. Now normally that would be fantastically fast and amazing for any large mature industry. But ask yourself is the demand increasing by 20% a year? No, the demand is increasing by 200% a year, maybe higher. So if you've got demand increasing much faster than supply then Economics 101 would suggest that the price increases. It does not decrease.”
The chart is one of the clearest rebuttals to the claim that AI compute is already being overbuilt.
GPU rental prices are not collapsing as new capacity comes online. They are rising across almost every generation. B200 rates have moved toward roughly $5.50 to $5.80 per GPU-hour, H200 rates are approaching $3, while even older H100 and A100 capacity continues to reprice upward.
The most important signal is not the absolute premium on B200. It is that older chips are also becoming more expensive. If this were merely temporary excitement around the newest architecture, B200 prices could remain high while H100 and A100 rates declined as customers substituted toward newer hardware. Instead, demand appears broad enough to absorb new systems without materially weakening pricing for the installed base. That suggests the industry is constrained by total usable compute, not simply by access to the latest chip.
This supports the argument that hyperscalers and neoclouds are currently under-earning on older contracts. Customers that secured GPU capacity in 2024 or 2025 at fixed rates are benefiting from pricing materially below today’s spot market. As those contracts expire, providers should have room to reprice capacity upward, improving cloud revenue and operating cash flow without requiring an equivalent increase in physical infrastructure.
It also strengthens the Jevons paradox thesis. Chips and models are becoming more efficient, but lower cost per task is expanding the number of economically viable workloads. Enterprises are deploying more coding agents, research agents, video models, longer-context inference and autonomous workflows. Efficiency per token is improving, yet total compute consumption is rising faster.
The investment implications remain constructive for $NVDA, HBM, networking, optical connectivity, cooling and power infrastructure. Rising rental prices indicate that the marginal value of energized compute remains high, which encourages operators to keep ordering equipment and accelerating data-centre deployment.
However, investors should not assume that high rental prices automatically produce high returns for every neocloud. Realized economics still depend on utilization, financing costs, electricity, networking quality, customer concentration and hardware depreciation. A provider renting B200s at $5.60 per hour can still destroy equity if the machines sit idle, were financed expensively or require rapid replacement.
The chart covers non-hyperscaler cloud providers, where spot pricing can be more volatile and scarcity premiums greater than in long-term enterprise contracts. It should therefore not be interpreted as the average rate across the entire market. But the direction is difficult to dismiss.
The AI overcapacity thesis requires supply to begin outrunning demand. For now, the opposite appears to be happening. More GPUs are entering the market, yet the price of renting them continues to rise. The industry is not struggling to find workloads for its compute. It is still struggling to find enough compute for its workloads.
Ken Griffin strikes so fast. Reminder that the day Enron filed for bankruptcy, he chartered a jet with 16 Citadel employees to Houston and “all we did was interview people at Enron for several days” to study energy business.
Citadel hired Enron quant research team while UBS bought rest of business. Team built Citadel’s commodity trading and made $30B+ so far (UBS shut down the Enron business it bought).
Travis is an apex predator put on ice in 2017 and only recently reactivated from cryogenic slumber.
Standards for male behavior have moved so far that he's actively perplexed by it.
Our current ecology was last home to performative frat bros nearly a decade ago.
So when Travis tells TBPN he jet skis to work, and Jordi + John ask whether he does his best thinking on the jet ski, he pauses.
You can visibly see him get confused, with an expression like "What the fuck are you guys talking about?"
Then he quite literally voices his frustration: "No guys...it's a fucking notch on the belt. Who else do you know who jet skis to work? Nobody. I'm the guy."
Even the literal technology brothers are incapable of fathoming a man who does things just for the bit.
Travis in 2026 is like the T-rex from Jurassic Park 2 who finds himself in modern day San Diego being like, yo, what the FUCK is this place?
And then just stomps everything in sight.
https://t.co/yv6Y3TJ4Co
🏴🤣 John Stones pretends his shoulder is hurt then begins to dance in the England changing room. 🕺
Tuchel was genuinely worried at one point that the injury was real. 😅
Ah yes, “global banks pulling cash out of Indonesia.” Terrifying. Except one of them left in 2023 and the other one is currently being fought over by three buyers. Some exodus.
Let’s be precise about what actually happened. Citi didn’t flee Prabowo. Citi sold its Indonesia consumer business to UOB and finished the migration in November 2023, part of a pre-planned retreat from consumer banking across 14 markets worldwide that was announced back in 2022, long before this administration existed. It walked away with about $1.1bn in regulatory capital benefit across the four ASEAN markets. The institutional business stayed.
HSBC is also the dane. It’s reviewing retail operations across Australia, Indonesia, Sri Lanka and more under a simplification plan announced October 2024, and is exploring a sale of the Indonesia retail unit, not bolting for the exit. DBS, SMFG and others are reportedly lining up for that book.
Funny kind of capital flight when there’s a bidding war for the asset. That same report notes foreign banks have been showing growing interest in Indonesia on the back of solid growth.
The real story is duller and more honest: profit repatriation timing. The corporate and institutional franchises stay put, which is the part that actually matters for capital markets. The remittance number and rupiah weakness are real. Pinning two pre-planned consumer exits on Prabowo is just reverse engineering a scary narrative from a soft currency chart.
Happy Monday to my fellow Indonesians who get bashed every Monday morning by Bloomberg 🤣
Imagine replacing 90% of your employees with a team of geniuses who have no idea how your company operates.
Total chaos. Nothing works.
That’s what AI feels like today.
The missing piece is extracting all the domain knowledge from people’s heads and providing that as structured context to the models.
we can all make pinanfarina slop.
we can all take the model trained on a million ferrari photos and make an image that looks like a ferrari.
we can all one-shot a better looking car than what jony ive just came up with.
i get the logic. look here, with close to zero effort i can conjure a thing of objective beauty.
but with even less effort than that, you can simply pull up an old photo of a 250 GTO and say, "this is beautiful."
there's something deeply insidious about LLM "creation":
the danger the luce portends, which i don't see abating anytime soon--is that in the domains in which we cannot create, have no notion of what it takes to create, creation now seems within our grasp.
we have been granted a tool whose output is by definition maximally derivative, yet convinces us we're each an inventor
If you pitched this as a screenplay, every studio would reject it for being too on-the-nose.
Fernando Mendoza was a 2-star recruit ranked 2,149th in his high school class. Zero FBS scholarship offers. Not one. He walked on at Cal, fought for a starting job, transferred to Indiana for his senior year, then led them to 16-0 and the first national title in school history. Heisman, Walter Camp, Maxwell, Davey O'Brien, Manning, Big Ten MVP. 41 TDs, 72% completion, 8-to-0 TD-to-INT ratio in the playoffs.
The Raiders took him #1 overall Thursday night. $54.56M fully guaranteed. Only the third player ever to win the Heisman, win a national championship, and go first overall the next spring. Burrow. Newton. Mendoza.
Then he skipped Pittsburgh.
The biggest stage in football, the moment every kid imagines from the second they pick up a ball, and Fernando watched the call from his living room in Florida because his mom Elsa is in a wheelchair and the travel is hard for her. She was diagnosed with MS when he was 4. She wrote a letter to her sons in The Players Tribune in 2015 promising the disease "won't affect us in the ways that matter."
The part nobody talks about: while every other top pick was on stage, Fernando announced the Mendoza Family Fund the same day. $500K personal donation to the National MS Society. Committed to raising $1M over three years. He hasn't taken an NFL snap and he's already given more to a cause than most players donate in a full career.
He and his brother Alberto have already raised $360K through the Mendoza Bros. Burger at BuffaLouie's in Bloomington. At Christmas, he handed four families dealing with MS $10,000 each for an Adidas shopping spree.
Both his parents are children of Cuban refugees who fled Castro. His dad rowed at Brown, won a Junior World Championship in 1987, and played high school football in Miami next to a teammate named Mario Cristobal. Fernando beat his dad's old teammate in the national championship game in January.
Every athlete talks about playing for their family. Fernando actually did it.
The slow mo, the first touch, the song, the fans reacting before the goal, the shot & how everyone just watches the ball go in
Would you take this upgraded Scott McTominay back at Manchester United? Yeah, I’d gladly take him back.
Indonesia's economy is genuinely unique. It doesn't move with the world the way most countries do.
Global economy crashes, Indonesia stays put. Global economy rockets, Indonesia can look flat. It runs on its own rhythm.
COVID 2020, global demand collapsed, oil went negative, trade volumes cratered. Indonesia's GDP dipped just 2%. Bounced back within a year.
Russia-Ukraine 2022, inflation ran hot everywhere, rate hikes hammered emerging markets. Indonesia ran a current account surplus. Coal and CPO exports were booming while the rest of the world was bleeding.
Same story now with Hormuz.
Your brain at 2 AM writing a paper you started at 10 PM is operating in a neurochemical state that most productivity systems spend thousands of dollars trying to replicate.
Sleep deprivation suppresses your prefrontal cortex. That's the region responsible for self-criticism, second-guessing, and the voice that says "this paragraph isn't good enough." At 2 AM, that voice goes quiet. Not because you've achieved some zen state. Because the hardware running it is shutting down for the night and you won't let it.
Meanwhile the deadline is dumping norepinephrine and cortisol into your system, which narrows your attention to a single point. Your brain physically cannot multitask in that state. No checking your phone. No opening a new tab. The stress response has commandeered every available resource and pointed it at the Google Doc.
Lowered inhibition plus chemically forced single-task focus. That combination is almost identical to what Csikszentmihalyi documented across 30 years of flow state research. Clear goal, immediate feedback, challenge matched to skill. A 12-page paper due in 8 hours hits all three criteria by accident.
The lo-fi beats matter more than people think. Repetitive audio at 60-70 BPM synchronizes with resting heart rate and suppresses novelty-seeking circuits. You stop hearing it within minutes. It becomes an auditory wall that blocks interruption without costing you any cognitive load. It's the cheapest sensory deprivation chamber ever built.
And the black coffee at midnight is pharmacologically different from your morning cup. Your adenosine levels have been building all day, so the caffeine is fighting a much stronger sleep signal. The subjective experience of "wired but calm" at 1 AM is a different drug interaction than alert-at-9-AM. Same molecule, completely different neurochemical environment.
Every semester, twice a semester, four years straight. That's 40 sessions of accidental deep work before anyone had a name for it.
The grade was an A- because the conditions were perfect. Not despite the chaos. Because of it.