GAFAM NewsMonthly roundup
Big Tech in August 2025: the light and the dark
August 2025: Apple pledged $100 billion more for US manufacturing and Amazon licensed NYT content for AI, while smishing surged and Meta's AI rules drew fire.
- First published in Japanese
- English edition
In August 2025, amid the ongoing chaos around AI, the direction of travel started to become a little clearer.
Put simply, AI as a tool of national strategy is being pushed ever further to the front. Meanwhile, the points of contention in regulation and in the courts are getting clearer, but ethical questions tend to be put off until later. Through it all, GAFAM still plays the central role.
This month, too, let’s look at the latest GAFAM moves from both sides of AI: the hopes and the problems.
The light side
Apple invests $100 billion in US production, speeding up “make it where you sell it”
Apple announced an additional $100 billion (about ¥14.7 trillion) investment to expand manufacturing in the United States.
The plan was unveiled in the presence of US President Trump, and it includes large-scale measures to strengthen production, such as building AI server factories in the US and hiring 20,000 new people for research and development. [1]
Behind it is the US government’s pressure of high tariffs on the iPhone; by shifting production home, Apple aims to avoid them. This bigger investment encourages local-for-local supply chains and helps revive US manufacturing. It may also push parts suppliers, Japanese companies included, to rethink where they produce, and it could open up new business opportunities for them.
At the same time, behind this move you can see the US–China split over AI accelerating, and Japanese companies that depend on Chinese demand are being pressed to choose their path forward.
Meta strengthens AR with “Hypernova,” its next-generation smart glasses
Reports that Meta will release “Hypernova,” a new pair of smart glasses with a built-in display, drew attention in the AR field.
The price is around $800, and the glasses are expected to be announced at Meta Connect in September 2025. Mass production is said to start in Q3, and sales are expected to be kept to around 150,000 to 200,000 units over the next two years, so Meta seems to position them as an experimental product.
On top of an AI-powered vision interface, the combination of fashionable frames through its partnership with EssilorLuxottica and a wristband for controlling the glasses looks like a strategy aimed at making AR wearables a mass-market product. [2]
Even while it pours effort into its AI business, Meta seems to be pushing its XR strategy forward step by step.
Amazon partners with The New York Times, paying for the value of content
Amazon signed a deal to pay The New York Times about ¥3.7 billion (roughly $25 million) a year. This is the Times’ first formal AI-related license, and it is also Amazon’s first AI deal with a major news organization.
With AI development requiring enormous amounts of text, partnerships like this are drawing attention as a new revenue source for newspapers and other traditional media, and as a mechanism that supports high-quality journalism. [3]
However, Japan has all but given up on developing its own large language models. If frameworks like this spread, it could mean Japanese media ending up under the influence of American players such as GAFAM and OpenAI.
I’ll be watching what Japan’s major media outlets do next.
Google’s AI search is coming to Japan
Google announced that it plans to bring new search features powered by generative AI to Japan within the year.
With the “AI Overviews” feature, AI automatically generates a summary of the search results, so users can quickly grasp the key points when they ask a question. [4]
Trials in the US and elsewhere have reportedly confirmed that visits to websites do not decline, and Google is making improvements that take into account the impact on the people who provide the information. [5]
If it fully launches in Japan this year, it should make search more convenient and encourage access to a wider range of information sources.
The dark side
SMS scams (“smishing”) are surging, and Japan is a prime target
Recently, scams that abuse SMS (text messages), known as “smishing,” have been surging.
Because texts reach so many people, it is easy to manufacture a sense of urgency and trust, which makes this a favorite method of attack. Recently there were also reports that more than 115 million credit card records in the US were leaked through smishing attacks, which shows how serious the damage can be. [6]
As these methods spread, the GAFAM companies are being targeted by scammers as brands to impersonate. A Check Point study reported that Microsoft is the most impersonated brand in phishing (38%), followed by Google (11%) and others. This suggests that smishing is expanding as part of a wider pattern of attacks that exploit brand trust. [7]
Especially striking is smishing that poses as Amazon. With messages like “There is a problem with your account” or “You need to complete a refund,” the scammers lure people to fake websites, and there are reports that this summer it rose by roughly 5,000% (50 times) or more compared with the previous year. [8]
At the same time, advances in AI have made it easy to write scam messages in natural Japanese, lowering the barrier for SMS scams aimed at Japan from overseas.
According to Proofpoint, in May 2025, messages targeting Japanese speakers reached 81% of all spam emails sent worldwide, suggesting a growing trend of scams written specifically in Japanese. [9]
Big sales and events from the fall onward are likely to bring even more of this scam activity, so consumers need to be more careful than ever.
The ethical collapse of Meta’s AI chatbots: “inappropriate conversations” with minors
An internal Meta document leaked, revealing that the rules for the company’s AI chatbots had allowed romantic or sensual conversations with minors.
The guideline document, more than 200 pages long, included specific example prompts. According to reports, it stated that “It is acceptable to engage a child in conversations that are romantic or sensual,” while drawing a line that prohibited role-play describing sexual acts.
This caused a big stir in the United States. Out of concern for children’s safety online, senators launched an investigation. The Kids Online Safety Act (KOSA), which failed to pass once in the face of heavy lobbying by the tech industry, is moving toward becoming law again, and the debate over AI ethics and child protection has suddenly sped up. [10] [11]
In Japan too, rules and guidelines on how AI treats underage users will likely be needed.
xAI accuses Apple of favoring ChatGPT and threatens to sue
xAI, the AI startup led by Elon Musk, announced that it intends to sue Apple and OpenAI for blocking competition in the conversational AI market.
On his social media account on X, Musk claimed that “Apple is behaving in a manner that makes it impossible for any AI company besides OpenAI to reach #1 in the App Store, which is an unequivocal antitrust violation,” and signaled that he would take legal action.
At the moment, ChatGPT is No. 1 in the US App Store’s free app ranking, while xAI’s Grok sits at No. 5. Combined with the fact that Apple has partnered with OpenAI and integrated ChatGPT into its own operating systems, this has raised suspicions that Apple favors one competitor.
No evidence has been presented, but the move has turned into a debate drawing in Sam Altman (CEO of OpenAI) and other industry leaders, and some point out it could become a courtroom battle over fairness in competition between AI platforms. [12]
The AI model race may end up “winner takes all,” just like the search market that Google once monopolized. Perhaps the people running at the very front of AI development are conscious of, and worried about, exactly that kind of future.
Microsoft fires protesting employees, sparking controversy as corporate ethics clash with workers’ rights
Protests against management’s policies grew inside Microsoft, bringing the conflict between employees and executives into the open.
The company has faced criticism from its own employees over its relationship with the Israeli government (including supplying technology used in the Gaza conflict). In August, things escalated to the point where employees who joined a protest outside headquarters were arrested by police. Then, at the end of the month, it was reported that Microsoft had fired four employees involved in the protests. Although the employees had voiced concern that the company’s technology might be complicit in the conflict, the company responded with the hard-line step of dismissing them. [13]
The incident has sparked debate about how to balance tech companies’ social responsibility with the voices of their employees. From here on, not only Microsoft but every company will be asked how it respects internal ethics and workers’ rights.
GAFAM keeps swinging back and forth between hope and concern, and I’m sure it will make big headlines again from September onward. As an owl, I may tire myself out swiveling my head to see which way they fly next.
References
[2] Meta reportedly has low sales expectations around Project Hypernova smart glasses (Android Central)
[3] Amazon to pay The New York Times ¥3.7 billion a year for AI use (Nikkei) (in Japanese)
[4] Google Search “AI Mode” to launch in Japan as early as this year (Nikkei) (in Japanese)
[5] Google executive says site visits do not fall even with AI search (Nikkei) (in Japanese)
[7] Phishing Statistics: The 21 Latest Phishing Stats to Know in 2024 (Hashed Out)
[8] 250+ Phishing Statistics and Trends You Must Know in 2025 (Keepnet)
[12] Musk’s xAI plans legal action against Apple over app rankings (Reuters) (in Japanese)
[13] US tech tightens grip on protests: Microsoft fires anti-Israel employees (Nikkei) (in Japanese)