Post

@KobeissiLetter If you are feeling a bit overwhelmed by the endless scrolling or heavy headlines on the feed today, pause for a second. Left a very smooth reality check for the family right here: x.com/i/status/208690097168732…

@KobeissiLetter duh bro 😎 are you low iq

@KobeissiLetter If that estimate is right, AI has moved from a software story to a macro demand story: data centers, chips, power and construction equal 2.4% of US GDP in capital spending. That can support growth, but it raises the bar for profits when depreciation and power bills hit.

@KobeissiLetter Maybe because AI didn’t exist in any other time in history.

@KobeissiLetter And this is just the reported CapEx—the real multiplier effect on AI infrastructure, chips, and power could push the actual economic footprint even higher. My market view ⬇️⬇️ premarketdesks.com/t/tier1-v1-…

$725 billion. That's what four of these companies alone are on pace to spend on capex in 2026, up 77% from $410 billion in 2025. Analysts already see that topping $1 trillion in 2027, right in line with the jump to 3.1% of GDP. Been mapping out where exactly that money is landing.

@KobeissiLetter ☝️ “Just hold” isn’t enough.

@KobeissiLetter today’s hyperscaler CapEx is largely self funded by enormous free cash flows and tied to real AI compute demand, not debt fueled speculation that produced massive overcapacity.

@KobeissiLetter $AMZN is the best positioned for this: x.com/LunarThinkTrade/status/2…

Comparing Big Tech CapEx to telecom in 1999 is a fair warning, but the revenue profiles are completely different. Cloud and AI have actual paying customers now, unlike dial up speculation. Still, 3.1% of GDP by 2027 means any AI monetization stumble hits $MSFT, $GOOGL, $META, $AMZN, and $ORCL hard. Watch the earnings revisions, not the headline spend.

@KobeissiLetter Oracle is the most undervalued stock right now

@KobeissiLetter what clicked

@KobeissiLetter Really appreciate this insight. Learning the process behind investment decisions is one of the best ways to grow. Thanks for sharing your experience.

@KobeissiLetter The dot-com parallel is fascinating, but today's spend is backed by massive free cash flow instead of risky telecom debt

@KobeissiLetter But what are they working on and why what is the goal?

@KobeissiLetter Good share, thanks. So straightforward and helpful. You and @meghaan_s are my go-to finance accounts.

@KobeissiLetter pop that bubble x.com/Roots19681/status/208691…

@KobeissiLetter Big tech companies are using their balance sheets as a weapon. Increasing AI capital expenditures to 3.1% of US GDP isn't just about building data centers; it's about permanently eliminating potential competition. The market is completely hosting their spending cycle.

@KobeissiLetter The real trade isn’t AI vs. no AI anymore,it’s who turns this enormous capital spending into actual cash flow first. That’s where the winners and losers will emerge.

This is becoming a macro story because AI capex is now large enough to influence the US economy itself. A capex boom of this scale can lift productivity and growth, but it also creates enormous demand for capital. That matters for Treasury yields, credit conditions and ultimately the Fed's reaction function. If AI investment keeps accelerating while inflation remains sticky, the Fed may have less room to ease than equity markets expect. The key isn't whether AI is a bubble. It's whether the productivity gains arrive fast enough to justify the capital being deployed.

@KobeissiLetter That pace of capital deployment has historically preceded major market inflection points, so I’m watching credit spreads and AI monetization signals closely. My market view ⬇️⬇️ premarketdesks.com/t/tier1-v1-…

@KobeissiLetter Spending a whole extra point of GDP just to make chatbots slightly faster.

@KobeissiLetter The scale of AI spending is unprecedented. The real question is whether these investments will generate enough revenue to justify the enormous capital flowing into the sector.

@KobeissiLetter The difference is that this spending is largely coming from free cash flow instead of debt

@KobeissiLetter x.com/i/grok/share/c9cb2df72e7… !!

@KobeissiLetter x.com/i/grok/share/c9cb2df72e7…

@KobeissiLetter There is no meat in that burger. Retail is buying the biggest hoax in stock market history. x.com/reasonus4/status/2086914…

@KobeissiLetter Will the bubble burst or will it not?

AI is no longer just a technology story. It’s becoming a capital spending cycle big enough to influence the entire economy. The key question for investors isn’t whether AI spending keeps growing. It’s whether the returns eventually justify this historic level of investment. That will separate the winners from the hype.

@KobeissiLetter The fact that only a few companies can spend more than 3% of GDP further deepens monopolization in the technology world, leaving small startups and competitors behind in the race.

@KobeissiLetter thats because it takes ever more to prop up the Ponzi

@KobeissiLetter My market view ⬇️⬇️ — these capital outlays are a clear bet on AI demand, but history warns us that record spending often peaks right before the cycle turns. premarketdesks.com/t/tier1-v1-…

This is no longer a technology investment cycle. It is becoming a macroeconomic one. When a handful of companies are spending several percentage points of GDP on data centers, chips, power and infrastructure, their decisions start shaping national growth, energy demand and credit markets. The comparison with the Dot Com era is important, but the bigger question is returns. If AI productivity scales with the spending, this becomes transformational. If monetization lags, the economy could discover it built enormous capacity ahead of demand.

The ratio measures size, not fragility. What broke in 2001 wasn't the level of telecom capex, it was that it was debt-funded by companies without the cash flow to service it. Most of this build is currently funded from operations, and that's the real difference. Whether it stays true is the thing to track. The $500B Nvidia financing consortium reported today is a step in the other direction.

@KobeissiLetter AI isn’t something leaders can just watch from the sidelines anymore. The real question is where it can save time without costing us the human connection that matters most.

@KobeissiLetter The scale is remarkable.

@KobeissiLetter Hyperscaler capex as a share of GDP finally crossing telecom is the whole story.

@KobeissiLetter Its a race to see who takes over the world!

@KobeissiLetter Each rung down, the same GPU is starting to get refinanced at a higher cost of capital; and it’s all in the new-issuance monthly calendar ⤵️

@KobeissiLetter Big Tech capex is doubling to 3.1% of GDP by 2027. My portfolio is projected to remain flat. Diversification

@KobeissiLetter Are you saying you are smarter than big tech?

@KobeissiLetter For context: AMZN options flow (Mon, Aug 10): call-heavy (89%), aggressive sweeps, near-dated; net $5.6M to calls; calls building 10 sessions; biggest $1.1M Wed $277.5C sweep.

@KobeissiLetter Capex booms are usually real. The 1999 fiber build was too — the cable got used, just not by the ones who paid for it. The trend and the returns are separate questions.

@KobeissiLetter x.com/manlikepelumi/status/208…

@KobeissiLetter Boards are requesting excecutives to control AI token costs accross most corporations around the world. The firsts dominos are rolling already.

@KobeissiLetter 这里要先统一口径：五家公司全球资本开支除以美国GDP，不等同于美国境内投资占比。真正决定回报的，是新增算力能否在折旧前转成收入与现金流。

@KobeissiLetter Very good breakdown from Jay Martin on this subject 👇 x.com/JayMartinBC/status/20866…

@KobeissiLetter And this is even before the agentic economy shows up in the numbers. Inference running at scale changes the demand curve entirely.

@KobeissiLetter the bigger the buildout gets, the less room there is for demand to disappoint

@KobeissiLetter Telecom in 1999 borrowed to lay fiber nobody leased yet. Big Tech's paying for AI capex out of cash flow, so there's no refinancing wall to trigger a bust, just a depreciation schedule that needs revenue to keep pace.