Google, Intel y Reddit se hunden… por culpa de la IA

Summary

The video argues that the recent sell-off in technology stocks is not clear evidence that AI demand is collapsing. Instead, AI’s rapid adoption is disrupting some existing business models and driving up the cost of scarce components. The longer-term risk, the presenter says, is whether traditional tech businesses can continue funding the investment boom if AI erodes their own cash flows.

Market moves and company results

• The Nasdaq had fallen for three weeks, roughly 8% from its early-October high, and recorded its worst month since March 2015, according to the video.
• Alphabet (Google) fell 12.6% after earnings, wiping out roughly $500 billion in market value in one session. Revenue grew 25% and operating profit 30%, but search revenue growth slowed to 12%, from 17% the prior quarter, and missed expectations.
  • Search generated about $63.4 billion in quarterly revenue—nearly half of Alphabet’s total—but AI-generated answers may earn less per query because users click fewer links and ads.
  • Google Cloud revenue rose 103% year over year to $30.7 billion. Contracted work awaiting execution exceeded $560 billion.
  • Alphabet planned to invest about $200 billion this year, mainly in data centers, and indicated spending would be higher in 2027. In the first nine months, capital investment of nearly $132 billion exceeded cash generated of $127 billion. Third-quarter free cash flow was negative by more than $9 billion.
• ServiceNow fell 27%, despite quarterly revenue of about $4.1 billion, up 20% year over year. Its remaining contract backlog expected to convert to revenue over the next 12 months was forecast to grow 18%, below analysts’ expectation of 20%. The concern is that AI agents could reduce the number of employees—and therefore software licences—customers need.
• Reddit fell nearly 30%. Revenue was up 53% year over year, with nearly $900 million in quarterly revenue, but management reported weakening visits from search engines. In October, Reddit averaged over one million fewer search-driven visits per week. It expects next-quarter revenue to come in below prior market expectations and is testing login requirements in some markets. Its share price was described as about 60% below its level at the start of 2026.
• Intel suffered its worst trading day since 1974, losing about $140 billion in market value. Revenue was around $16.5 billion, up 21%, and its data-center and AI business grew 78%. However, its fourth-quarter revenue forecast of $15–16 billion was below analysts’ estimate of roughly $17 billion.
  • Intel remains exposed to PCs. The video cites global PC sales down about 20% in the third quarter year over year, with declines of 12% in 2026 and another 10% in 2027 forecast.
• Higher memory costs also weighed on hardware-related stocks: Dell fell nearly 11%, HP about 10%, and AMD and Qualcomm about 11% each.
• Even some perceived AI beneficiaries declined: Nvidia fell about 4%, Micron more than 5%, and Sandisk 14%, despite the latter reporting 61% growth in its data-center business.

Why AI growth can hurt some companies

The presenter describes two channels of disruption:

1. AI cannibalizes existing business models.

   • AI answers can reduce clicks on Google search ads and make visits to Reddit threads unnecessary.
   • AI agents may automate work previously done by employees, reducing demand for per-user software licences such as ServiceNow’s.
   • Companies may need to adopt AI to avoid losing users to competitors, even if AI features currently monetize less effectively.

2. AI competes for scarce inputs, raising costs elsewhere.

   • AI servers require large volumes of high-bandwidth memory (HBM). The video names Samsung, SK Hynix, and Micron as major suppliers.
   • It says AI data centers are absorbing roughly 70% of global high-end memory. HBM production uses about three times as many silicon wafers per gigabyte as conventional memory.
   • Memory prices were said to be more than 10 times their January 2025 level in some cases. Another 10–15% price increase was anticipated for the quarter.
   • HP reportedly said memory had risen to about 35% of a laptop’s materials cost, versus roughly 15% three months earlier. Higher device prices can delay PC and phone purchases, hurting manufacturers and chip suppliers.
   • A major memory producer’s capacity was described as sold out through 2029, with shortages not expected to ease before 2030.

Investment framework and risk

The video frames the sell-off as a redistribution of value: end users and AI infrastructure providers may benefit, while businesses whose existing revenues are displaced—or whose input costs rise—may lose.

The key risk is financing. If AI weakens the traditional businesses that currently generate cash for data-center investment, tech companies may borrow more or reduce planned spending. That could ultimately hurt suppliers such as Nvidia, Micron, and Sandisk.

The presenter’s conclusion is that the current risk looks more like “dying of success” than “dying of failure”: AI demand appears strong, but the sector’s investment plans may not be self-financing if legacy revenue declines faster than new AI income grows.

No valuation multiples or specific stock-price targets are given.

ETF mention and disclosure

The video includes a promotional segment for the iShares AI Infrastructure UCITS ETF, described as investing in companies involved in semiconductors, cloud computing, and big-data technologies. The presenter says it had gained about 50% year to date at the time of recording and explains how to buy it through Freedom24.

This is a promotional mention, not an individualized portfolio recommendation. The video does not state a separate “not financial advice” disclaimer.

Presenter/source: Juan Ramón Rallo, on the Juan Ramón Rallo channel.