<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Neocloud on Korea Invest Insights</title><link>https://koreainvestinsights.com/tags/neocloud/</link><description>Recent content in Neocloud on Korea Invest Insights</description><generator>Hugo -- gohugo.io</generator><language>en</language><copyright>koreainvestinsights.com · @korea_invest_insights</copyright><lastBuildDate>Mon, 02 Nov 2026 00:03:29 +0900</lastBuildDate><atom:link href="https://koreainvestinsights.com/tags/neocloud/feed.xml" rel="self" type="application/rss+xml"/><item><title>Provisional Verdict on AI Demand: Supply Caps 2027, the Bill Lands in 2028, and the Market Sold the Wrong Year</title><link>https://koreainvestinsights.com/post/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/</link><pubDate>Sun, 01 Nov 2026 23:47:00 +0900</pubDate><guid>https://koreainvestinsights.com/post/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/</guid><description>&lt;img src="https://koreainvestinsights.com/" alt="Featured image of post Provisional Verdict on AI Demand: Supply Caps 2027, the Bill Lands in 2028, and the Market Sold the Wrong Year" /&gt;&lt;p&gt;The market&amp;rsquo;s story is short: &amp;ldquo;AI demand is rolling over. That is why SK hynix is down 51% from its high.&amp;rdquo;&lt;/p&gt;
&lt;p&gt;We take the other side. Over the past two weeks this series asked the same question from several angles: &lt;a class="link" href="https://koreainvestinsights.com/post/dram-shortage-part-inventory-buyers-stop-when-cpus-arrive-2026-10-30/" &gt;part of the DRAM shortage is buyer inventory&lt;/a&gt;, &lt;a class="link" href="https://koreainvestinsights.com/post/ai-bullwhip-grid-cloud-memory-2026-10-31/" &gt;the AI bullwhip map&lt;/a&gt;, &lt;a class="link" href="https://koreainvestinsights.com/post/ai-roi-beyond-token-price-2026-10-31/" &gt;AI ROI clears the bar through volume and other ledgers&lt;/a&gt;, &lt;a class="link" href="https://koreainvestinsights.com/post/ai-hardware-selloff-supplier-profit-shelf-life-2026-11-01/" &gt;the market sold the shelf life of supplier profits&lt;/a&gt;, and &lt;a class="link" href="https://koreainvestinsights.com/post/samsung-per-3x-prices-60-percent-2028-profit-drop-2026-11-01/" &gt;what Samsung&amp;rsquo;s 3x multiple prices into 2028&lt;/a&gt;. Today we tie those pieces into one call. It is provisional, and the conditions that would flip it are listed with dates.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Our view: In 2027, AI compute demand exceeds supply, and it stays that way all year. Supply, not demand, sets the 2027 token count. So &amp;ldquo;demand slowdown&amp;rdquo; is not the 2027 risk. The risk sits where the money flows. Token prices fall 45-55% a year while compute rents hold. That squeeze lands on labs that rent their compute, and the real variable is whether it travels through finance into 2028 orders. The bullwhip is real but small. It sits in buyers&amp;rsquo; commodity DRAM buffers, and its bill comes in 2028. Because the supply chain is behaving conservatively, a 2028 commodity DRAM correction is coming, but it will be shallower than what the market has priced. The market sold 2028 deep and has not bought any of 2027&amp;rsquo;s earnings.&lt;/strong&gt;&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;We put 80% odds on paying demand exceeding capacity for all of 2027. Flat growth, announced list-price hikes and sold-out capacity are the shape of a ceiling.&lt;/li&gt;
&lt;li&gt;The odds that 2027 memory profits break through the demand path are 10% or less. The odds that the finance path cuts 2027 shipments are 15% or less. The squeeze lands on 2028 orders.&lt;/li&gt;
&lt;li&gt;Non-repeating buyer-buffer demand is 8-13% of 2027 shipments. Moved into 2028 demand, it is about 10%, one scenario notch.&lt;/li&gt;
&lt;li&gt;We put 55% odds on a 2028-29 commodity DRAM correction. Because the big four memory makers add only 15-17% more wafers in 2028, we expect it to be shallow.&lt;/li&gt;
&lt;/ul&gt;
&lt;p&gt;The reference date is November 1, 2026. All probabilities are our subjective estimates. We move them only in 5-point steps, and only on a primary source or a threshold break.&lt;/p&gt;
&lt;h2 id="the-numbers-draw-a-ceiling-not-fatigue"&gt;The numbers draw a ceiling, not fatigue
&lt;/h2&gt;&lt;p&gt;In a market with tiring demand, quarterly growth keeps slowing and prices and bookings fall with it. Every current number points the other way.&lt;/p&gt;
&lt;p&gt;&lt;img alt="A ceiling, not fatigue. Google first-party API tokens per minute grew 60%, 37% and 40% QoQ in Q1, Q2 and Q3 2026, flat near +40%. Combined quarterly revenue of two frontier labs was $10.4bn, $18.2bn and $26.4bn, up 45% QoQ despite falling prices. Per Epoch, fixed-price demand grows about 10x a year against 3.4x for inference supply, a gap of about 3x. Google announced API list-price hikes, North American data center vacancy is 1.4%, H100 one-year contract prices are up 38% YoY, and most of Amazon’s 2027 capacity is already booked." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/ceiling.en.svg"&gt;
&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Indicator&lt;/th&gt;
 &lt;th&gt;Value&lt;/th&gt;
 &lt;th&gt;How to read it&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Google first-party API tokens per minute, QoQ&lt;/td&gt;
 &lt;td&gt;Q1 2026 +60%, Q2 +37%, Q3 +40%. Two quarters annualized: 3.7-4.0x&lt;/td&gt;
 &lt;td&gt;It stepped down once and then flattened. It is not bending every quarter&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Google API list prices&lt;/td&gt;
 &lt;td&gt;Hike announced (effective Dec 1 and Jan 1, 2027)&lt;/td&gt;
 &lt;td&gt;Demand fatigue would push prices down. They are going up&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Two frontier labs, combined quarterly revenue&lt;/td&gt;
 &lt;td&gt;Q1 2026 $10.4bn, Q2 $18.2bn, Q3 $26.4bn&lt;/td&gt;
 &lt;td&gt;+45% QoQ despite falling unit prices&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Amazon Bedrock tokens (YoY), Microsoft Foundry&lt;/td&gt;
 &lt;td&gt;17x, 5x&lt;/td&gt;
 &lt;td&gt;Wherever capacity frees up, it fills&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;North American DC vacancy, H100 one-year contract price&lt;/td&gt;
 &lt;td&gt;1.4%, +38% YoY&lt;/td&gt;
 &lt;td&gt;The scarce thing is energized GW&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Amazon 2027 capacity&lt;/td&gt;
 &lt;td&gt;Mostly booked; a sizable share of 2028 too&lt;/td&gt;
 &lt;td&gt;2027 is already sold&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Epoch: fixed-price demand vs inference supply&lt;/td&gt;
 &lt;td&gt;About 10x vs 3.4x a year&lt;/td&gt;
 &lt;td&gt;A ~3x gap, toward shortage&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;Google&amp;rsquo;s token growth is flat near +40%. Over the same stretch Google announced list-price hikes and Amazon closed out 2027. That combination is the shape of a supply curve, not a demand curve. The Q3 token figure is in Alphabet&amp;rsquo;s earnings release (&lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/1652044/000165204426000080/googexhibit991q32026.htm" target="_blank" rel="noopener"
 &gt;Alphabet 8-K, Q3 2026&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;Anthropic&amp;rsquo;s second quarter showed the same thing. When capacity freed up, pent-up demand appeared at once. We used Anthropic&amp;rsquo;s Claude for research and editing, so we carry Anthropic&amp;rsquo;s numbers as reported and do not judge their impact.&lt;/p&gt;
&lt;h2 id="solve-2027-token-volume-with-the-supply-formula"&gt;Solve 2027 token volume with the supply formula
&lt;/h2&gt;&lt;p&gt;When demand is pinned against a ceiling, demand surveys cannot tell you 2027 volume. The supply formula can. Multiply supply GW growth of 40-50% a year by Nvidia Rubin&amp;rsquo;s realized efficiency.&lt;/p&gt;
&lt;p&gt;If Rubin delivers 5-7x Blackwell on real workloads, hardware alone gives 2.5-3.1x the 2026 level. Add 1.3x from software and the ceiling is 3.3-4.0x. Our previous base range of 2.5-3.0x sat at the bottom of that ceiling. We move the base to 2.8-3.3x. That range is GW +45% times 1.9-2.3x tokens per GW. Because volume is supply-bound, GW demand is also near the same +45%.&lt;/p&gt;
&lt;p&gt;&lt;img alt="The supply formula for 2027 token volume. Hardware alone: 2.5-3.1x. Ceiling with 1.3x software: 3.3-4.0x. Old base: 2.5-3.0x. New base: 2.8-3.3x. Scenario weights move from bear 30 / base 50 / bull 20 to bear 20 / base 55 / bull 25. This is a sensitivity example that excludes power, HBM, CPU and substrate bottlenecks." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/supply-math.en.svg"&gt;
&lt;/p&gt;
&lt;p&gt;Demand does not decide anything above 3.5x. Rubin does. That is why the bull case struggles to go above 20-25% no matter how strong demand is. A result below 2.5x has two routes. One is demand breaking. The other is smaller throughput from a Rubin delay, weak realized efficiency, or power and substrate bottlenecks. In that second case unit prices and rents hold, which is the opposite signal from weak demand. There is no evidence that demand has broken, so we cut the bear case from 30% to 20%. We track the supply route separately through measured Rubin efficiency and power.&lt;/p&gt;
&lt;p&gt;One condition withdraws this call. If Alphabet&amp;rsquo;s Q4 API tokens per minute grow less than 15% QoQ, the problem is demand, not price, and that number alone is enough. If token volume in corporate spend samples like Ramp shrinks, we review; three months of no growth plus another drop in the top 1% sends us back to the previous base. The first check is the Ramp data on November 13. The decisive check is Alphabet&amp;rsquo;s results in early February 2027.&lt;/p&gt;
&lt;h2 id="the-risk-lives-where-the-money-flows"&gt;The risk lives where the money flows
&lt;/h2&gt;&lt;p&gt;Demand can grow without money growing at the same rate. We call this the token-compute scissors. Over the same four weeks, the blended token price in corporate card data fell 26.8% and Nvidia B200 rental rates rose 10.6%.&lt;/p&gt;
&lt;p&gt;&lt;img alt="The token-compute scissors. Over the same four weeks, B200 rental rose 10.6% and the Ramp blended token price fell 26.8%. The space between the lines is where lab margins get pinched. Annualized, token prices fall 45-55% while compute rents hold." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/scissors.en.svg"&gt;
&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Variable&lt;/th&gt;
 &lt;th&gt;Value&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Ramp blended token price, 4 weeks&lt;/td&gt;
 &lt;td&gt;−26.8%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;B200 rental, same period&lt;/td&gt;
 &lt;td&gt;+10.6%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;2027 lab revenue (4Q27/4Q26, previous base of 2.5-3.0x volume)&lt;/td&gt;
 &lt;td&gt;+40-80%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;2027 GW demand (previous base)&lt;/td&gt;
 &lt;td&gt;+25-90%, center +53%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Lab revenue at the new 2.8-3.3x base and unit prices −40-45%&lt;/td&gt;
 &lt;td&gt;+54-98%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;GW demand at 2.8x Rubin efficiency and 4.2x volume&lt;/td&gt;
 &lt;td&gt;+50%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Lab revenue per GW under the same conditions&lt;/td&gt;
 &lt;td&gt;+12-40%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;Compute scarcity protects rental rates. It does not protect token prices, because labs chase volume with cheap tiers. So the squeeze hits the margins of labs that rent compute before it hits hyperscalers that own it.&lt;/p&gt;
&lt;p&gt;If lab funding stalls, the backlog that leans on labs wobbles. Oracle&amp;rsquo;s prepayments, the OpenAI share of Microsoft&amp;rsquo;s and Oracle&amp;rsquo;s backlog, and the neoclouds sit there. Neocloud and developer problems reach hyperscaler balance sheets through guarantees: Alphabet&amp;rsquo;s $48 billion data center payment guarantees and Meta&amp;rsquo;s $41.5 billion residual value guarantees.&lt;/p&gt;
&lt;p&gt;&lt;img alt="Where the risk lives. Token prices falling 45-55% a year squeeze the margins of labs that rent compute. That pressure runs into Oracle prepayments, the OpenAI share of Microsoft and Oracle backlog, and neoclouds, and then into hyperscaler balance sheets through Alphabet’s $48bn data center guarantees and Meta’s $41.5bn residual value guarantees. Odds: neocloud partial funding block 35%, hyperscaler funding block 15%. 2027 shipments are locked by LTAs, prepayments and the power cap, so the odds the finance path cuts them are under 15%; the squeeze lands on 2028 orders. +100bp is 0.28% of hyperscaler operating cash flow; 2027 funding need is about $200bn." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/money-path.en.svg"&gt;
&lt;/p&gt;
&lt;p&gt;We put 15% odds on hyperscaler funding stalling and 35% on neoclouds being partly blocked. This path can arrive inside 2027. LTAs carry price bands, volume-cut and undelivered-volume termination clauses, and SK hynix&amp;rsquo;s contracts cover more than 50% of 2027-28 capacity, not all of it. Still, hyperscalers are the ones buying 2027 shipments, and prepayments and the power cap tie up volume. So this path mainly cuts 2028 orders rather than 2027 shipments. How firmly volume is locked follows the counterparty&amp;rsquo;s cash ranking: hyperscalers first, then labs, then neoclouds.&lt;/p&gt;
&lt;h3 id="more-tokens-do-not-mean-proportionally-more-memory-bits"&gt;More tokens do not mean proportionally more memory bits
&lt;/h3&gt;&lt;p&gt;Here is the paradox. Even if tokens grow faster than expected, GW and memory bits do not grow at the same rate. At 4.2x token volume, GW demand is +50%, the same as the center of the base case. What rises is lab revenue per GW.&lt;/p&gt;
&lt;p&gt;&lt;img alt="Token upside is not memory-bit upside. 4.2x token volume equals 1.5x GW demand times 2.8x Rubin efficiency. If tokens beat, lab revenue per GW rises 12-40%, lab margins recover and finance-path risk shrinks. If tokens miss, 2027 GW stays put because LTAs, prepayments and the power cap bind it, so 2027 memory profit is insensitive to the token count." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/token-gw.en.svg"&gt;
&lt;/p&gt;
&lt;p&gt;So token upside does not arrive as memory-bit upside. It arrives as recovering lab margins, and it shrinks the finance-path risk. In reverse, token downside cannot cut 2027 GW by much either, because most 2027 volume is tied up by long-term agreements (LTAs), prepayments and the power cap. 2027 memory profit is insensitive to the token count. That is the precise meaning of &amp;ldquo;2027 is locked.&amp;rdquo;&lt;/p&gt;
&lt;p&gt;So here is where we stand. The odds that 2027 memory profits break through the demand path are 10% or less. The odds that the finance path cuts 2027 shipments are 15% or less. In 2027, the squeeze will show up in lab margins and in Oracle and neocloud spreads, not in memory earnings.&lt;/p&gt;
&lt;p&gt;Rates are second-order. A 100bp rise costs hyperscalers 0.28% of operating cash flow. The issue is funding volume, not the rate level. The 2027 funding need is about $200 billion. As our &lt;a class="link" href="https://koreainvestinsights.com/post/ackman-ai-rates-fed-mistake-marginal-dollar-is-debt-2026-11-01/" &gt;same-day Ackman post&lt;/a&gt; showed, Big Tech&amp;rsquo;s combined Q3 free cash flow is already negative. The marginal AI dollar is debt, and the variable is whether that debt gets funded on time.&lt;/p&gt;
&lt;h2 id="the-bullwhip-is-small-and-it-sits-in-buyers-commodity-dram"&gt;The bullwhip is small, and it sits in buyers&amp;rsquo; commodity DRAM
&lt;/h2&gt;&lt;p&gt;The bullwhip has a different character at each stage. Paper orders are largest in grid connection requests and Chinese GPU orders. But they build no inventory, and forecasts already filter them out (&lt;a class="link" href="https://koreainvestinsights.com/post/ai-bullwhip-grid-cloud-memory-2026-10-31/" &gt;AI bullwhip map&lt;/a&gt;). Committed forward demand is largest in multi-year lab-to-cloud commitments (RPO). Its downside path is renegotiation, not cancellation.&lt;/p&gt;
&lt;p&gt;The bullwhip that builds inventory is in memory, specifically buyers&amp;rsquo; buffers of commodity server and PC DRAM. The cause is the server CPU shortage. Without CPUs, buyers cannot build servers, yet they bought DRAM first (&lt;a class="link" href="https://koreainvestinsights.com/post/dram-shortage-part-inventory-buyers-stop-when-cpus-arrive-2026-10-30/" &gt;part of the DRAM shortage is inventory&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;&lt;img alt="The bullwhip differs by stage. Paper orders (grid requests, China GPUs) are the most inflated but build no inventory and are already filtered, so the bill is near zero. Committed forward demand (lab-to-cloud RPO) is largest in dollars, its downside path is renegotiation, and its bill is 2028 orders. The inventory-building whip (commodity server and PC DRAM buyers) is a 3-5 week buffer; buyers stop when CPUs and substrates land, costing about 10% of 2028 demand and about 10 points of price. 8-13% of 2027 shipments do not repeat. The first to take an HBM-side whip is Nvidia’s $279bn of supply commitments, 96% in FY27-29." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/bullwhip.en.svg"&gt;
&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Item&lt;/th&gt;
 &lt;th&gt;Value&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Buyer DRAM buffer, excess over normal&lt;/td&gt;
 &lt;td&gt;1-5 weeks estimated (cumulative shipments minus installs 2-9 weeks, center 5; end-2025 opening stock unconfirmed)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Non-repeating buffer demand in 2027 shipments&lt;/td&gt;
 &lt;td&gt;About 8-13%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Share of 2028 demand taken by a 5-week excess buffer&lt;/td&gt;
 &lt;td&gt;About 10%. One scenario notch, about 10 points of price&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Threshold for a deep correction&lt;/td&gt;
 &lt;td&gt;Excess buffer of 8-12 weeks at end-2027, 2028 orders down 15-23%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Trigger for unwinding&lt;/td&gt;
 &lt;td&gt;2H27 server CPU and substrate supply additions&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;First to take an HBM-side whip&lt;/td&gt;
 &lt;td&gt;Nvidia supply commitments of $279bn, 96% in FY27-29 (&lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm" target="_blank" rel="noopener"
 &gt;NVIDIA 10-Q&lt;/a&gt;, Note 10)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;It is small, and we know where it sits. The 2027 profit figure does not change. The bill comes in 2028, and its size is about the difference between a balanced 2028 and a mildly oversupplied one.&lt;/p&gt;
&lt;h2 id="a-conservative-supply-chain-keeps-the-2028-correction-shallow"&gt;A conservative supply chain keeps the 2028 correction shallow
&lt;/h2&gt;&lt;p&gt;The big four memory makers grow 2028 wafers by 15-17%. Samsung runs five-year rolling LTAs, SK hynix typically five years, and the three makers&amp;rsquo; $38 billion of prepayments stretch to 2030 (&lt;a class="link" href="https://koreainvestinsights.com/post/samsung-sk-hynix-lta-volume-locked-terms-flexible-2026-10-29/" &gt;LTAs lock in volume, not profit&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;The 38% drop in the equipment basket discounts lower equipment orders in 2027-28. If that is the main cause, 2028 supply is more disciplined than the market fears. HBM is only about 10% of DRAM revenue at all three makers (estimates: SK hynix 13-14%, Samsung 11-12%, Micron 7-10%), so they ration capacity like commodity-cycle companies.&lt;/p&gt;
&lt;p&gt;Power points the same way. The US can build at most 30-40GW of data centers a year. Transformers are about 15% short. Gas turbines ordered today deliver in 2029-31. So 2027 supply cannot grow beyond the build-out already set (GW +40-50%), and extra supply arrives only modestly in 2028.&lt;/p&gt;
&lt;p&gt;&lt;img alt="The supply chain is conservative. Memory LTAs roll for five years and $38bn of prepayments run to 2030. Big-four wafers grow only 15-17% in 2028. On power, the US build cap of 30-40GW a year and a 15% transformer shortfall run through 2028, and gas turbines deliver in 2029-31. Buyer buffers build through 1H27 and are drawn down after CPUs arrive in 2H27. Amazon has booked most of 2027 and a sizable share of 2028. Our call: 55% odds of a 2028-29 commodity DRAM correction, shallow, and 20% odds of an HBM glut." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/lock-timeline.en.svg"&gt;
&lt;/p&gt;
&lt;p&gt;Here is how we see 2028. A commodity DRAM correction is coming; we put the cumulative 2028-29 odds at 55%. If CPUs and substrates arrive in 2H27, buffer building stops and installs absorb the stock, it will not be a deep 2021-22-style correction, because the supply increment is small and volume is tied up by prepayments and LTAs.&lt;/p&gt;
&lt;p&gt;The shallow call is conditional. In 2021 the excess buffer was 3-7 weeks, which overlaps today&amp;rsquo;s 1-5 week estimate. If the excess buffer reaches 8-12 weeks by end-2027, 2028 orders fall 15-23%, a collapse-level cut. The threshold is buyer inventory days rising all through 2027 and not turning in 2H27.&lt;/p&gt;
&lt;p&gt;We put the odds of an HBM glut at 20%. Samsung, with its larger commodity share, is more exposed. The first signal is mobile DRAM contract prices.&lt;/p&gt;
&lt;h2 id="the-market-sold-2028-deep-and-bought-none-of-2027"&gt;The market sold 2028 deep and bought none of 2027
&lt;/h2&gt;&lt;p&gt;This is where the market is wrong. From September 21 to October 30, about 83% of foreign net selling in Taiwan was AI hardware. Financials, telecoms and steel were not sold. What got sold was the shelf life of supplier profits, not AI demand (&lt;a class="link" href="https://koreainvestinsights.com/post/ai-hardware-selloff-supplier-profit-shelf-life-2026-11-01/" &gt;our November 1 post&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;Prices make it sharper. Samsung trades at 3.1x 2027 expected earnings. Plug in a normal 8x multiple, and the price assumes 2028 earnings about 60% below 2027 (&lt;a class="link" href="https://koreainvestinsights.com/post/samsung-per-3x-prices-60-percent-2028-profit-drop-2026-11-01/" &gt;the Samsung 3x post&lt;/a&gt;). Almost none of 2026-27&amp;rsquo;s earnings are in the price; 2028 is priced deep. The 2028 we see is shallower.&lt;/p&gt;
&lt;p&gt;Reading the drops in Intel and Chinese AI stocks as a demand slowdown today is not supported by current demand data. The gate to the demand-slowdown scenario (S3) is a single one: hyperscalers slowing capex. On their Q3 calls, all four raised 2027 capex. Amazon called 2027 a &amp;ldquo;meaningful increase,&amp;rdquo; and Meta even turned to equity financing. Prices alone cannot rule out that the market is pricing a future demand slowdown starting with suppliers. The tell is the order in which equipment and memory relative strength turn.&lt;/p&gt;
&lt;p&gt;Our call rests on 2027 earnings that are not in the price. The fastest way that gap closes is through flows, not fundamentals. Deleveraging is at the start of stage 4 of 6, and forced selling by leveraged retail investors is still ahead. That means one more shakeout before the bottom is possible.&lt;/p&gt;
&lt;h2 id="the-stance-in-numbers-less-demand-risk-more-credit-risk"&gt;The stance in numbers: less demand risk, more credit risk
&lt;/h2&gt;&lt;p&gt;&lt;img alt="The stance in numbers. 2027 token-volume weights move from bear 30 / base 50 / bull 20 to bear 20 / base 55 / bull 25, and the base range rises from 2.5-3.0x to 2.8-3.3x. Memory-cycle scenarios move from S1 rebound 55 / S2 credit squeeze 15 / S3 demand slowdown 15 / S4 macro shock 15 to S1 55 / S2 20 / S3 10 / S4 15, shifting 5 points from S3 to S2. New calls: 80% that paying demand exceeds capacity all of 2027, 10% or less that demand breaks 2027 memory profits, 55% for a shallow 2028-29 commodity DRAM correction." loading="lazy" sizes="(max-width: 767px) calc(100vw - 30px), (max-width: 1023px) 700px, (max-width: 1279px) 950px, 1232px" src="https://koreainvestinsights.com/post-assets/ai-demand-provisional-verdict-2027-supply-capped-2028-bill-2026-11-01/weights.en.svg"&gt;
&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Call&lt;/th&gt;
 &lt;th&gt;Previous&lt;/th&gt;
 &lt;th&gt;This stance&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;2027 token volume&lt;/td&gt;
 &lt;td&gt;Bull 20 / base 50 / bear 30, base 2.5-3.0x&lt;/td&gt;
 &lt;td&gt;Bull 25 / base 55 / bear 20, base 2.8-3.3x&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Memory-cycle scenarios&lt;/td&gt;
 &lt;td&gt;S1 55 / S2 15 / S3 15 / S4 15&lt;/td&gt;
 &lt;td&gt;S1 55 / S2 20 / S3 10 / S4 15&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Paying demand exceeds capacity all of 2027&lt;/td&gt;
 &lt;td&gt;New&lt;/td&gt;
 &lt;td&gt;80%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;2027 memory profits break via demand&lt;/td&gt;
 &lt;td&gt;New&lt;/td&gt;
 &lt;td&gt;10% or less&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;2028-29 commodity DRAM correction&lt;/td&gt;
 &lt;td&gt;2028 45%, cumulative 55%&lt;/td&gt;
 &lt;td&gt;Cumulative 55%, shallow&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Character of the bullwhip&lt;/td&gt;
 &lt;td&gt;Temporary 35 / structural 50 / negligible 15&lt;/td&gt;
 &lt;td&gt;Unchanged, leaning structural&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;The scenario names follow our &lt;a class="link" href="https://koreainvestinsights.com/post/chip-selloff-four-scenarios-week-ahead-2026-10-24/" &gt;October 24 post&lt;/a&gt;: S1 rebound, S2 credit squeeze, S3 demand slowdown, S4 macro shock.&lt;/p&gt;
&lt;p&gt;Moving 5 points from S3 to S2 has one meaning: we move the risk&amp;rsquo;s address from demand to credit. After Amazon&amp;rsquo;s results we also weighed raising S2 to 35%, but we kept the move smaller. The company borrowing cash fastest is also the one with the longest booked backlog, and we give those two facts equal weight.&lt;/p&gt;
&lt;h2 id="answering-the-obvious-objections-first"&gt;Answering the obvious objections first
&lt;/h2&gt;&lt;p&gt;&lt;strong&gt;&amp;ldquo;People use tokens because they got cheap. That is not real demand.&amp;quot;&lt;/strong&gt; Cheap tokens did pull in demand. That is exactly why we put the risk in the money flow. But demand that rests only on price would give Google no reason to announce list-price hikes. In quarters when unit prices fell, volume still grew 40% and lab revenue grew 45%.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;&amp;ldquo;Intel and Chinese AI stocks fell. Demand is cooling.&amp;quot;&lt;/strong&gt; The gate that decides a demand slowdown is hyperscalers&amp;rsquo; 2027 capex. All four raised it in Q3. While that gate is shut, drops in peripheral stocks are explained by flows and discount rates.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;&amp;ldquo;If Rubin is 7x more efficient, there will be spare GW.&amp;quot;&lt;/strong&gt; Higher efficiency raises the ceiling on 2027 token volume. It does not leave GW idle. 2027 GW is already tied up by LTAs and the power cap, and paying demand fills it. If Rubin stays at 2-4x, volume comes in smaller and unit prices hold. Either way, 2027 memory profits barely move.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;&amp;ldquo;When the bullwhip snaps, it will be as deep as 2021-22.&amp;quot;&lt;/strong&gt; What differs is supply&amp;rsquo;s hands. 2028 wafer growth is 15-17%, and volume is tied up in prepayments and LTAs. A 5-week excess buffer takes about 10% of 2028 demand. That is one scenario notch, not a whole cycle. We also wrote down when this answer fails: an excess buffer of 8-12 weeks by end-2027 means a deep correction.&lt;/p&gt;
&lt;h2 id="what-would-flip-this-call"&gt;What would flip this call
&lt;/h2&gt;&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Area&lt;/th&gt;
 &lt;th&gt;Signal&lt;/th&gt;
 &lt;th&gt;Threshold&lt;/th&gt;
 &lt;th&gt;When&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Demand&lt;/td&gt;
 &lt;td&gt;Alphabet API tokens per minute, QoQ&lt;/td&gt;
 &lt;td&gt;Under +15% means a demand problem. Under +30% means review a lower base range. +40% or more for two quarters confirms a higher range&lt;/td&gt;
 &lt;td&gt;Early Feb 2027&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Demand&lt;/td&gt;
 &lt;td&gt;Token volume and spend in samples like Ramp&lt;/td&gt;
 &lt;td&gt;Volume +20% a month plus a spend rebound locks in the 25% bull case. Three months of no growth plus another drop in the top 1% restores the bear case&lt;/td&gt;
 &lt;td&gt;Nov 13&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Finance&lt;/td&gt;
 &lt;td&gt;High-yield spread, Oracle CDS, CoreWeave bonds&lt;/td&gt;
 &lt;td&gt;350bp+, 300bp+ and below 75 cents together mean a shift to S2&lt;/td&gt;
 &lt;td&gt;Early-November issuance window; CoreWeave Nov 11&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Finance&lt;/td&gt;
 &lt;td&gt;Two labs&amp;rsquo; combined quarterly revenue, QoQ&lt;/td&gt;
 &lt;td&gt;Below +9% for two straight quarters means the 2030 payback path breaks&lt;/td&gt;
 &lt;td&gt;Every quarter in 2027&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Supply&lt;/td&gt;
 &lt;td&gt;Rubin customer-measured tokens per MW&lt;/td&gt;
 &lt;td&gt;5x or more raises the volume ceiling. Stuck at 2-4x means smaller 2027 volume and firm unit prices&lt;/td&gt;
 &lt;td&gt;Nov 19 earnings is the starting point; independent checks in 1H27&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Bullwhip&lt;/td&gt;
 &lt;td&gt;Rising buyer DRAM inventory days, falling same-generation rents, slowing follow-on orders&lt;/td&gt;
 &lt;td&gt;All three together move 2028 correction depth from &amp;ldquo;shallow&amp;rdquo; to &amp;ldquo;moderate&amp;rdquo;&lt;/td&gt;
 &lt;td&gt;Quarterly&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;2028&lt;/td&gt;
 &lt;td&gt;Mobile DRAM contract prices, 2028 HBM specs&lt;/td&gt;
 &lt;td&gt;A turn to falling contract prices pulls the commodity correction forward&lt;/td&gt;
 &lt;td&gt;Late-October contract prices onward; Nov 19 specs&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;h2 id="limits-and-interests"&gt;Limits and interests
&lt;/h2&gt;&lt;p&gt;Google&amp;rsquo;s mid-2025 monthly tokens, Microsoft Foundry, OpenRouter charts and Anthropic&amp;rsquo;s listing-document figures are secondary reports. Token definitions differ by metric. The capacity formula is a sensitivity example that leaves out power, HBM, CPU and substrate bottlenecks. The equipment drawdown mixes China exposure and position liquidation, so we do not read all of it as supply discipline. Without the end-2025 opening inventory, we cannot yet tell whether the buyer buffer is one week or five.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Do you think demand or supply sets the 2027 ceiling? Which number will shake this call first, the November 13 Ramp data or Alphabet&amp;rsquo;s token growth in February 2027? Tell us why.&lt;/strong&gt;&lt;/p&gt;
&lt;p&gt;For research and information only. Not investment advice. Probabilities and scenarios are analytical assumptions, not trading instructions or individual portfolio decisions. Figures carry reference dates, and some come from public reports and secondary tallies. We used Anthropic&amp;rsquo;s Claude for research and editing. Anthropic appears in the AI demand chain discussed here as a customer and a compute buyer.&lt;/p&gt;</description></item><item><title>The Marginal AI Dollar Is Now Debt: Ackman's 'Near-Infinite ROI' Math Is Right, His Fed-Mistake Verdict Is Not</title><link>https://koreainvestinsights.com/post/ackman-ai-rates-fed-mistake-marginal-dollar-is-debt-2026-11-01/</link><pubDate>Sun, 01 Nov 2026 23:40:00 +0900</pubDate><guid>https://koreainvestinsights.com/post/ackman-ai-rates-fed-mistake-marginal-dollar-is-debt-2026-11-01/</guid><description>&lt;img src="https://koreainvestinsights.com/" alt="Featured image of post The Marginal AI Dollar Is Now Debt: Ackman's 'Near-Infinite ROI' Math Is Right, His Fed-Mistake Verdict Is Not" /&gt;&lt;p&gt;On September 24 (US time), Pershing Square CEO Bill Ackman wrote on X that &amp;ldquo;winning the race for super intelligence has a near infinite ROI.&amp;rdquo; Higher rates, he argued, would not reduce AI investment or the demand for compute. They would only push inflation higher, because interest costs are embedded in everything. His conclusion: &amp;ldquo;I think the Fed might have just made a mistake&amp;rdquo; (&lt;a class="link" href="https://finance.yahoo.com/economy/policy/articles/billionaire-investor-bill-ackman-says-160620045.html" target="_blank" rel="noopener"
 &gt;Business Insider via Yahoo, Sep 26&lt;/a&gt;). On September 16 the Fed had raised its policy rate by 25bp to 3.75-4.00%. It was the first hike since July 2023, and the vote was 12-0 (&lt;a class="link" href="https://www.federalreserve.gov/newsevents/pressreleases/monetary20260916a.htm" target="_blank" rel="noopener"
 &gt;Fed statement, Sep 16&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;In Korea the argument spread as a simpler line: if the expected return is 100%, what difference does a 5% versus 5.5% rate make? The original post contains neither 100% nor 5.5%. The logic is the same, though. When expected returns far exceed the cost of capital, a small change in that cost does not change the investment decision.&lt;/p&gt;
&lt;p&gt;A month later, Fed Chair Kevin Warsh rejected the view at his press conference: &amp;ldquo;Interest rates get in the cracks.&amp;rdquo; Who is right matters for AI capex, for memory chips, and for the stocks of companies that borrow to build data centers.&lt;/p&gt;
&lt;p&gt;&lt;strong&gt;Our view: Ackman&amp;rsquo;s arithmetic is right. In our hypothetical examples, a 50bp rise cuts the NPV of short-lived, high-return investments such as GPUs by only 1-4%. But the conclusion that the Fed made a mistake does not follow from that arithmetic. In Q3 2026 the combined free cash flow (FCF) of the five hyperscalers was -$31.8 billion. The marginal AI dollar now comes from outside operating cash, and the largest identified source is debt. Borrowers are sensitive to rates. Ackman himself said so about neoclouds in October. And if AI investment truly ignores rates while inflation pressure persists, the Fed has to press harder on other sectors, such as housing, autos and small business. Insensitivity alone is not a reason to stop hiking.&lt;/strong&gt;&lt;/p&gt;
&lt;p&gt;Data are as of November 1, 2026. The sources are the September 24 X post (as quoted in press reports), CNBC&amp;rsquo;s unofficial transcript of an October 22 interview, and coverage of Warsh&amp;rsquo;s October 28 press conference. Company figures are from each firm&amp;rsquo;s latest quarterly results (Microsoft&amp;rsquo;s September quarter, Oracle&amp;rsquo;s August quarter). Rates and credit spreads are FRED values for October 29.&lt;/p&gt;
&lt;h2 id="the-original-post-contains-neither-100-nor-55"&gt;The original post contains neither &amp;ldquo;100%&amp;rdquo; nor &amp;ldquo;5.5%&amp;rdquo;
&lt;/h2&gt;&lt;p&gt;Put the popular summary next to the original and the gap is clear.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Item&lt;/th&gt;
 &lt;th&gt;Popular summary&lt;/th&gt;
 &lt;th&gt;Original and facts&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Expected return&lt;/td&gt;
 &lt;td&gt;100%&lt;/td&gt;
 &lt;td&gt;&amp;ldquo;winning the race for super intelligence has a near infinite ROI&amp;rdquo; (X, Sep 24)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Rates&lt;/td&gt;
 &lt;td&gt;5% vs 5.5%&lt;/td&gt;
 &lt;td&gt;The Fed raised the short rate by 25bp to 3.75-4.00%. On CNBC on Oct 22, Ackman said a 5% 10-year is still a relatively low cost of capital&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Conclusion&lt;/td&gt;
 &lt;td&gt;The rate gap is meaningless&lt;/td&gt;
 &lt;td&gt;&amp;ldquo;I think the Fed might have just made a mistake.&amp;rdquo; Hikes fail to cut investment and raise inflation through pass-through of interest costs&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;The &amp;ldquo;5%&amp;rdquo; probably comes from the October CNBC interview. Ackman said that since intelligence is so valuable, &amp;ldquo;you can earn an attractive return even with a risk-free rate of 5 percent&amp;rdquo; (&lt;a class="link" href="https://www.cnbc.com/2026/10/22/cnbc-exclusive-transcript-pershing-square-founder-ceo-bill-ackman-speaks-with-cnbcs-sara-eisen-on-squawk-on-the-street-today.html" target="_blank" rel="noopener"
 &gt;CNBC transcript, Oct 22&lt;/a&gt;). The &amp;ldquo;100%&amp;rdquo; turns &amp;ldquo;near infinite&amp;rdquo; into an example number. The numbers differ, but the logic is the same, so we test the logic of the original.&lt;/p&gt;
&lt;p&gt;The original runs like this. If the ROI of the super-intelligence race is near infinite, higher rates will not reduce demand or investment. Because interest costs are embedded in everything, a hike then raises inflation. The more the Fed hikes, the more inflation rises and the more the Fed must hike. The old models may not apply to the current paradigm. The next day, answering critics who said the 2022-23 hikes brought inflation down, he replied: &amp;ldquo;The world has changed since 2023.&amp;rdquo;&lt;/p&gt;
&lt;h2 id="ackmans-post-stacks-claims-about-investment-prices-and-policy"&gt;Ackman&amp;rsquo;s post stacks claims about investment, prices and policy
&lt;/h2&gt;&lt;p&gt;The post reads as one sentence, but it stacks claims that need different conditions to hold.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Layer&lt;/th&gt;
 &lt;th&gt;Claim&lt;/th&gt;
 &lt;th&gt;What it needs&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;A. Investment ignores rates&lt;/td&gt;
 &lt;td&gt;Expected ROI overwhelms the cost of capital, so AI investment does not respond to rates&lt;/td&gt;
 &lt;td&gt;Investors fund with their own cash, and expected ROI stays unchanged as rates move&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;B. Hikes raise prices&lt;/td&gt;
 &lt;td&gt;Interest costs are in every price, so each hike pushes inflation up&lt;/td&gt;
 &lt;td&gt;The cost channel is larger and more durable than the demand channel&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;C. So the hike was a mistake&lt;/td&gt;
 &lt;td&gt;The Fed should stop hiking&lt;/td&gt;
 &lt;td&gt;A and B both hold, and the Fed&amp;rsquo;s goal is to curb AI investment&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;A is conditionally true. B has weak evidence. C points the other way the more A is true. We take them in order.&lt;/p&gt;
&lt;h2 id="50bp-is-small-for-a-gpu-and-large-for-a-15-year-data-center-building"&gt;50bp is small for a GPU and large for a 15-year data center building
&lt;/h2&gt;&lt;p&gt;First, the arithmetic. We computed how much the net present value (NPV) of an investment falls when the discount rate rises from 5.0% to 5.5%, using hypothetical projects. Each invests 100 today.&lt;/p&gt;
&lt;div class="fig-scroll" style="overflow-x:auto;-webkit-overflow-scrolling:touch;margin:1.4em 0"&gt;&lt;img src="https://koreainvestinsights.com/post-assets/ackman-ai-rates-fed-mistake-marginal-dollar-is-debt-2026-11-01/npv-50bp.en.svg" alt="NPV and equity-multiple change when the discount rate rises 50bp from 5.0% to 5.5%. 200 back in one year (IRR 100%) -1.0%; 100 a year for five years (IRR 96.6%) -1.8%; GPU server over five years (IRR 19.9%) -4.4%; data center building over 15 years (IRR 10.0%) -12.3%; perpetual-growth equity multiple (9% discount rate, 4% growth) 20.0x to 18.2x, -9.1%." loading="lazy" style="display:block;width:100%;min-width:720px;max-width:none;height:auto;margin:0 auto"&gt;&lt;/div&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Hypothetical project (invest 100)&lt;/th&gt;
 &lt;th style="text-align: right"&gt;IRR&lt;/th&gt;
 &lt;th style="text-align: right"&gt;NPV (5.0%)&lt;/th&gt;
 &lt;th style="text-align: right"&gt;NPV (5.5%)&lt;/th&gt;
 &lt;th style="text-align: right"&gt;Change&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;200 back in one year&lt;/td&gt;
 &lt;td style="text-align: right"&gt;100.0%&lt;/td&gt;
 &lt;td style="text-align: right"&gt;90.5&lt;/td&gt;
 &lt;td style="text-align: right"&gt;89.6&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-1.0%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;100 a year for five years&lt;/td&gt;
 &lt;td style="text-align: right"&gt;96.6%&lt;/td&gt;
 &lt;td style="text-align: right"&gt;332.9&lt;/td&gt;
 &lt;td style="text-align: right"&gt;327.0&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-1.8%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;GPU server: 33.4 a year for five years&lt;/td&gt;
 &lt;td style="text-align: right"&gt;19.9%&lt;/td&gt;
 &lt;td style="text-align: right"&gt;44.6&lt;/td&gt;
 &lt;td style="text-align: right"&gt;42.6&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-4.4%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Data center building: 13.15 a year for 15 years&lt;/td&gt;
 &lt;td style="text-align: right"&gt;10.0%&lt;/td&gt;
 &lt;td style="text-align: right"&gt;36.5&lt;/td&gt;
 &lt;td style="text-align: right"&gt;32.0&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-12.3%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;For a project with a 100% IRR, 50bp trims NPV by 1-2%. Ackman&amp;rsquo;s arithmetic holds. The direct interest cost is also small. The five hyperscalers&amp;rsquo; latest-quarter operating cash flow totals about $198.6 billion, or about $795 billion annualized. Even $500 billion of new debt at 50bp costs $2.5 billion a year, 0.3% of annual operating cash flow.&lt;/p&gt;
&lt;p&gt;The same 50bp cuts the NPV of a 15-year data center building with a 10% IRR by 12%. Measured as the present value of its cash inflows, the drop is 3.3%. The thinner the margin over cost, the more the same rate change weighs on the decision. A perpetual-growth equity multiple, 1/(discount rate - growth), falls from 20.0x to 18.2x, or 9%, at a 9% discount rate and 4% growth. AI capex is a bundle of GPUs (short-lived, high expected return) and buildings, power and land (long-lived, lower expected return). Under these assumptions, &amp;ldquo;rates don&amp;rsquo;t matter&amp;rdquo; holds only for the GPU part.&lt;/p&gt;
&lt;h2 id="in-q3-the-marginal-ai-dollar-became-debt"&gt;In Q3, the marginal AI dollar became debt
&lt;/h2&gt;&lt;p&gt;Ackman&amp;rsquo;s arithmetic has a hidden premise: investors fund with their own money, the cash their operations generate. In Q3 2026 that premise broke.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Company (quarter)&lt;/th&gt;
 &lt;th style="text-align: right"&gt;Operating cash flow&lt;/th&gt;
 &lt;th style="text-align: right"&gt;Purchases of property and equipment&lt;/th&gt;
 &lt;th style="text-align: right"&gt;FCF&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Amazon (Q3)&lt;/td&gt;
 &lt;td style="text-align: right"&gt;42.42&lt;/td&gt;
 &lt;td style="text-align: right"&gt;77.26&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-34.84&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Alphabet (Q3)&lt;/td&gt;
 &lt;td style="text-align: right"&gt;42.29&lt;/td&gt;
 &lt;td style="text-align: right"&gt;51.36&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-9.07&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Microsoft (Sep quarter)&lt;/td&gt;
 &lt;td style="text-align: right"&gt;58.00&lt;/td&gt;
 &lt;td style="text-align: right"&gt;39.37&lt;/td&gt;
 &lt;td style="text-align: right"&gt;+18.63&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Meta (Q3)&lt;/td&gt;
 &lt;td style="text-align: right"&gt;32.79&lt;/td&gt;
 &lt;td style="text-align: right"&gt;33.94&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-1.16&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Oracle (Aug quarter)&lt;/td&gt;
 &lt;td style="text-align: right"&gt;23.1&lt;/td&gt;
 &lt;td style="text-align: right"&gt;28.5&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-5.40&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Total&lt;/td&gt;
 &lt;td style="text-align: right"&gt;198.6&lt;/td&gt;
 &lt;td style="text-align: right"&gt;230.4&lt;/td&gt;
 &lt;td style="text-align: right"&gt;-31.83&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;Figures are in US$ billions. FCF is operating cash flow minus purchases of property and equipment; finance-lease principal is not deducted. Amazon, Alphabet and Meta are from their 8-K earnings exhibits (&lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/1018724/000101872426000038/amzn-20260930xex991.htm" target="_blank" rel="noopener"
 &gt;Amazon&lt;/a&gt;, &lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/1652044/000165204426000080/googexhibit991q32026.htm" target="_blank" rel="noopener"
 &gt;Alphabet&lt;/a&gt;, &lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/1326801/000162828026066700/meta-09302026xexhibit991.htm" target="_blank" rel="noopener"
 &gt;Meta&lt;/a&gt;) and Microsoft from its &lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/789019/000119312526436133/msft-20260930.htm" target="_blank" rel="noopener"
 &gt;10-Q&lt;/a&gt;. Oracle&amp;rsquo;s figures come from press aggregates; we could not confirm them in the filing.&lt;/p&gt;
&lt;p&gt;Capex is 116% of operating cash flow. Of the five, only Microsoft has positive FCF. The gap can be filled with cash on hand, share issuance or debt. The largest identified source is debt: Amazon alone raised $48.1 billion of long-term debt in Q3, more than the five companies&amp;rsquo; combined $31.8 billion shortfall. Its own-definition quarterly FCF is -$33.5 billion (net of incentives).&lt;/p&gt;
&lt;p&gt;For a marginal investor that borrows, rates work first through access to funding. Hyperscalers borrow at investment-grade rates; neoclouds borrow at high-yield rates. The spread widening and investor selectivity Warsh described run through this channel. The second channel is equity value. For a company that funds capex with share issuance and stock compensation, a 9% lower multiple means 9% less funding capacity.&lt;/p&gt;
&lt;p&gt;Ackman conceded the crack himself on CNBC on October 22. He compared the neocloud play to &amp;ldquo;investing in a real estate developer in a highly levered format&amp;rdquo; and said they are very sensitive to their cost of capital. In the same interview he said SpaceX and the Mag 7 are &amp;ldquo;competing with the U.S. government to issue debt, and that&amp;rsquo;s kind of bearish for rates.&amp;rdquo; That is an observation that the investment boom pushes rates up. In the textbook, an investment boom raises the natural rate of interest (the rate that neither heats nor cools the economy), which is a reason for a central bank to raise rates.&lt;/p&gt;
&lt;h2 id="the-winners-roi-is-not-the-marginal-dollars-roi"&gt;The winner&amp;rsquo;s ROI is not the marginal dollar&amp;rsquo;s ROI
&lt;/h2&gt;&lt;p&gt;The original speaks of the ROI of &amp;ldquo;winning the race.&amp;rdquo; The winner&amp;rsquo;s payoff and the expected return of the marginal dollar are different things. In a race with several entrants, each one&amp;rsquo;s expected ROI is the probability of winning times the winner&amp;rsquo;s payoff, minus the cost.&lt;/p&gt;
&lt;p&gt;In an arms race no one can stop first, even when the combined ROI of all entrants is negative, because stopping means losing for sure. So the observation that AI investment ignores rates is correct. But if an arms race is why it ignores rates, the risk that much of the invested capital is destroyed rises with it. US railroads in the 1870s and fiber-optic networks in 1999-2001 are the precedents.&lt;/p&gt;
&lt;p&gt;When capital destruction begins, rates become decisive again, because rates decide how long the indebted entrants can hold on. A single cycle can contain, in sequence, a phase where rates do not matter and a phase where they decide the outcome.&lt;/p&gt;
&lt;h2 id="the-case-that-interest-cost-pass-through-beats-demand-restraint-is-weak"&gt;The case that interest-cost pass-through beats demand restraint is weak
&lt;/h2&gt;&lt;p&gt;Now layer B, the claim that each hike pushes prices up. Researchers do measure a short-run effect in which hikes raise firms&amp;rsquo; working-capital costs and therefore prices. Barth and Ramey named it the cost channel (&lt;a class="link" href="https://www.nber.org/books-and-chapters/nber-macroeconomics-annual-2001-volume-16/cost-channel-monetary-transmission" target="_blank" rel="noopener"
 &gt;Barth &amp;amp; Ramey, NBER Macroeconomics Annual 2001&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;Measured, it is small. US nonfinancial corporate debt was $15.7 trillion at the end of Q2 (&lt;a class="link" href="https://fred.stlouisfed.org/series/BCNSDODNS" target="_blank" rel="noopener"
 &gt;FRED BCNSDODNS&lt;/a&gt;). Passing all of a 25bp increase into prices adds $39 billion a year, 0.12% of $33.0 trillion nominal GDP. If 30% of the debt reprices in the first year (floating rate and refinancing), it is 0.04%. The effect raises the price level while debt reprices at the new rate, and stops when refinancing is done.&lt;/p&gt;
&lt;p&gt;The demand channel behaves differently. A common rough estimate is that a 100bp hike lowers inflation by 0.1-0.4 percentage points after one to two years. For 25bp that is 0.03-0.1 points, and the effect persists while rates stay high. The two figures are not estimated on the same basis, so we use them only to gauge size. Nothing in them suggests the cost channel beats the demand channel for years.&lt;/p&gt;
&lt;p&gt;For a &amp;ldquo;hike more, get more inflation&amp;rdquo; spiral, the cost channel would have to keep beating the demand channel. The recent record points the other way. While the Fed raised rates 525bp in 2022-23, core PCE inflation fell from the 5% range to around 3%.&lt;/p&gt;
&lt;p&gt;Another Ackman remark weakens layer B. On October 22 he said that &amp;ldquo;if we can resolve Iran and Russia-Ukraine, I think rates are coming down, inflation&amp;rsquo;s coming down.&amp;rdquo; That places the cause of inflation in war and energy. If the cause is a supply shock, the question is whether to look through the shock. That is a different question from whether AI has cut the transmission of rates.&lt;/p&gt;
&lt;h2 id="the-more-ai-ignores-rates-the-harder-the-fed-must-press-elsewhere"&gt;The more AI ignores rates, the harder the Fed must press elsewhere
&lt;/h2&gt;&lt;p&gt;Layer C. Suppose layer A is true. AI investment does not respond to rates and bids up power, construction labor and memory prices. Core PCE inflation was 3.1% year over year in September and has been above 2% for 67 straight months since March 2021 (&lt;a class="link" href="https://fred.stlouisfed.org/series/PCEPILFE" target="_blank" rel="noopener"
 &gt;FRED PCEPILFE&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;If inflation pressure persists and current restraint is not enough, the Fed then has to press harder on rate-sensitive housing, autos and small business to free labor, power and materials for AI, or accept higher inflation. The effect Ackman criticized, that hikes hurt only the weaker industries, is what the first path looks like when it works. Monetary policy normally moves resources through the rate-sensitive sectors.&lt;/p&gt;
&lt;p&gt;The second path has a precedent. In 1996-99 the Greenspan Fed judged that the internet was lifting productivity and, apart from one move in March 1997, held off on hikes. That path ended with 175bp of hikes from June 1999 to May 2000 and the investment bust that followed. Either way, the verdict that the Fed erred does not come from layer A. The truer layer A is, the bigger the Fed&amp;rsquo;s job.&lt;/p&gt;
&lt;h2 id="the-fed-raised-the-short-rate-ai-firms-cost-of-capital-is-set-by-the-long-rate"&gt;The Fed raised the short rate; AI firms&amp;rsquo; cost of capital is set by the long rate
&lt;/h2&gt;&lt;p&gt;The &amp;ldquo;5% versus 5.5%&amp;rdquo; summary folds the Fed&amp;rsquo;s short rate and the market&amp;rsquo;s long rate into one variable. The Fed raised the short rate by 25bp. A hyperscaler&amp;rsquo;s cost of capital is set by 10- to 30-year yields and credit spreads. The 30-year Treasury yield hit 5.67% on October 7, the highest since 2002, and was 5.32% on October 29. The 10-year was 4.98% on October 29 (&lt;a class="link" href="https://fred.stlouisfed.org/series/DGS30" target="_blank" rel="noopener"
 &gt;FRED DGS30&lt;/a&gt;, &lt;a class="link" href="https://fred.stlouisfed.org/series/DGS10" target="_blank" rel="noopener"
 &gt;DGS10&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;Long yields move with growth expectations, fiscal policy, the term premium and the volume of bond issuance. Ackman named Big Tech and SpaceX issuance as a driver of higher rates. Fed Governor Christopher Waller was reported on October 13 as citing growth expectations, fiscal concerns and competition for investor dollars from massive AI bond issuance as the main drivers of higher government yields (&lt;a class="link" href="https://www.scotsmanguide.com/news/feds-waller-says-ai-fueled-inflation-could-persist-for-years/" target="_blank" rel="noopener"
 &gt;Scotsman Guide, Oct 13&lt;/a&gt;). No estimate splits the contributions. Still, the Fed alone did not make the &amp;ldquo;5% versus 5.5%&amp;rdquo; gap, and AI&amp;rsquo;s own bond issuance has been named as one driver.&lt;/p&gt;
&lt;h2 id="ackmans-portfolio-is-built-from-rate-sensitive-assets"&gt;Ackman&amp;rsquo;s portfolio is built from rate-sensitive assets
&lt;/h2&gt;&lt;p&gt;Whether a claim is right is a separate question from where its author stands, but both matter. Ackman runs Pershing Square and has an economic interest in the stocks it holds.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Exposure&lt;/th&gt;
 &lt;th&gt;Source&lt;/th&gt;
 &lt;th&gt;Link to rates&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Microsoft, Amazon and Meta, 31.63% combined&lt;/td&gt;
 &lt;td&gt;&lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/2026053/000117266126003790/xslForm13F_X02/infotable.xml" target="_blank" rel="noopener"
 &gt;13F, Jun 30&lt;/a&gt;&lt;/td&gt;
 &lt;td&gt;Equity multiples built on long-dated growth&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Uber 12.72%, Brookfield 12.58%, Howard Hughes 10.23%&lt;/td&gt;
 &lt;td&gt;Same 13F&lt;/td&gt;
 &lt;td&gt;Brookfield is real infrastructure and alternatives; Howard Hughes is real estate development&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Fannie Mae and Freddie Mac, &amp;ldquo;well north of 200 million shares&amp;rdquo; combined&lt;/td&gt;
 &lt;td&gt;CNBC, Oct 22&lt;/td&gt;
 &lt;td&gt;Direct exposure to mortgage rates and spreads&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Pershing Square Holdings (PSH) NAV -6.9% year to date&lt;/td&gt;
 &lt;td&gt;&lt;a class="link" href="https://pershingsquareholdings.com/?p=5978" target="_blank" rel="noopener"
 &gt;PSH weekly NAV, Oct 20&lt;/a&gt;&lt;/td&gt;
 &lt;td&gt;Relative underperformance in 2026&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;On October 22 Ackman said &amp;ldquo;our relative underperformance this year in large part is driven by Fannie and Freddie.&amp;rdquo; Fannie Mae fell 72%, from $11.25 on January 5 to $3.15 on October 14, and closed at $4.80 on October 30. The manager, Pershing Square (PS), listed on the NYSE on April 29 together with the closed-end fund Pershing Square USA (PSUS) (&lt;a class="link" href="https://www.sec.gov/Archives/edgar/data/0002026053/000114036126018103/ny20040230x30_424b4.htm" target="_blank" rel="noopener"
 &gt;SEC 424B4&lt;/a&gt;). The manager&amp;rsquo;s share price is also tied to fund performance and inflows. Rate hikes press on his whole book in the same direction.&lt;/p&gt;
&lt;p&gt;His Fed commentary has also moved with his positions.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;When&lt;/th&gt;
 &lt;th&gt;Remark&lt;/th&gt;
 &lt;th&gt;Position at the time&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Oct 2021&lt;/td&gt;
 &lt;td&gt;The Fed should start raising rates as soon as possible (&lt;a class="link" href="https://www.cnbc.com/2021/10/29/bill-ackman-calls-for-the-fed-to-start-raising-interest-rates-as-soon-as-possible.html" target="_blank" rel="noopener"
 &gt;CNBC&lt;/a&gt;)&lt;/td&gt;
 &lt;td&gt;Held hedges against rising rates; made $2.7 billion on rate trades in 2022 (&lt;a class="link" href="https://www.kitco.com/news/off-the-wire/2023-02-10/ackmans-pershing-square-reaps-27-bln-rate-trade-down-9-2022-letter" target="_blank" rel="noopener"
 &gt;Reuters via Kitco&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Aug 2023&lt;/td&gt;
 &lt;td&gt;Disclosed a short in 30-year Treasuries&lt;/td&gt;
 &lt;td&gt;Long-bond short&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Oct 2023&lt;/td&gt;
 &lt;td&gt;Covered the short: too much risk in the world to stay short bonds (&lt;a class="link" href="https://www.cnbc.com/2023/10/23/bill-ackman-covers-bet-against-treasurys-says-too-much-risk-in-the-world-to-bet-against-bonds.html" target="_blank" rel="noopener"
 &gt;CNBC&lt;/a&gt;)&lt;/td&gt;
 &lt;td&gt;Covered&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Nov 2023&lt;/td&gt;
 &lt;td&gt;Hard-landing risk unless the Fed cuts soon (&lt;a class="link" href="https://fortune.com/2023/11/28/bill-ackman-fed-interest-rate-cuts-first-quarter-hard-soft-landing" target="_blank" rel="noopener"
 &gt;Fortune&lt;/a&gt;)&lt;/td&gt;
 &lt;td&gt;Long equities&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Aug 2024&lt;/td&gt;
 &lt;td&gt;The Fed was too slow to raise and is too slow to cut (&lt;a class="link" href="https://investinglive.com/centralbank/bill-ackman-says-fed-was-too-slow-to-raise-rates-now-too-slow-to-lower-them-20240805/" target="_blank" rel="noopener"
 &gt;investinglive&lt;/a&gt;)&lt;/td&gt;
 &lt;td&gt;Long equities&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Sep 2026&lt;/td&gt;
 &lt;td&gt;The 25bp hike might be a mistake&lt;/td&gt;
 &lt;td&gt;Long megacaps, Fannie, Freddie and real estate; behind for the year&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;In the record we checked, his Fed remarks have never run against the direction of his book. That alignment alone does not make them wrong. But readers should weigh them knowing his interests. We did not check his 2025 remarks.&lt;/p&gt;
&lt;p&gt;His worldview points the same way. On CNBC in October he said &amp;ldquo;we&amp;rsquo;re asymptotically approaching superintelligence.&amp;rdquo; What follows is our interpretation. Treat AI as a contest between nations and the cost of capital looks like a wartime-finance variable, with the central bank cast as an institution that should support the effort. The Fed&amp;rsquo;s 1942-51 cap on Treasury yields is the historical form of that view.&lt;/p&gt;
&lt;p&gt;The politics overlap too. Ackman endorsed Donald Trump in July 2024, and Warsh is the chair Trump nominated, confirmed by the Senate in May. Yet the Fed&amp;rsquo;s center of gravity sits on the hawkish side of Warsh. The October 28 vote to hold was 10-2, and both dissenters, Governor Lisa Cook and Cleveland Fed President Beth Hammack, wanted a 25bp hike (&lt;a class="link" href="https://www.federalreserve.gov/newsevents/pressreleases/monetary20261028a.htm" target="_blank" rel="noopener"
 &gt;Fed statement, Oct 28&lt;/a&gt;).&lt;/p&gt;
&lt;h2 id="a-month-later-ackman-named-ai-capex-as-a-cause-of-inflation-and-prescribed-more-supply"&gt;A month later, Ackman named AI capex as a cause of inflation and prescribed more supply
&lt;/h2&gt;&lt;p&gt;Line up Ackman&amp;rsquo;s remarks after the September 24 post and the emphasis shifts.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;When&lt;/th&gt;
 &lt;th&gt;Venue&lt;/th&gt;
 &lt;th&gt;Gist&lt;/th&gt;
 &lt;th&gt;Versus Sep 24&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Sep 25&lt;/td&gt;
 &lt;td&gt;X reply&lt;/td&gt;
 &lt;td&gt;Answered critics who said the 2022-23 hikes cut inflation: &amp;ldquo;The world has changed since 2023&amp;rdquo;&lt;/td&gt;
 &lt;td&gt;No change&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Oct 22&lt;/td&gt;
 &lt;td&gt;CNBC interview&lt;/td&gt;
 &lt;td&gt;A 5% 10-year is still a low cost of capital. Neoclouds are very sensitive to their cost of capital. Big Tech and SpaceX issuance is bearish for rates. Resolve Iran and Russia, and rates and inflation come down&lt;/td&gt;
 &lt;td&gt;Conceded that marginal investors are rate-sensitive. Placed the cause of inflation in war and energy&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Oct 28&lt;/td&gt;
 &lt;td&gt;&lt;a class="link" href="https://shows.acast.com/equity-mates/episodes/bill-ackman-on-building-the-next-berkshire-ai-his-next-10-ba" target="_blank" rel="noopener"
 &gt;Equity Mates podcast&lt;/a&gt; (chapter at 16:42)&lt;/td&gt;
 &lt;td&gt;Said the 25bp hike may have signaled resolve and may have been a mistake, without committing. Named war-driven energy and unprecedented AI infrastructure spending as causes of inflation. Proposed more supply, through cheaper housing permits and development&lt;/td&gt;
 &lt;td&gt;Added &amp;ldquo;a signal of resolve&amp;rdquo; to &amp;ldquo;may have been a mistake.&amp;rdquo; Named AI investment as a cause of inflation. Moved the prescription from monetary to supply policy&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;We confirmed the October 28 podcast remarks from the episode&amp;rsquo;s chapter titles and a summary by a Korean listener; we did not hear the audio. If the summary is right, Ackman named AI infrastructure spending as a cause of inflation, which is the Fed hawks&amp;rsquo; premise. We found no rebuttal from Ackman after Warsh&amp;rsquo;s remarks.&lt;/p&gt;
&lt;h2 id="the-fed-hawks-accepted-ackmans-premise-and-reversed-his-conclusion"&gt;The Fed hawks accepted Ackman&amp;rsquo;s premise and reversed his conclusion
&lt;/h2&gt;&lt;p&gt;The reactions split on whether they accept the premise, and on what they conclude from it.&lt;/p&gt;
&lt;div class="fig-scroll" style="overflow-x:auto;-webkit-overflow-scrolling:touch;margin:1.4em 0"&gt;&lt;img src="https://koreainvestinsights.com/post-assets/ackman-ai-rates-fed-mistake-marginal-dollar-is-debt-2026-11-01/debate-map.en.svg" alt="Map of reactions. Rebuttal: Chair Warsh says AI capex is rate-sensitive too; Cullen Roche rebuts the claim that hikes raise inflation. Accept the premise, critical of hikes: Ackman, Mark Zandi, Gregory Daco. Accept the premise, stay restrictive: Hammack, Waller, Cook. Data: Q3 nonresidential fixed investment +12.7%, equipment +24.7%, though most of Q3 predates the hike." loading="lazy" style="display:block;width:100%;min-width:720px;max-width:none;height:auto;margin:0 auto"&gt;&lt;/div&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Camp&lt;/th&gt;
 &lt;th&gt;Who&lt;/th&gt;
 &lt;th&gt;Gist&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Rebuttal: capex is rate-sensitive&lt;/td&gt;
 &lt;td&gt;Kevin Warsh (Fed Chair, Oct 28)&lt;/td&gt;
 &lt;td&gt;&amp;ldquo;Interest rates get in the cracks.&amp;rdquo; On the new technology and the investment surge: &amp;ldquo;do interest rates matter? I believe they do,&amp;rdquo; and &amp;ldquo;I don&amp;rsquo;t believe that they&amp;rsquo;re somehow divorced from the rest of the economy.&amp;rdquo; He noted wider credit spreads for some assets after the September hike (&lt;a class="link" href="https://finance.yahoo.com/economy/policy/live/fed-meeting-live-updates-interest-rates-held-steady-in-split-vote-warsh-says-politics-played-no-role-ahead-of-midterm-elections-191946928.html" target="_blank" rel="noopener"
 &gt;Yahoo live, Oct 28&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Rebuttal: hikes do not raise inflation&lt;/td&gt;
 &lt;td&gt;Cullen Roche (Discipline Funds)&lt;/td&gt;
 &lt;td&gt;Cited post-2022 disinflation and Turkey as counterexamples to &amp;ldquo;hikes raise inflation&amp;rdquo; (&lt;a class="link" href="https://www.tipranks.com/news/bill-ackman-asks-if-the-ai-boom-is-breaking-the-feds-inflation-playbook" target="_blank" rel="noopener"
 &gt;TipRanks, Sep 25&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Accepts premise, critical of hikes&lt;/td&gt;
 &lt;td&gt;Mark Zandi (Moody&amp;rsquo;s)&lt;/td&gt;
 &lt;td&gt;&amp;ldquo;AI is running on its own dynamic and expectations for future profits are quite high&amp;rdquo;; hyperscalers can pay higher rates (&lt;a class="link" href="https://finance.yahoo.com/economy/policy/article/is-ai-investment-sensitive-to-interest-rates-warsh-says-yes-but-others-disagree-162933738.html" target="_blank" rel="noopener"
 &gt;Yahoo, Oct 29&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Accepts premise, critical of hikes&lt;/td&gt;
 &lt;td&gt;Gregory Daco (EY-Parthenon)&lt;/td&gt;
 &lt;td&gt;More hikes &amp;ldquo;could create further strain for already constrained, interest-sensitive sectors&amp;rdquo; (&lt;a class="link" href="https://www.tikr.com/blog/bill-ackman-ais-near-infinite-roi-could-undo-fed-strategy" target="_blank" rel="noopener"
 &gt;TIKR, Sep 28&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Data&lt;/td&gt;
 &lt;td&gt;Joe Brusuelas (RSM)&lt;/td&gt;
 &lt;td&gt;On the idea that AI investment is rate-sensitive: &amp;ldquo;Sure does not look that way following the September rate hike.&amp;rdquo; Q3 nonresidential fixed investment rose 12.7% annualized and equipment 24.7% (BEA advance, &lt;a class="link" href="https://fred.stlouisfed.org/series/A008RL1Q225SBEA" target="_blank" rel="noopener"
 &gt;FRED&lt;/a&gt;). Most of Q3 predates the hike, so this does not show a response to rates&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Accepts premise, stay restrictive&lt;/td&gt;
 &lt;td&gt;Beth Hammack (Cleveland Fed, Oct 13)&lt;/td&gt;
 &lt;td&gt;&amp;ldquo;I&amp;rsquo;m under no illusion that raising interest rates will reopen the Strait of Hormuz&amp;rdquo; or materially stop data center demand. Still judges a restrictive stance appropriate; dissented for a hike in October (&lt;a class="link" href="https://www.scotsmanguide.com/news/feds-waller-says-ai-fueled-inflation-could-persist-for-years/" target="_blank" rel="noopener"
 &gt;Scotsman Guide&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Accepts premise, stay restrictive&lt;/td&gt;
 &lt;td&gt;Christopher Waller (Fed Governor, Oct 13)&lt;/td&gt;
 &lt;td&gt;Hyperscalers are &amp;ldquo;still planning on spending trillions of dollars over the next three years,&amp;rdquo; so demand pressure may remain and &amp;ldquo;we just have to decide how we want to respond to that&amp;rdquo; (&lt;a class="link" href="https://www.scotsmanguide.com/news/feds-waller-says-ai-fueled-inflation-could-persist-for-years/" target="_blank" rel="noopener"
 &gt;Scotsman Guide&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Accepts premise, stay restrictive&lt;/td&gt;
 &lt;td&gt;Lisa Cook (Fed Governor)&lt;/td&gt;
 &lt;td&gt;Sees AI as already causing inflationary pressure; dissented for a hike in October (&lt;a class="link" href="https://investorplace.com/2026/10/fed-holds-ais-bigger-risk-isnt-rates/" target="_blank" rel="noopener"
 &gt;InvestorPlace, Oct 28&lt;/a&gt;)&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;Hammack and Waller also accept that core AI capex does not move much on 25bp. The split is over the conclusion. Ackman said the hike might be a mistake; Zandi and Daco see the burden of hikes landing on other sectors more than on AI. Hammack, Waller and Cook draw the opposite conclusion from the same observation: stay restrictive. It is the same structure as layer C above.&lt;/p&gt;
&lt;p&gt;Warsh&amp;rsquo;s rebuttal aims at a different layer. He said rates get into the cracks, and in the Q3 numbers the cracks are the marginal investors who borrow. Our view takes the hawks&amp;rsquo; conclusion and Warsh&amp;rsquo;s diagnosis together. We accept that core investment funded with own cash ignores rates. But in Q3 a share of capex exceeded operating cash, that share was mostly debt-funded, and it is sensitive to rates.&lt;/p&gt;
&lt;h2 id="verdict-the-dominant-variable-is-where-the-marginal-dollar-comes-from"&gt;Verdict: the dominant variable is where the marginal dollar comes from
&lt;/h2&gt;&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;Variable&lt;/th&gt;
 &lt;th&gt;Ackman&amp;rsquo;s claim&lt;/th&gt;
 &lt;th&gt;Verdict&lt;/th&gt;
 &lt;th&gt;Evidence&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Core capex funded with own cash (within a year)&lt;/td&gt;
 &lt;td&gt;Ignores rates&lt;/td&gt;
 &lt;td&gt;True&lt;/td&gt;
 &lt;td&gt;In hypothetical examples, 50bp moves GPU NPV -1% to -4%; interest cost is 0.3% of operating cash flow&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Debt-funded marginal capex&lt;/td&gt;
 &lt;td&gt;Not in the X post; on CNBC, conceded neoclouds are sensitive&lt;/td&gt;
 &lt;td&gt;Sensitive&lt;/td&gt;
 &lt;td&gt;Five-company FCF -$31.8B, capex/operating cash flow 116%, building NPV -12%, equity multiple -9%&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Does the cost channel beat the demand channel?&lt;/td&gt;
 &lt;td&gt;Yes (spiral)&lt;/td&gt;
 &lt;td&gt;Not supported&lt;/td&gt;
 &lt;td&gt;Cost channel is 0.04-0.12% of GDP in size; inflation fell during the 525bp of hikes in 2022-23&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Fed&amp;rsquo;s right move in a rate-insensitive boom&lt;/td&gt;
 &lt;td&gt;Stop hiking&lt;/td&gt;
 &lt;td&gt;Opposite&lt;/td&gt;
 &lt;td&gt;With insensitive investment and persistent inflation pressure, restraining rate-sensitive sectors is the normal working of policy&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Short rate vs long rate&lt;/td&gt;
 &lt;td&gt;One variable&lt;/td&gt;
 &lt;td&gt;Must separate&lt;/td&gt;
 &lt;td&gt;Long yields move with growth, fiscal policy, the term premium and issuance; Ackman named issuance himself&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;The dominant variable is the second row. If the marginal dollar is own cash, &amp;ldquo;rates don&amp;rsquo;t matter&amp;rdquo; holds for all AI capex. Even then, the fourth row stops it short of &amp;ldquo;the Fed erred.&amp;rdquo; If the marginal dollar is debt, &amp;ldquo;rates don&amp;rsquo;t matter&amp;rdquo; breaks at the margin first, and Warsh&amp;rsquo;s &amp;ldquo;cracks&amp;rdquo; are right. The Q3 2026 numbers point to debt.&lt;/p&gt;
&lt;p&gt;What would change the verdict:&lt;/p&gt;
&lt;ul&gt;
&lt;li&gt;Toward Ackman: the five-company FCF total turns positive in H1 2027, so the marginal dollar is own cash again. Or core PCE falls quickly below 2.5% without further hikes.&lt;/li&gt;
&lt;li&gt;Further toward Warsh: high-yield spreads move above 350bp or a neocloud bond deal fails. Or one hyperscaler lowers its 2027 capex plan. High-yield spreads widened to 340bp on October 22 and were 326bp on October 29 (&lt;a class="link" href="https://fred.stlouisfed.org/series/BAMLH0A0HYM2" target="_blank" rel="noopener"
 &gt;FRED&lt;/a&gt;).&lt;/li&gt;
&lt;/ul&gt;
&lt;h2 id="what-investors-keep-risk-shows-up-in-the-leveraged-layer-before-the-core"&gt;What investors keep: risk shows up in the leveraged layer before the core
&lt;/h2&gt;&lt;p&gt;A useful observation survives from Ackman&amp;rsquo;s post. If the Fed&amp;rsquo;s 25bp does not change hyperscaler core capex, one-year visibility for memory and AI-chip demand looks better. On Q3 calls, Amazon said it expects to &amp;ldquo;spend meaningfully more in 2027,&amp;rdquo; Alphabet&amp;rsquo;s CFO said capex will increase significantly in 2027 (as reported), and Microsoft said capex will accelerate in the second half of its fiscal year (&lt;a class="link" href="https://www.fool.com/earnings/call-transcripts/2026/10/30/amazon-amzn-q3-2026-earnings-call-transcript/" target="_blank" rel="noopener"
 &gt;Amazon call&lt;/a&gt;, &lt;a class="link" href="https://www.fool.com/earnings/call-transcripts/2026/10/29/microsoft-msft-q1-2027-earnings-call-transcript/" target="_blank" rel="noopener"
 &gt;Microsoft call&lt;/a&gt;).&lt;/p&gt;
&lt;p&gt;Risk appears first where rates seep into the cracks: neoclouds building data centers with debt, companies that depend on equity issuance, and buildings and power tied up in long leases. The next point is inference. If the leveraged layer wobbles while core capex holds, suppliers&amp;rsquo; pricing power can weaken before demand falls, because buyers with negative FCF negotiate next year&amp;rsquo;s supply prices.&lt;/p&gt;
&lt;table&gt;
 &lt;thead&gt;
 &lt;tr&gt;
 &lt;th&gt;When&lt;/th&gt;
 &lt;th&gt;What to check&lt;/th&gt;
 &lt;th&gt;Link to the verdict&lt;/th&gt;
 &lt;/tr&gt;
 &lt;/thead&gt;
 &lt;tbody&gt;
 &lt;tr&gt;
 &lt;td&gt;Mid-November&lt;/td&gt;
 &lt;td&gt;Pershing Square 13F as of Sep 30: weights in Fannie, Freddie and megacaps&lt;/td&gt;
 &lt;td&gt;Size of the interest&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;December FOMC&lt;/td&gt;
 &lt;td&gt;Hike or not, and the number of dissents&lt;/td&gt;
 &lt;td&gt;The Fed majority&amp;rsquo;s conclusion&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Jan-Feb 2027&lt;/td&gt;
 &lt;td&gt;Sign of five-company FCF in Q4 results and 2027 capex totals&lt;/td&gt;
 &lt;td&gt;Does the marginal dollar return to own cash?&lt;/td&gt;
 &lt;/tr&gt;
 &lt;tr&gt;
 &lt;td&gt;Ongoing&lt;/td&gt;
 &lt;td&gt;High-yield spreads vs 350bp; neocloud bond yields and issuance&lt;/td&gt;
 &lt;td&gt;Are the cracks widening?&lt;/td&gt;
 &lt;/tr&gt;
 &lt;/tbody&gt;
&lt;/table&gt;
&lt;p&gt;&lt;strong&gt;Do you expect the marginal AI dollar to return to own cash in 2027, or the debt share to keep rising? Tell us the sign you expect for the five hyperscalers&amp;rsquo; combined FCF in H1 2027, and why.&lt;/strong&gt;&lt;/p&gt;
&lt;p&gt;This is research and information, not investment advice. The NPV calculations use hypothetical cash flows, and the sizes of the cost and demand channels are rough estimates. We confirmed the X post through press quotes and the October 28 podcast remarks through chapter titles and a secondary summary. Ackman runs Pershing Square and has an economic interest in its holdings. We used Anthropic&amp;rsquo;s Claude for research and editing. Anthropic appears here as a buyer of AI compute.&lt;/p&gt;</description></item></channel></rss>