Aggregator
Wealthy neighbors outraged after Airbnb party erupts in gunfire, wild video shows
'Caution Is Warranted': Ed Dowd Warns Wall Street's AI CapEx Party Is Ending
Authored by Ed Dowd via 'Beyond The Narrative' substack,
The signs are piling up faster than the hype can spin them. AI capex has been the rocket fuel for markets, but the second derivative is turning. Factors ending the party:
-
Private credit stalled — flows reversing, redemptions surging, industry effectively paused.
-
Enterprise demand cracking — ROI skepticism, token costs biting, data/alpha extraction backlash.
-
Power constraints hitting hard — the grid can’t scale without massive, long lead time builds or dystopian reallocation.
-
Open-source pressure — Chinese based DeepSeek and now the new open-source frontier model Kimi K3 are rivaling OpenAI and Anthropic frontiers labs at fraction of the price, commoditizing the economics.
Credit markets always end the party. We’re watching it live.
Private Credit: The Silent Pause Button on AI CapexThis is where financing reality bites. Morgan Stanley estimated private credit could fund up to 50% of the external financing needs for the massive AI data center buildout. That channel is now under serious stress.
Flows in private credit are going the wrong way. The industry is effectively paused. Redemption requests are surging, funds are gating, and high-profile bankruptcies plus underwriting scrutiny are flashing warnings. Private credit has become the new junk bond market…except it lacks transparency, liquidity, and is now being stress-tested in real time.
With outflows accelerating, near-term funding from Private credit for AI data center buildouts looks less likely. NVIDIA and others keep popping up in private credit loan books. What happens to repossessed GPUs in a stressed environment?
Continued capex relies on credit markets keeping the spigot open. When that spigot slows or gets expensive, the capex math breaks. Financing could turn prohibitively costly, pausing or dramatically slowing the cycle that has supported the S&P 500 index with roughly 45% of the market cap being AI or AI-adjacent.
Additionally AI infrastructure inflation itself is credit driven. This reflexive credit driven surge in demand has caused cost inflation for chips and data center construction making the past cost projections moot. These inflated costs make ROI hurdles even harder. Current creditors are reassessing their exposure as the Goldman Sachs credit desk has recently highlighted.
Enterprise Demand Is CrackingThe people who are supposed to use this stuff are slowing down. Companies that rushed AI tools into workers’ hands are now reining them in because costs at scale are biting. That’s not theory…it’s the second derivative showing up in real budgets. They are stepping back and trying to assess the ROI from this investment as their AI budgets come in above initial cost projections.
Alex Karp of Palantir laid it out bluntly. Enterprises are livid. They’re paying for tokens that create no reliable value. They’re watching their own workflows, customer data, and competitive alpha get extracted and potentially sold back to competitors. Token pricing itself is the confession…if these models delivered durable, defensible productivity gains at scale, the labs would price on value or take equity cuts, not meter compute.
Open-source pressureDeepSeek and other cheaper open-source models were already pressuring token pricing for those users who didn’t need the frontier models. To make matters worse, recent newcomer Moonshot AI’s Kimi K3 is a Chinese open-weight model rivaling top US frontier leaders OpenAI and Anthropic at a fraction of the cost. It’s putting direct pressure on closed-source pricing and exposing how over-hyped the token economics have become. Why pay premium rates when open-source alternatives deliver competitive performance? Competitive pricing hurts revenue growth for the two US frontier leaders which is likely to push their IPOs into next year (if at all) and raises the cost of their debt capital. The downstream effect is an eventual capex slowdown for the pick-and-shovel crowd (semiconductors etc.). Simple as that.
Power Constraints: The Physical Wall Nobody Wants to Talk AboutEven if the money were flowing freely, the electricity isn’t. AI compute requires tremendous amounts of energy and water. Currently installed electrical capacity can’t handle the aggressive projections without massive new builds, which take years of capital, permitting, and construction or drastic reallocation of existing power away from other uses.
One path is slow and practical: add real capacity. The other gets dystopian fast: reduce human usage or rolling blackouts to free up juice. The reality will be somewhere in between and will play out politically. Local opposition to data centers is already rising over electricity rates, water use, and land. Towns are pushing back. This isn’t abstract futurism…it’s a hard physical constraint on timelines and costs.
Power adds another multiplier to the capex problem. Data centers need gigawatts. Hyperscalers are already forecasting enormous spending just to keep up. When private credit tightens and power infrastructure lags, the combined effect is a slower, more expensive buildout than the bull-case spreadsheets assume. The second derivative doesn’t just slow…it can stall.
Market Concentration and Cyclical RealitySemiconductor stock valuations have hit record levels recently around 19-20% of the entire S&P 500. That’s not healthy diversification; it’s concentration in a notoriously cyclical industry whose recent boom has been funded by debt to business models that remain unproven at scale.
The biggest AI capex spenders have seen their stocks pull back meaningfully from highs even as broader indices hover near records. ROI concerns are finally showing up in price action. Even the Bank of International Settlements (BIS) has been more sober: ‘AI has boosted confidence via productivity expectations, but it’s also raising job fears, supply bottlenecks, and the risk of overinvestment boom-bust cycles we’ve seen before.’ Power and credit constraints make that overinvestment risk even more acute.
Bottom LineWe’ve seen this movie before. Credit questions profitability first. Physical limits and cheaper open-source alternatives (hello Kimi) force the timeline and pricing reckoning. Circular deals, negative free cash flow, sky-high chip prices, and now power realities all point in the same direction.
The party isn’t over tomorrow but closing time signals are everywhere: semis at peak gross margins, enterprises pausing, private credit tightening, power wall rising, open-source commoditization accelerating. Stock market AI concentration at extremes.
Skepticism isn’t denial of eventual AI value. It’s calling the current valuation and frenzy for what it is…priced for perfection that customers, credit markets, the electrical grid, and open-source competition aren’t delivering.
Watch the flows. Watch power builds. Watch Kimi-style pricing pressure. Watch the second derivative across money, megawatts, and model costs. The unwind in these concentrated, debt-fueled, physically constrained narratives tends to be swift once the marginal equity buyer and/or lender steps away.
Caution is warranted…the bright bar lights are about to be turned on. The distance to a real repricing is shrinking fast. Protect capital. The math doesn’t lie.
Tyler Durden Fri, 07/24/2026 - 15:00SpaceX Reportedly Turns Away Falcon Customers As Starship Gamble Comes Into Focus
Yet another corporate media report based on anonymous sources is likely to draw an immediate response from Elon Musk. He has repeatedly used X this year to challenge reporting on Tesla and SpaceX.
Bloomberg reports that SpaceX has begun turning away satellite operators seeking dedicated Falcon 9 launches after 2028 and is no longer accepting future reservations. The report was based on people familiar with the matter, and the company has not confirmed it.
The report continued:
Engineers at Musk's rocket, satellite and artificial intelligence juggernaut have halted building some non-reusable components for the Falcon family, such as the rocket's massive upper stage, said one of the people.
. . .
SpaceX's plans could change for a number of reasons, including development setbacks with the futuristic Starship vehicle, the people said. The company is likely to still use the Falcon 9 for launches for the Department of Defense and NASA, some of the people said.
If Starship isn't operational by the end of 2028 and Falcon production isn't extended, satellite operators could face a shortage of heavy-lift launch capacity. That means a lot is riding on the mega-rocket Starship getting through the testing phase and achieving commercial viability.
Starship's execution risks weighed on SpaceX shares this week after last week's test-launch delay. Over the past several weeks, more than $1 trillion in market capitalization has been wiped out.
Shares have fallen about 16% below the $135 IPO price in recent days.
The CIO of Tigress Financial Partners noted that the latest Starship launch abort "underscores ongoing execution risk around ramping Starship to high-cadence, reusable operations, and reinforces that repeated delays could push out revenue and margin trajectories."
Could be wrong, but it seems that 4 engines didn’t light causing the abort sequence
Unfortunate scenario, but better safe than sorry! Hoping for another attempt tomorrow pic.twitter.com/o0BQ8Ekt7i
Starship's Thursday launch attempt was postponed due to adverse weather conditions, with another attempt scheduled for later this evening. Check back for updates.
Tyler Durden Fri, 07/24/2026 - 14:45‘Like magic in 10 minutes’ — People say they’ve found a drug-free $12 fix for headaches
Livery cabbie shot in face by bystander after NYC fender-bender: cops, sources
John Mayer is coming to Connecticut’s Mohegan Sun. What do tickets cost?
San Francisco home prices surge to average $1.7M after sinking to record lows — for 1 big reason
California party promoter charged in bizarre attack, robbery of lawmaker at her own home
Florida woman accused of murder was ‘master manipulator’ of alleged accomplices: new arrest reports
Tony Romo arrested and booked for DUI in Milwaukee County
11 indicted in smuggling deaths of 7 migrants who hid inside Texas cargo container
Bunnie Xo speaks out after Jelly Roll’s daughter hints at explosive reason for shock split
Bunnie Xo speaks out after Jelly Roll’s daughter hints at explosive reason for shock split
Melt into this best-selling Boll & Branch Sheet Set for a bedroom luxury boost
Inside Hulk Hogan’s $5 million estate drama — including Mike Tyson’s involvement — a year after wrestler’s death
Inside Hulk Hogan’s $5 million estate drama — including Mike Tyson’s involvement — a year after wrestler’s death
A Quantum Roadmap Would Push Bitcoin Much Higher
Authored by Ciaran Lyons via CoinTelegraph.com,
Bitcoin developers need to swallow their pride and outline a clear plan to harden the blockchain against quantum computing attacks, according to Capriole Investments founder Charles Edwards. He says the day they finally bite the bullet, the price will respond very quickly.
“If the Bitcoin core team says in two or three months: ‘this is our roadmap, we’re gonna solve it in the next two years, these are the rough steps we’ll take,’ that would be amazing news,” Edwards tells Cointelegraph on Trade Secrets.
“I think that would discount a lot of the risk pretty much overnight,” Edwards says.
The question of whether Bitcoin developers should modify the network to make its cryptography quantum-resistant has sparked heated debate within the Bitcoin community, with some arguing that major changes could conflict with Bitcoin’s core ethos. Others claim quantum computers are many years away, and a rushed cure could be worse than the disease.
Charles Edwards says a clear roadmap could push price up “very quickly”Edwards often highlights the risk of quantum computing to Bitcoin to his 132,800 X followers. The fear is that, one day, powerful enough quantum computers could break the cryptography that protects the Bitcoin network and potentially compromise Bitcoin wallets.
The uncertainty has impacted investor sentiment, and some analysts say it has contributed to the downfall in Bitcoin’s price. The world’s largest asset manager, BlackRock recently pointed to quantum computing as a potential long-term risk in materials for spot Bitcoin ETF investors.
However, Edwards says if Bitcoin developers outline a clear roadmap to address the quantum threat, as some other chains have already done, it could send Bitcoin’s price higher “very quickly.”
Source: Charles Edwards
“Double digits probably,” Edwards predicts.
He adds the quantum issue is “somewhat counterintuitively an upside catalyst potential,” because it is currently on the back burner and the Bitcoin Improvement Proposals (BIPs) to date are “not really” a genuine solution.
Edwards is no stranger to making high-conviction calls on Bitcoin. Based in Melbourne, Australia, he founded Capriole Investments in 2019, a hedge fund focused on Bitcoin and digital assets. The firm uses a combination of quantitative models, AI, and macroeconomic analysis to guide its investment strategy across crypto markets.
Charles Edwards says Bitcoin is 40% below its fair valueA growing number of observers worry the risk could become more serious if Bitcoin developers fail to make the necessary changes to the network before 2030. Ethereum is due to complete it’s post quantum overhaul by 2029, which will shine a spotlight on Bitcoin’s own preparations.
Bitcoin is trading at $65,270 at the time of publication. Source: CoinMarketCap
Edwards estimates that Bitcoin is currently around 40% below what he considers its fair value based on energy value, while arguing that quantum risk accounts for roughly a 30% discount. “That means it’s more than priced in,” Edwards said. Bitcoin is trading at $65,270 at the time of publication, roughly 49% below its October all-time highs of $126,100.
Edwards clarifies that Bitcoin’s current price reflects the quantum risk based on the information available today, rather than any unknown future developments that could accelerate the threat and tank the price further.
His estimate is based on the timelines outlined by leading quantum computing companies and researchers for when “Q Day” could arrive, the point at which quantum computers become powerful enough to reverse engineer private keys from public keys.
“That sits in that four to five year range, give or take, a few years,” Edwards says.
Edwards says he also factors in the time Bitcoin would need to develop and implement a solution, which BIP-360 author Ethan Heilman estimates could take years.
“If we’re gonna get into maths, it’s pretty simple; it is just an aggregation of those expert opinions. So it’s based on that, and based on the fact that there’s currently no solution for Bitcoin.”“That risk again falls significantly if there’s a solution or if there’s a roadmap to a solution. But it also could grow if tomorrow we find out that Google is, you know, twice as far ahead on their roadmap to Q Day or some other major company,” he said.
“It’s priced in today, but it’s not to say that it can’t get worse or better. It’s just I think it’s skewed more probabilistically to the upside from here,” Edwards says.
Tyler Durden Fri, 07/24/2026 - 14:20