Alphabet, Amazon, Meta, Microsoft and Oracle had $32.7 billion of free cash flow over the four quarters to the third quarter, down about 84% from $201.1 billion a year earlier, according to a 15 November report by the Hyundai Research Institute, a Seoul think tank. The institute says the US AI investment boom could enter a correction around 2028.
The report, titled “We Must Prepare for a Possible Correction in the US AI Investment Boom” in the English rendering by BigGo Finance, projects the five companies’ AI-related capital spending rising from $789.6 billion this year to $1.3079 trillion in 2029, up 65.6%.
Business Pill · Free Cash Flow
An evergreen Business Engineer explainer on free cash flow. It teaches the general idea only and does not discuss this report.
The key insight: As we read the report, the institute separates demand from capacity to pay: it expects demand and profitability to stay sound through 2028 while financial health, financing and regulation worsen, so its correction case rests on cash and constraints rather than on AI use falling.
The institute scored AI investment against five conditions: demand and profitability, financial health, financing, regulation, and infrastructure and supply chains, according to Edaily’s account of the report. It expects demand and profitability to stay sound through 2027 and 2028, financial health and regulation to move from neutral to unfavourable, financing to slip from favourable to neutral, and infrastructure and supply chains to stay unfavourable.
Over the same four quarters, capital spending equalled 95.5% of the five companies’ operating cash flow, the report says, and in the third quarter alone all of them except Microsoft had negative free cash flow.
Combined free cash flow of Alphabet, Amazon, Meta, Microsoft and Oracle over four quarters: $201.1 billion to Q3 2025 vs $32.7 billion to Q3 2026, per the Hyundai Research Institute; the about-84% fall is FourWeekMBA arithmetic.
The Borrowing and the Build Constraints
The five issued $182 billion of corporate bonds from January to October, about double the $93.25 billion they issued in all of 2025, according to the report. It says the spread between new bond yields and the 10-year US Treasury yield is 0.65 percentage points for Microsoft, while Oracle’s widened from 1.45 points in February to 3.10 points this month, more than double by our arithmetic.
About 70% of announced US data-centre capacity has not broken ground, the report says, and 38.4% of planned capacity sits in Texas, Georgia and Ohio, where it says tighter regulation came into focus after this month’s midterm elections. We have mapped the levers states hold over the data-centre build.
The report expects AI investment to keep growing through next year, citing rising cloudrevenue and plans already committed. Around 2028, it says, delayed profitability, financing burdens and regulatory and infrastructure limits could combine, leading companies to postpone or trim new spending case by case. It compares this to how investment adjusted after earlier general-purpose technologies such as the internet. Earlier this week we covered a Federal Reserve warning about an AI correction.
The institute says a cut in US AI spending could hit South Korea’s chip exports, production, investment and jobs, and could spill into Korean financial markets.
The Structural Read
As we read the cash figures, the five companies spent 95.5% of their operating cash flow on capital spending over four quarters, which leaves little room before outside funding becomes the marginal source.
As we read the comparison with our own count, the single-quarter shortfall we measured from 10-Qs in October is about the size of the whole four-quarter free cash flow the institute reports, though the two use different periods.
Hyundai Research Institute report title (English rendering by BigGo Finance)
“We Must Prepare for a Possible Correction in the US AI Investment Boom”
Three Implications
FOR CHIP AND MEMORY SUPPLIERS The institute says a cut in US AI spending could hit South Korea’s chip exports, production, investment and jobs.
FOR LENDERS The report says the five issued $182 billion of bonds from January to October, about double their full-year 2025 total.
FOR DATA-CENTRE BUILDERS The report says about 70% of announced US capacity has not broken ground and 38.4% sits in Texas, Georgia and Ohio.
The framework says: “This is not a continuation of the cloud capex cycle. This is a different cycle, with different physics.”
As we read it, the institute’s report tests whether that cycle can keep being funded from cash; the framework does not discuss this report.
What Is Not Established
The report’s full text was not available to us; figures here come from Edaily’s Korean account, which we translated, and BigGo Finance’s English account. Edaily defines free cash flow as cash from operations minus capital spending; the report’s underlying data source is not stated in either account. The 2028 timing is the institute’s assessment, not a forecast by this publication, and none of the five companies has commented on the report.
This Week In The AI Supercycle
AI Lost Its Central Banker, the Lab Oligopoly, and the Future of Open Source
This week’s issue: the Nvidia leadership change, read as the AI build-out losing its central banker: who now finances the GPU landlords, backstops the data centres and sponsors open models, and why two labs are becoming the bank.
The Hyundai Research Institute puts five major US AI spenders’ four-quarter free cash flow at $32.7 billion, down about 84%, and says the investment boom could enter a correction around 2028, according to Edaily’s account of its 15 November report.
A note on sourcing. Figures come from the Hyundai Research Institute’s 15 November 2026 weekly economic review as reported by Edaily (Korean, translated by us) and BigGo Finance (English); we did not read the full report. Comparisons with our earlier filing counts and percentage changes are our arithmetic. None of the five companies was contacted for this piece.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
