The company wants to scale first, refine later, and lock users into an AI-driven system before competitors can meet up.

By the end of last year, around 400 million Samsung devices already carried Gemini-powered features. That figure will double as the company extends AI beyond smartphones to tablets, televisions and home appliances. 

Beyond a feature upgrade, Samsung is enhancing how its products work and how users interact with them.

“We will apply AI to all products, all functions, and all services as quickly as possible,” T M Roh said in his first interview since becoming co-CEO.

As the largest supporter of Google’s Android platform, Samsung is also giving Google a massive consumer advantage at a time when AI models are fighting for everyday relevance. 

Every Samsung Galaxy phone shipped with Gemini baked in is another front opened in Google’s move against OpenAI and others.

Global Smartphone Shipments to Fall 2.1% in 2026 as High Memory Prices Hit Low-End Devices Hard

The strategy is already visible in Samsung’s flagship devices. The Galaxy S25 series, launched in early 2025, arrived with solid Gemini integration, including side-button access that replaces Bixby in some tasks. 

Samsung wants AI to feel native, not optional. The aim is to make Gemini a default layer across the Galaxy ecosystem, not just a chatbot buried in an app.

This comes as competition in AI is growing. Google’s Gemini 3, released in November 2025, set new performance records, becoming the first model to cross the 1500 Elo threshold on reasoning benchmarks. 

It also led in maths, coding, multimodal tasks and long-context understanding. The response was quick. OpenAI launched GPT-5.2 weeks later, following reports that Sam Altman had declared an internal “code red” to enhance development.

Samsung believes consumer adoption is meeting up with the technology. Roh said internal surveys show awareness of its Galaxy AI brand has jumped from about 30% to 80% in just one year.

“Even though the AI technology might seem a bit doubtful right now, within six months to a year, these technologies will become more widespread,” he said.

On phones, search is the most used AI feature. But usage is spreading to image editing, productivity tools, translation and summaries. These are small actions, repeated daily, and that is where Samsung thinks loyalty will be built.

Still, the aggressive AI rollout is happening against a tougher market backdrop. A global shortage of memory chips is lifting prices for key components, helping Samsung’s semiconductor unit but squeezing margins in its smartphone business.

“As this situation is unprecedented, no company is immune to its impact,” Roh said.

He acknowledged that price increases may be unavoidable, calling some impact “inevitable”, even as Samsung works with partners to soften the blow over the longer term. 

Counterpoint Research revised its 2026 smartphone shipment forecast in December, predicting a 2.1% decline as higher DRAM prices increase device costs. Analysts estimate memory price hikes have added between 10% and 25% to the bill of materials across devices.

Samsung is also managing expectations around foldable phones, a category it pioneered in 2019. Growth has been slower than hoped, held back by engineering challenges and a lack of apps designed for foldable screens. 

Roh believes the format will break through within two to three years, noting that repeat purchase rates among foldable users are “very high”.

For now, Samsung is firmly in control. It held about 64% of the global foldable smartphone market in the third quarter of 2025, far ahead of Huawei and Motorola.

Apple is expected to launch its first foldable iPhone in 2026, increasing the stakes in a segment Samsung once had to itself.

For the second quarter, the chipmaker reported $46.7 billion in revenue, representing a 56% jump from the same period last year. Net income also surged to $26.4 billion, up 59% year-on-year. 

The results, which is more than the initial projection, reveal just how much the company has become the backbone of the artificial intelligence growth.

Data centre sales, the largest driver of growth, pulled in $41.1 billion, with the new Blackwell generation of chips accounting for $27 billion alone. “Blackwell is the AI platform the world has been waiting for,” CEO Jensen Huang said. “The AI race is on, and Blackwell is the platform at its centre.”

Huang spoke on his long-term outlook, projecting a $3 to $4 trillion wave of AI infrastructure spending by 2030. “$3 to 4 trillion is fairly sensible for the next five years,” he told analysts on the call.

Beyond the financials, Nvidia underlined its central role in OpenAI’s release of open-source gpt-oss models earlier this month, noting that its Blackwell GB200 NVL72 rack system processed 1.5 million tokens per second during the launch.

However, not all markets are open doors, as the company admitted that it sold zero units of its China-focused H20 chip in the past quarter. Instead, $650 million worth of those devices went to a buyer outside China. 

The lack of shipments comes from a murky U.S. policy under President Trump, which currently allows Nvidia to sell advanced GPUs to China if it pays a 15% export tax. Legal scholars have criticised the arrangement as an “unconstitutional abuse of power.”

Nvidia’s Chief Financial Officer, Colette Kress, stressed the company’s hesitation: “While a select number of our China-based customers have received licences over the past few weeks, we have not shipped any H20 devices based on those licences.”

Adding to the issue, Beijing has discouraged local firms from using Nvidia’s chips, a move that reportedly pushed the company to halt production of the H20 earlier this month.

In future plans, Nvidia has guided for $54 billion in revenue in the third quarter, though it warned that figure does not include any Chinese shipments. The guidance was broadly in line with Wall Street expectations, but fell short of more bullish analyst forecasts of over $60 billion.

Despite the blockbuster earnings and the announcement of a $60 billion stock buyback, Nvidia’s shares slid about 3% in post-market trading.

The dip reflected disappointment over a narrow miss in data centre revenue, a segment that investors have been watching closely as a proxy for the strength of the AI boom.