Big Tech in October 2025: the light and the dark

In October 2025, tech stocks rose across the board on the back of AI expectations, and the growth story’s second act came into sharp focus.

At the same time, Amazon began a large round of layoffs, and, following Microsoft, “workforce optimization for the AI era” moved out into the open. Meanwhile, Meta’s huge AI investments squeezed its profits and sent its stock tumbling, and investors are turning back to a sober question: does AI really make money?

With expectations and reality crossing paths, I’d say this was the month the “payback phase” of AI strategy finally began.

Now, as every month, join me, the Owl, for a look back at the major GAFAM news.

---

The light side

Meta makes a bold shift toward AI infrastructure and talent

Meta is accelerating large-scale investment and organizational change to strengthen its position in the AI race.

In Louisiana, it is moving ahead with plans to build Hyperion, a next-generation AI data center, and has raised $27 billion (about ¥4 trillion) in private capital for it. The company announced that the site will have 2GW-class computing power and will be used to train large language models such as ChatGPT and Google’s Gemini. [1]

Meta also recently cut about 600 jobs in its AI division, while moving people from its metaverse unit into AI and poaching an AI executive from Apple, reallocating talent to concentrate on its priorities. Even as it speeds up AI investment, it is rushing to rebuild its organization. [2] [3]

In its third-quarter 2025 results on October 29, revenue reached $51.24 billion, up 26% year over year. But costs driven up by AI infrastructure investment and other items left earnings per share (EPS) below market expectations. Meta raised its 2025 capital expenditure outlook to $70–72 billion and said plainly that it expects spending growth to be “notably larger” in 2026. [4]

The stock fell as much as 9% at one point, but Meta has not backed away from its bold stance of investing for growth. [5]

Apple boosts performance with new M5 devices

Apple announced three products, the iPad Pro, Vision Pro and 14-inch MacBook Pro, as the first to carry M5, its new in-house SoC.

The designs are unchanged, but processing speed, graphics performance and AI processing have improved substantially. The new iPad Pro’s AI performance, for example, is up to 3.5 times that of the previous generation.

Prices are unchanged, and sales began on October 22. The extra performance should make creative work and everyday AI use even smoother, and in Japan it looks likely to spur upgrade demand ahead of the year-end shopping season. [6]

Google’s researchers win Nobel Prizes two years running as quantum moves toward practical use

Google has produced Nobel Prize–linked achievements in quantum computing research for the second year in a row, a major step toward making a next-generation computing platform practical.

In 2025, a scientist from a Google-backed research project won the Nobel Prize, once again confirming Google’s position as a global leader in quantum science. It has won high praise for speeding up the bridge from basic research in quantum physics to industrial application. [7]

In October, Google also announced its quantum processor Willow and Quantum Echoes, a quantum algorithm said to run up to 13,000 times faster than conventional supercomputers. The findings were published in Nature and drew attention for showing a path past hard problems such as quantum error correction and the stable operation of large numbers of qubits.

This puts commercial use of quantum computing within view across a wide range of fields: drug discovery, materials science, climate simulation, and even making AI model training more efficient.

Google has set a goal of practical use within the next few years and is positioning quantum technology, alongside its AI strategy, as its next growth engine. [8][9]

In Japan, too, collaboration with universities and research institutions is moving forward, which could help train quantum talent and create new industries. Google is steadily strengthening its presence at the center of the “next computing revolution” after AI.

Windows evolves into an “AI-native OS”

Microsoft announced a plan to turn Windows 11 into a platform where you work together with an AI assistant on every PC.

With the voice command “Hey, Copilot,” you can talk to AI built into the whole OS, and the company is also strengthening Copilot Vision, a feature in which the AI recognizes what is on screen and helps you operate it. A new experience of running your PC by voice and AI, on top of keyboard and mouse, should improve productivity and accessibility. [10]

For Japanese companies, this could be the push that speeds up the move to Windows 11 and for bringing generative AI into everyday work.

---

The dark side

Amazon’s largest-ever job cuts

Amazon appears to have begun large-scale layoffs affecting about 14,000 people.

That is equivalent to 4% of its corporate workforce, spanning several divisions including devices, advertising, video streaming and AWS. The stated goals are to correct pandemic-era overhiring and cut costs, but the move also reflects an aim to cut office jobs by automating work with generative AI. CEO Andy Jassy has suggested that AI will take over routine work and drive further streamlining. [11]

In Japan, too, there are concerns about the impact on Amazon group companies and on distribution, and debate is growing over how technological progress affects employment.

The global AWS outage exposes the risks of cloud dependence

On October 20, AWS, Amazon’s cloud service, suffered a major outage lasting about 15 hours, taking down hundreds of web services around the world.

A wide range of services were affected, from payment apps and stock trading apps to Apple’s music streaming, Zoom, Salesforce, McDonald’s ordering system and even the game Fortnite. Amazon’s own Alexa and Ring stopped working as well.

The outage laid bare how heavily the world depends on a single company (AWS), and experts pointed out that it was the worst since the 2021 outage. [12]

In Japan, too, corporate websites and apps ran into problems one after another during the same hours, raising concerns about how concentrated the cloud has become. Japanese companies will likely be pushed to strengthen their own preparations too, through multi-cloud setups and backup systems.

The irony of “AI rescuing Google” in its antitrust case

In the United States, the first major antitrust case in decades against Google, which dominates the search market, moved forward.

The court first found that Google held an illegal monopoly, but in the October ruling on remedies, Judge Mehta said that “the emergence of GenAI changed the course of this case” and held back from imposing harsh corrective measures on Google.

On the view that emerging AI chatbots such as ChatGPT could become competitors to search engines, no forced breakup or opening of key assets was ordered, and Google effectively came out the winner. [13]

It is an ironic outcome in which technological progress shapes the direction of regulation. In Japan, where debate continues over the new law to promote competition in digital markets (including rules on app stores), policymaking needs to take into account the impact of rapid AI innovation.

The EU moves against Meta over illegal content

The European Union (EU) took its first enforcement step against Meta (Facebook and Instagram), saying the company has not complied with new rules on tackling illegal content.

An investigation under the EU Digital Services Act (DSA) found that the system for users to report illegal content such as child sexual abuse material and terrorism is hard to understand, and that it includes unnecessary steps known as “dark patterns.” It also said removal after reports was insufficient and that the means of appealing account suspensions were ineffective.

Meta counters that it complies with the law, but if it does not improve, it could face fines of up to 6% of its global revenue. [14]

This case tests the balance between freedom of expression and safety, and in Japan, too, pressure may grow on platform companies to provide transparency and protect users.

---

This month, the expectations and the realities around AI came into sharper focus for GAFAM. Stock prices rose, while layoffs and the burden of investment kept growing, and the companies entered a phase where growth has to become leaner and more muscular. Google’s quantum research was a strong sign of what’s ahead, and the next chapter of AI has quietly begun.

Speaking of muscle, we owls are more muscular than we look: flight muscles are said to make up about 30–40% of our bodies. Quiet on the outside, but well trained underneath… perhaps that is exactly the stage tech companies are at right now.

---

References

[1] Meta raises $27 billion for its data center project in Louisiana (Reuters) (in Japanese)

[2] Meta names its metaverse chief to lead AI products after job cuts (Bloomberg) (in Japanese)

[3] Apple’s AI search executive to move to Meta as talent drain continues (Bloomberg) (in Japanese)

[4] Meta forecasts bigger capital costs next year as Zuckerberg lays out aggressive AI buildout

[5] Meta earnings recap: Stock tumbles 9% after big tax charge and plans to spend even more on AI

[6] Apple announces three products with its new M5 chip, boosting performance ahead of the holiday season (Bloomberg) (in Japanese)

[7] Google researchers win the Nobel Prize for the second year in a row for basic research in AI and quantum (Nikkei) (in Japanese)

[8] Google unveils a new quantum computing method 13,000 times faster than a supercomputer (Nikkei) (in Japanese)

[9] Google makes a major advance in quantum computing, with hopes of practical use within five years (Bloomberg) (in Japanese)

[10] Microsoft’s aim in “making every Windows 11 PC an AI PC”: pushing the move from 10 to 11 through “working together with AI” (Business Insider Japan) (in Japanese)

[11] Amazon to cut about 14,000 corporate jobs in AI push

[12] Amazon’s cloud outage is resolved after affecting hundreds of sites, exposing dependence on one company (Bloomberg) (in Japanese)

[13] How AI Upended a Historic Antitrust Case Against Google | TechPolicy.Press

[14] Meta found in breach of EU law over ‘ineffective’ complaints system for flagging illegal content