Dear Dr. Burry @michaeljburry , This question has been debated by the investment community and it deserves a data driven answer rather than a rhetorical one. The more important question is not when CapEx ends, but whether the incremental capital being deployed is producing higher operating cash flow, stronger earnings power, and durable strategic advantages. If it is, then the spending is not inherently a problem, it is an investment cycle. The "accounting tricks" question is important and I'd genuinely like to engage with you on this topic. You allude to companies "tortuously adjusting earnings in sinister ways" but you don't specify what you mean. As a shareholder who has analyzed these balance sheets at length, I want to understand what I may be missing. Are you referring to capitalization of software development costs, depreciation schedule adjustments, or lease accounting? These are legitimate areas to evaluate but without specific datapoints it doesn't give shareholders or the market anything actionable to analyze. The meat of your post was “When does it End?” The reality is that it ends when demand for compute stops growing. These aren't speculative bets on a technology that might work. MSFT, GOOGL, AMZN, and META are deploying capital in response to existing paying customer demand that is currently outpacing supply. That's not a bubble dynamic that's a capacity constraint. The spending slows when supply meets demand, not when someone rings a bell. I have tremendous respect for your track record but the data here suggests that the CEOs who built the largest, most profitable enterprises in human history are deploying capital to not just meet demand but drive future profitability for shareholders. The burden of proof for "sinister" accounting sits with the claim not the companies. I'm genuinely open to being wrong but here is the data I am looking at. I am openly inviting you to have an intellectual friendly conversation about this with me anytime, publicly or privately. Let me walk through the five companies most central to this conversation: $AAPL, $AMZN, $META, $MSFT, and $GOOGL. As a disclaimer for everyone who is unfamiliar with corporate filings $MSFT and $AAPL do not report on a calendar year so I will be using the TTM data for them to incorporate the latest quarterly reports. The range of data that will be used is fiscal year 2021 – TTM. AAPL · $62.47 billion CapEx allocated 2021 – TTM · Annual Capex Growth 9.59% ($1.06 billion) · CapEx spent in the TTM $12.15 billion MSFT · $264.74 billion CapEx allocated 2021 – TTM · Annual Capex Growth 302.94% ($62.47 billion) · CapEx spent in the TTM $83.09 billion AMZN · $392.25 billion CapEx allocated 2021 – TTM · Annual Capex Growth 115.91% ($70.77 billion) · CapEx spent in the TTM $131.82 billion GOOGL · $232.39 billion CapEx allocated 2021 – TTM · Annual Capex Growth 271.13% ($66.81 billion) · CapEx spent in the TTM $91.45 billion META · $183.87 billion CapEx allocated 2021 – TTM · Annual Capex Growth 272.88% ($51 billion) · CapEx spent in the TTM $69.69 billion The reason this is significant is because AAPL isn’t expanding their CapEx allocation at similar rates as MSFT, GOOGL, AMZN, and META and we see a direct correlation on what happens to profitability when companies reinvest in themselves rather than prioritizing buybacks and dividends. Here is the data regarding how much AAPL, META, AMZN, MSFT and GOOGL have increased their Cash from Operations over the same time period. AAPL · 2021 Cash from Ops $104.04 billion · TTM Cash from Ops $135.47 billion · Cash from Ops growth 30.21% ($31.43 billion) MSFT · 2021 Cash from Ops $76.74 billion · TTM Cash from Ops $160.51 billion · Cash from Ops growth 109.16% ($83.77 billion) AMZN · 2021 Cash from Ops $46.33 billion · TTM Cash from Ops $139.51 billion · Cash from Ops growth 201.15% ($93.19 billion) GOOGL · 2021 Cash from Ops $91.65 billion · TTM Cash from Ops $164.71 billion · Cash from Ops growth 79.72% ($73.06 billion) META · 2021 Cash from Ops $57.68 billion · TTM Cash from Ops $115.80 billion · Cash from Ops growth 100.75% ($58.12 billion) There is a direct correlation between increased CapEx and increased profitability. AAPL has increased their CapEx since the 2021 fiscal year by 9.59% and their cash from operations has grown by 30.21%. In the 2021 fiscal year AAPL was the only company that generated more than $100 billion in cash from operations. Now MSFT, GOOGL, AMZN, and META all generate over $100 billion in cash from operations and MSFT, GOOGL, AMZN have surpassed AAPL. 2021 Cash from Ops · AAPL $104.04 billion · GOOGL $91.65 billion · MSFT $76.74 billion · META $57.68 billion · AMZN $46.33 billion TTM Cash from Ops · GOOGL $164.71 billion · MSFT $160.51 billion · AMZN $139.51 billion · AAPL $135.47 billion · META $115.80 billion By AAPL not allocating at similar rates to their CapEx they have fallen behind in their profitability growth and this can’t be denied. If I use EBITDA also the same pattern holds up: AAPL · 2021 EBITDA $120.23 billion · TTM EBITDA $152.90 billion · EBITDA growth 27.17% ($32.67 billion) MSFT · 2021 EBITDA $80.82 billion · TTM EBITDA $175.26 billion · EBITDA growth 116.86% ($94.44 billion) AMZN · 2021 EBITDA $59.31 billion · TTM EBITDA $145.73 billion · EBITDA growth 145.70% ($86.42 billion) GOOGL · 2021 EBITDA $88.99 billion · TTM EBITDA $150.18 billion · EBITDA growth 68.76% ($61.19 billion) META · 2021 EBITDA $54.72 billion · TTM EBITDA $101.89 billion · EBITDA growth 86.21% ($47.17 billion) Here are the condensed versions of this data from 2021 to today: AAPL · $1.06 billion CapEx Growth · $32.67 billion EBITDA Growth · $31.43 billion Cash from Ops Growth MSFT · $62.47 billion CapEx Growth · $94.44 billion EBITDA Growth · $83.77 billion Cash from Ops Growth AMZN · $70.77 billion CapEx Growth · $86.42 billion EBITDA Growth · $93.19 billion Cash from Ops Growth GOOGL · $66.81 billion CapEx Growth · $61.19 billion EBITDA Growth · $73.06 billion Cash from Ops Growth META · $51.00 billion CapEx Growth · $47.17 billion EBITDA Growth · $58.12 billion Cash from Ops Growth These companies have the cash flow and cash on hand to fund this. Whether they are borrowing or not doesn’t matter. The way the balance sheets stand AMZN GOOGL MSFT and META are in positive net liquidity positions. GOOGL · Cash and short-term investments $126.84 billion · Long-term investments $68.68 billion · Long-term debt $46.55 billion · Net Liquidity $148.98 billion MSFT · Cash and short-term investments $89.46 billion · Long-term investments $21.20 billion · Long-term debt $35.38 billion · Net Liquidity $75.29 billion AMZN · Cash and short-term investments $123.03 billion · Long-term investments $19.56 billion · Long-term debt $73.45 billion · Net Liquidity $69.14 billion META · Cash and short-term investments $81.59 billion · Long-term investments $27.53 billion · Long-term debt $58.74 billion · Net Liquidity $50.38 billion In my opinion the data leads to one conclusion. These companies are not burning cash recklessly. Every dollar of incremental CapEx deployed since 2021 has been met with multiple dollars of incremental operating cash flow and EBITDA. That is not the signature of a bubble. It’s the signature of infrastructure that is working. To your broader concern about when it ends the reality is that the companies closest to the demand signal will know before it shows up in the data available to the rest of us. When MSFT, GOOGL, AMZN, and META start to scale back on their CapEx allocations it will be the signal that everyone is waiting for. Dr. Burry you were right about 2008 when almost no one else was and no matter what happens you will always be a legend in the investment community for it. The level of credibility that you have amassed from your previous predictions is exactly why specificity matters here. Broad warnings without solid evidence don't protect investors as they just create noise. If there are accounting irregularities worth exposing please name them and I’ll look at every number with you. That is a conversation that would be worth having a public forum about. @amitisinvesting @FunOfInvesting @RealMattMoney @Futurenvesting @KrisPatel99 @Sam_Badawi @Kross_Roads

Dear AI bears: this is not the dot-com bubble 2.0 Why today’s AI build-out looks nothing like 1999 and what that means for investors Every time the Nasdaq makes new highs, bears compare the AI revolution to the dot-com bubble. It’s an easy comparison, but if you actually look under the hood, it doesn’t hold up. Yes, there are speculative corners within AI and some of those stocks will get crushed but that can be said about any market. The core of today’s AI build-out is being driven by some of the most profitable companies in history, funding long-duration infrastructure out of cash flow—not fragile startups living off endless equity raises. If you want to understand whether this is really 1999 again, you have to start by remembering what dot-com actually was. What really drove the dot-com bubble The 1990s were a genuine technological revolution. The internet went from an academic network to a mainstream platform thanks to browsers like Mosaic and Netscape. Between 1995 and 2000, the market surged 400–500%, and suddenly anything with a “.com” in its name could command a multi-billion-dollar valuation. The problem wasn’t the underlying technology. It was the way capital was deployed on top of it. Many of the high-flying names had no profitability, minimal revenue, and in some cases literally no revenue at all. Traditional valuation metrics were discarded and the mantra became scale first, worry about profits later. Investors convinced themselves that the internet would create winner-take-all markets, so any price could be justified for the perceived winner. Venture capital firms raised enormous funds and had to put money to work. A competitive race among VCs led to faster dealmaking, thinner diligence, and ever-higher valuations. The playbook was simple and reckless: spend aggressively on marketing, undercut prices, and acquire users at any cost. Burning cash wasn’t a red flag; it was almost a badge of honor. The assumption was that once dominance was achieved, profitability would magically follow. Heading into 2000cmany dot-com companies had deeply negative operating cash flow and no credible path to positive earnings. Their survival depended on an endless stream of fresh capital which included venture funding before an IPO, secondary offerings and convertible debt after the IPO. That made the whole ecosystem overly sensitive to shifts in sentiment and macro conditions. A fragile business model + extreme valuation is the exact recipe for a bubble. How the dot-com bubble actually popped Bubbles can survive a long time before it bursts. In 1999–2000, several catalysts hit almost at once. The Federal Reserve raised rates by 175 basis points between June 1999 and May 2000, increasing discount rates and making safer assets more attractive relative to unprofitable dot-com companies. The Nasdaq peaked on March 10, 2000, at 5,048. Soon after, Japan slipped back into recession, rattling global markets. MicroStrategy’s accounting restatement highlighted how aggressive some earnings claims had become. Barron’s ran its now-famous “Burning Up” piece, pointing out that many dot-com companies were literally going to run out of cash. At the same time, the Microsoft antitrust case reminded the market that even dominant tech firms operated under regulatory risk. When the mood shifted, the IPO window slammed shut. Secondary offerings dried up. Venture capital became more selective. Companies that had been reliant on constant capital infusions suddenly had no room to operate. The Nasdaq fell roughly 78% from its March 2000 peak to its October 2002 low. Leverage made everything worse. Many investors had bought hot tech names on margin, and as prices fell, margin calls forced them to sell into a declining market. Hedge funds running leveraged momentum strategies were forced to unwind. The result was not a gentle re-pricing; it was a cascade of forced selling. On top of that, the telecom and infrastructure side of the boom turned out just as fragile. Telecom operators had borrowed heavily to build out networks. When demand and pricing disappointed, they were left with excess capacity and crushing debt loads. WorldCom, Global Crossing, NorthPoint and others went bankrupt, wiping out equity holders and inflicting major losses on creditors. That feedback loop between collapsing equity values and credit stress deepened the downturn and amplified the dot-com crash. If you’re trying to map that entire setup onto 2025 AI, you’re starting from the wrong base case. Why the AI build-out is fundamentally different The single biggest difference between the dot-com era and now is simple: who is doing the spending, and how they’re funding it. The core of the AI boom is not a swarm of tiny, unprofitable companies betting everything on a single idea. It’s a handful of massive, hyper-profitable platforms including $MSFT $AMZN $GOOGL $META deploying record capex into cloud and AI infrastructure which is largely being funded by internally generated cash flow. If you want to stress-test the AI story, don’t look at the tiny speculative names first. Look at MSFT, AMZN, GOOGL, and META. Their earnings reports tell you more about the durability of this cycle than any meme stock catching an elevated valuation. In the most recent quarter (Q3 2025 for AMZN/GOOGL/META, Q1 for Microsoft), these three cloud/AI giants MSFT, AMZN, and GOOGL generated about $97.3 billion in revenue from their cloud/AI divisions alone. Collectively, they reported roughly $755 billion in backlog tied to these businesses: approximately $400 billion at Microsoft, $200 billion at Amazon, and $155 billion at Alphabet. Backlog is critical as it’s not capex. It’s contracted or committed customer demand dollars customers are planning to spend on these services over time. That demand is what’s driving these companies to commit unprecedented sums of capital to data centers, networking, and AI infrastructure. The big four by the numbers You don’t have to guess about demand; you can see it in the operating results. I believe that the bears who are trying to correlate the AI revolution to the dot-com bubble haven’t gone through the latest earnings reports which reveal just how much demand there really is. MSFT: Microsoft Cloud revenue reached $49.1 billion in Q3, up 26% year-over-year. The Intelligent Cloud segment, which includes Azure, generated $30.9 billion in revenue, up 28%, and is now running at roughly a $123 billion annualized rate with $13.4 billion in quarterly operating income. Commercial remaining performance obligation (RPO) has grown more than 50% to nearly $400 billion, with a weighted average contract duration of just two years. On the product side, Microsoft is using Azure AI Foundry to help customers build AI applications and agents. Foundry already has tens of thousands of customers, including the vast majority of the Fortune 500. Across its ecosystem, Microsoft has hundreds of millions of monthly active users of AI features, with first-party Copilots representing a rapidly growing share. GOOGL: Google Cloud revenue grew 34% year-over-year to $15.2 billion, and cloud backlog increased to roughly $155 billion, up sharply from a year ago. Google has signed more billion-dollar cloud deals in the first nine months of 2025 than in the previous two years combined. Its Gemini models are handling massive volumes of tokens via API and consumer interfaces, with the Gemini app and AI Overviews touching hundreds of millions to billions of users. Enterprise adoption is also ramping, with millions of paid seats across a broad base of companies. Amazon (AWS): AWS revenue increased 20% year-over-year to $33.0 billion and is now operating at an annualized run rate of roughly $132 billion. Over the past 12 months, AWS has added several gigawatts of power capacity which is more than any other cloud provider and plans to double total power capacity by the end of 2027. Backlog has climbed to around $200 billion. AWS launched AgentCore, a set of infrastructure building blocks for secure, scalable AI agents, and has already seen hefty developer adoption. On the retail side, Amazon is deploying AI throughout the business from tools that help millions of sellers create better product listings to Rufus, its AI shopping assistant. Rufus now serves hundreds of millions of customers and is on track to deliver billions in incremental annualized sales. Meta: Meta is a reminder that AI isn’t just about cloud and infrastructure it’s also about monetization on top of those platforms. The annualized revenue run rate flowing through Meta’s AI-powered ad tools has already surpassed $60 billion. Meta is working on unifying three major AI systems which include Facebook, Instagram, and ads recommendations into a single model architecture (Lattice). More than a billion people are already interacting with Meta AI features across the family of apps. These are not companies telling “maybe someday” stories. They are reporting real revenue, operating income, and contracted demand tied directly to AI and cloud. Follow the cash, not the hot takes The strongest argument against “AI = dot-com 2.0” isn’t just the top-line growth, it’s the cash flow. Look at the core financials of these four businesses over the trailing twelve months: Together, MSFT, AMZN, GOOGL, and META have generated roughly $1.54 trillion in revenue, $549.71 billion in EBITDA, $536.73 billion in cash from operations, and $206.97 billion in free cash flow (FCF). Over that same period, they spent $329.76 billion on capex and still produced over $200 billion in free cash flow after that spending. In 2024 alone, they spent $217.27 billion on capex, an increase of about $77 billion (55%) year-over-year. Over the trailing twelve months, capex climbed to $329.76 billion, up another $112.49 billion (about 52%) above 2024 levels. Based on recent management commentary, combined capex for these four is expected to reach roughly $354.55 billion in 2025 which is more than 60% above 2024 levels. Bears look at those capex numbers and see a bubble. They try to map today’s spending directly onto the reckless telecom/infrastructure build-out of the late 1990s. The problem with that analogy is that this spending is not being funded by speculative venture debt or serial equity offerings. It’s being funded by rising operating cash flow at four of the most profitable companies in the world. This is not one highly leveraged player betting the farm on a single unproven idea. This is a group of trillion-dollar platforms, with diversified revenue streams and fortress balance sheets, responding to clear, measurable customer demand in cloud and AI. Ask yourself: what are the odds that the bears sitting on the sidelines have a better read on optimal data center spend than Andy Jassy, Mark Zuckerberg, Sundar Pichai, and Satya Nadella? The valuation puzzle If this really were a full-blown bubble, you’d expect the core AI infrastructure names to be trading at nosebleed valuations that imply perfection for a decade. That’s not what the market is actually pricing. Based on forward EPS estimates, META, GOOGL, MSFT, $NVDA, AMZN, $AVGO, and $AAPL do not look wildly expensive relative to high-quality non-tech staples. In fact, when you look at the 2027 estimates, many of these AI leaders trade at lower forward P/Es than retailers like $WMT and $COST. The market is comfortable assigning more than 30x earnings to companies like COST and WMT, while assigning under 25x 2027 earnings to companies like META, GOOGL, MSFT, NVDA, and AMZN that are sitting at the center of the AI and cloud build-out. You can reasonably argue about whether all those earnings estimates will be hit but calling that setup a “bubble” stretches the word beyond meaning. For instance, AMZN is expected to generate $7.17 of EPS in 2025 and $9.53 in 2027 which is an increase of 32.91% over the next 2-years and trades at 24.63 times 2027 earnings. Does that seem unreasonable considering WMT has 25.67% EPS growth and COST has 21.24% EPS growth over the same period yet trade at 31.24 and 37.87 times 2027 earnings? NVDA has 78.98% of projected EPS growth over the same period and trades at 23.61 times 2027 earnings. When I look at the projections from The street the argument that we are in an AI bubble loses credibility. Where the real bubble risk actually is There are absolutely pockets of the market which that look and feel expensive. We are seeing companies with tiny revenue bases, no profits, enormous TAM slides receiving sky-high valuations as investors pile into them. Some of those names are going to experience brutal round-trips. But that’s very than the AI cycle being equivalent of a dot-com bubble. Most of these speculative names are not in major indices. Those that are often have such small weights that even a 50–90% drawdown in their share prices would barely move the S&P 500 or the Nasdaq. Traders and speculators in those names will feel real pain if and when the air comes out. What it won’t do is suddenly erase Microsoft’s $400 billion cloud backlog, or Amazon’s $200 billion AWS backlog, or Alphabet’s $155 billion cloud backlog. It won’t retroactively make Meta’s $60+ billion AI ad run rate a mirage. And it won’t stop these companies from continuing to build out infrastructure as long as they see durable demand and attractive returns on capital. The Bottom Line If your entire bear case rests on “this looks like 1999 again,” you may be anchored to the wrong historical analogy. Dot-com was a story of fragile, cash-burning business models financed by an endless chain of new investors. Today’s AI cycle is a story of cash-rich incumbents reinvesting enormous free cash flow into infrastructure to meet committed demand. None of that means AI is risk-free as Capex cycles have the ability to overshoot while competition intensifies. Multiples can compress even if earnings grow. There will be individual winners and losers, and owning the wrong high-flyer at the wrong time will still hurt. If you’re waiting for a 2000-style wipeout in mega-cap tech as your entry point, you might be waiting for a crash that belongs to a different era. Today’s market is one where the world’s most profitable businesses are trading at reasonable forward multiples while funding one of the largest infrastructure build-outs in history from their own cash flow. That’s not the dot-com bubble 2.0. and to put it nicely, the bears are wrong trying to correlate the AI revolution to the dot-com bubble. @amitisinvesting @KrisPatel99 @RealMattMoney @Futurenvesting @FunOfInvesting @dhurstell @_financeken @DanIvesprivate @fundstrat @Kross_Roads @StockMarketNerd @sam_badawi

QXO: Deconstructing The Playbook To Consolidate Building Products Distribution Market $QXO ⁦@SeekingAlpha⁩ seekingalpha.com/article/483…