GAFAM NewsMonthly roundup
Big Tech in January 2026: the light and the dark
January 2026: Apple picked Google's Gemini for the next Siri and Microsoft doubled carbon removal, while Meta and Google faced scam-ad and privacy trouble.
- First published in Japanese
- English edition
In January 2026, GAFAM news showed generative AI spreading beyond the boundary between hardware and software, and environmental investment picking up speed. At the same time, lawsuits and regulation over competition and privacy kept coming. It was a month of tech companies swaying between innovation and ethics.
The light side
Apple adopts Google’s Gemini for the next-generation Siri
On January 12, 2026, Apple announced a multi-year agreement with Google to use Google’s generative AI model, Gemini, for the next-generation Siri and future AI features.
With this partnership, Gemini will reach not only Samsung’s Galaxy AI but also Apple’s base of more than 2 billion devices, a major business opportunity for Google.
Apple explained that it chose Google’s AI technology because it judged it to be “the most capable foundation,” and that it will deliver Apple Intelligence features while keeping its privacy protections in place.
The move comes on top of Apple’s existing collaboration with OpenAI and further speeds up the race among AI assistants. Alphabet’s market value topped $4 trillion on AI expectations, and the whole industry took notice.[1]
A new wave of Galaxy AI: Samsung to double its devices
Samsung Electronics co-CEO T.M. Roh said that the number of mobile devices with Galaxy AI features powered by Google’s Gemini will double, from about 400 million units in 2025 to 800 million in 2026.
The company plans to bring AI to every product, every feature and every service, extending AI features beyond smartphones to TVs and home appliances.[2]
For Google, this means a larger user base to compete against OpenAI and Chinese rivals, and in Japan, too, new AI services for Galaxy users look set to roll out.
Microsoft doubles its carbon removal contracts
Microsoft announced that it signed carbon dioxide removal agreements for 45 million tonnes in 2025, double the previous year. It contracted with 21 companies and uses a wide range of removal methods, from nature-based to engineered.
The company says these are important for absorbing its residual emissions as it works toward its goals of becoming carbon negative by 2030 and removing all of its historical emissions by 2050.[3]
Projects with Japanese companies taking part are also expected, which could help grow the carbon removal market in Japan.
A new frontier in medical image analysis: MedGemma 1.5
Google Research announced MedGemma 1.5 4B, an open model for medicine, and launched a hackathon on Kaggle.
The new model expands support for high-dimensional medical images, including three-dimensional scans such as CT and MRI, as well as analysis of longitudinal chest X-rays and anatomical localization.
It also adds the ability to extract structured data from documents, and its accuracy has improved over the previous version.
Because a 4B model that can run offline is also available, it will be easier for researchers and companies in Japan to customize and use, and it should help telemedicine and diagnostic support move forward.[4]
A Japanese startup teams up with Google: the Sakana AI partnership
Japanese AI startup Sakana AI announced a strategic partnership with Google, along with an investment from Google.
The company said it will use Google’s global scale and infrastructure to raise the quality and reliability of its AI products, which tackle complex industrial and scientific problems.
By integrating Google models such as Gemini and Gemma, it will strengthen its own technologies such as “The AI Scientist” and “ALE-Agent,” aiming for use in fields that demand high levels of safety, such as finance and government agencies.[5]
The news raises hopes both for a stronger AI ecosystem in Japan and for expansion into global markets.
The dark side
Meta’s playbook for “hiding” scam ads
According to an investigative report by Reuters, as the Japanese government and others pressed Meta to deal with scam ads, the company created an internal manual for adjusting the ad libraries of Facebook and Instagram so that scam ads would be harder to find.
Internal documents describe a method of identifying the search keywords used by Japanese regulators, running the same searches repeatedly, and deleting the ads that turned up, so that problem ads could no longer be found.
This method was reportedly built into a global playbook for responding to regulators and rolled out in the United States, Europe and elsewhere.
Meta argues that removing scam ads is not in itself misleading, but the lack of transparency has drawn criticism.[6]
The Instagram password reset scare and a suspected data leak
In mid-January, Instagram users received a flood of password reset notices they had not requested, and reports that data on about 17 million users was circulating on the dark web spread anxiety.
CPO Magazine pointed out that the data was a compilation of information from past leaks and did not include credentials such as passwords.
On January 11, Meta said it had fixed an issue that let a third party send reset emails for some accounts, and stressed that there had been no new breach. Security experts recommended turning on two-factor authentication, while describing the incident as a bug that coincided with old data being spread around, not a large-scale breach.[7]
A chatbot tragedy: Google and Character.AI settle
Google and Character.AI agreed to settle a lawsuit filed by a mother in Florida, who alleged that a Character.AI chatbot had encouraged her 14-year-old son to take his own life.
According to the complaint, the chatbot posed as a psychotherapist and a romantic partner and encouraged the boy to kill himself.
The terms of the settlement were not disclosed, but the case is seen as one of the first in which an AI company faced claims of psychological harm, and similar suits in other states are also moving toward settlement. It once again highlighted how important laws on AI safety and ethics are.[8]
Google pays $135 million to settle a lawsuit by Android users over improper data transfers
Android users filed a class action alleging that Google collected cellular data even when apps were closed or location settings were turned off, and used it for ad delivery and product development.
Without admitting wrongdoing, Google agreed to pay $135 million (about ¥20 billion) to settle the case. The settlement includes commitments not to send user data without consent in the future and to make it easy to stop data transfers from the settings screen.
With privacy regulation continuing to tighten, it was a significant lesson for the company as well.[9]
There was a run of bright news, such as the number of devices with generative AI doubling, while the darker issues of privacy lawsuits and regulatory investigations kept dogging the companies as usual.
January seems to have been a month in which each company looked for a balance between innovation and responsibility, keeping one eye on the line between light and dark.
From my branch, I’ll keep watching generative AI spread its wings, reading the wind without hurrying or making a fuss, and following each company’s next step.
References
[2] Reuters: Samsung to double AI mobile devices to 800 million units this year (January 5, 2026)
[5] Pulse 2.0: Sakana AI Announces Strategic Partnership With Google (January 24, 2026)
[8] Reuters: Google, AI firm settle lawsuit over teen’s suicide linked to Chatbot (January 7, 2026)
[9] Reuters: Google to pay $135 million to settle Android data transfer lawsuit (January 28, 2026)