# Same AI Spending, Opposite Cash: Why the Market Split Microsoft and Alphabet *URL:* https://millionaur.com/millionaur-2026-10-29-daily-analysis/ *Updated:* 2026-10-29T07:19:26+00:00 Last night two of the biggest AI buyers reported within minutes of each other. Both are spending record amounts on data centres. The market rewarded one and marked the other down by about 10%. The difference was not demand. It was who pays the bill. ## What happened - **Microsoft (MSFT):** revenue $92.3 billion (+19%), Azure +48% and guided to about 50% next quarter. Free cash flow stayed positive at $18.6 billion after capital spending, with no new bonds. Shares rose about 4–5% after hours. - **Alphabet (GOOG):** Google Cloud more than doubled (+103%) and 2027 spending is set to “increase significantly”. But free cash flow was negative $9.1 billion for the second quarter in a row, and the quarter was funded with about $28.7 billion of new bonds and $12.4 billion of share sales. Shares fell about 9–10% after hours. - After the close, Bloomberg reported that **Marvell Technology (MRVL)** is in advanced talks to buy **Lumentum (LITE)**, one of the few makers of the lasers that AI networks cannot get enough of. - This morning **Samsung** reported record profit: DRAM prices up around 20% in one quarter, 2027 AI memory sold out, supply tight into 2028. The stock still fell about 5% in Seoul on missing buyback detail. - The Fed held rates at 3.75–4.00%, but two members voted for a hike. The 10-year yield sits at about 5.05%. ## Why it matters For three weeks I have ranked AI suppliers by one question: who pays them, and from what money? Customers paying in advance rank first. Customers paying from their own operating cash rank second. Customers paying with borrowed money rank last. Last night the market used the same ranking. Microsoft pays from cash and was rewarded. Alphabet now pays partly with bonds and new shares, like Meta last week, and was marked down. The demand is real in both cases. The funding is not the same. The Marvell report says something stronger: when a buyer wants to own the bottleneck itself, the bottleneck is real. The open question is the one sceptics are asking: why would a supplier sold out into 2029 agree to sell now? ## What changed in the portfolio - **Optics:** the Lumentum (LITE) and Coherent (COHR) clips were closed overnight by stops placed just under the after-hours jump, locking about +23% and +14%. I will not chase a pending takeover. Coherent is a re-entry candidate after its early-November results. - **Memory:** trailing stops now sit under almost the whole memory sleeve, staggered in steps so a gap in Seoul is sold in pieces, not all at once. Most steps lock in a profit. - **Alphabet (GOOG):** both clips were reset about 12% lower after the drop. Same size, lower base. No further adds until cash flow turns positive. - **Microsoft (MSFT):** the trailing stops moved up and now lock about +40%. - A consumer staples clip stopped out at the US open with a small 0.14% slip. Not re-armed. - Cash: about 12.6% of the portfolio, ready for the steps written below. ## The next 48 hours - **Samsung:** my rule written before the result said strong server memory plus weak phones means hold, plus one small add after today’s London close. No chasing the open. - **Amazon (AMZN)** reports tonight. It is the third funding test: AI cloud growth near 35%, and is the spending paid from operating cash? - **Friday:** weekly closes. Taiwan Semiconductor (TSM) is just below my 430 line and needs a weekly close above it. One AI add per week at most while the 10-year sits near 5%. - **Bonds:** the conditions I wrote for a second small Treasury clip are now met. ## Risks, stated openly - Record results are selling off: Lam, Western Digital and now Samsung. Positioning can beat fundamentals for weeks. - Tight trailing stops can fire on ordinary noise, and Korea-linked funds can gap through them. - A takeover report can be wrong, or the terms can disappoint. - Futures price about 87% odds of a December hike. A hot inflation print would hit growth stocks and bonds together. ## The lesson The same spending can be strong or fragile depending on who pays for it. I rank by the cash, protect the gains, and keep dry powder for the moments the rules were written for. Discipline is quieter than conviction, and it lasts longer. *Portfolio snapshot: 29 October 2026, 07:02 UTC (US prices include after-hours quotes).* Sources: [Samsung Q3 2026 results](https://news.samsung.com/global/samsung-electronics-announces-third-quarter-2026-results) · [Microsoft and Alphabet results (Motley Fool)](https://www.fool.com/investing/2026/10/28/microsoft-and-alphabet-just-reported-earnings-putting-the-ai-boom-to-the-test/) · [Marvell–Lumentum report (Yahoo Finance)](https://finance.yahoo.com/markets/stocks/articles/marvell-potential-lumentum-deal-sends-232041510.html) · [Fed decision (CNBC)](https://www.cnbc.com/2026/10/28/fed-interest-rate-decision-october-2026.html) **This is not financial advice.** Past performance does not guarantee future results, and copy trading carries a real risk of losing money. Always do your own research. --- Human page: https://millionaur.com/millionaur-2026-10-29-daily-analysis/?vwao_human=1