Apple iPhone teams up with Google AI to shake up the market! Musk criticizes unfair monopoly

Apple and Google have reached a partnership, with the new version of Siri to be rebuilt based on Google AI. Alphabet’s market capitalization has surpassed 4 trillion USD. Elon Musk criticized: “Google already controls Android and Chrome, which is an unreasonable concentration of power.” Google AI has 1.2 trillion parameters, far exceeding Apple’s 150 billion.

Why Elon Musk Criticizes the Apple-Google Partnership

Apple and Google issued a joint statement announcing a deep, multi-year cooperation agreement. The next-generation Apple foundation model will be built directly on Google’s AI models and cloud technology. This means that the core brain driving Apple Intelligence in the future will have Google’s influence flowing through it. The Siri that was once criticized as “artificial stupidity” will soon undergo a thorough rebuild based on AI models.

Following this news, both Apple and Google stocks rose, and Alphabet’s market cap directly broke through the 4 trillion USD mark. It seems the capital markets have approved this “facing reality” collaboration. But Elon Musk was the first to react, directly criticizing on social media: “Considering Google already owns Android and Chrome, this seems like an unreasonable concentration of power.”

Musk’s criticism hits the mark. Google currently controls the world’s largest mobile operating system Android (about 70% market share), the largest browser Chrome (about 65%), and the largest search engine (over 90% market share). Now, it will also control the AI core of iPhone. Such vertical integration of power is extremely rare in tech history and is almost equivalent to monopoly.

Three Major Risks of Power Concentration in Apple-Google Collaboration

Mobile Ecosystem Monopoly: Android + iOS dual systems’ AI layers are both provided by Google, controlling 95% of global smartphones

Data Channel Control: AI interactions may collect deeper user behavior data

Choking Competitors: Startups like OpenAI, Anthropic lose access to the largest distribution channel—iPhone

As the founder of xAI, Musk’s criticism also carries commercial interests. The Grok model from xAI could have been a partner for Apple, but ultimately lost to Google. This disappointment may have amplified Musk’s concerns about “power concentration.” However, setting aside personal interests, his criticism indeed points to a real antitrust issue.

Google AI 1.2 Trillion Parameters Outperform Apple’s 150 Billion in the Technology Gap

As early as August last year, Bloomberg reported that Apple was in early negotiations with Google, planning to use customized AI models to support the new Siri. In November, the media revealed that Apple planned to pay about 1 billion USD annually to use Google’s AI technology. According to the joint agreement, Google provides Apple with an AI model containing 1.2 trillion parameters, far surpassing Apple’s current 150 billion parameters model.

This huge parameter scale difference directly reflects in the model’s reasoning ability, knowledge breadth, and task complexity. For reference, in July last year, the open-source Kimi-K2-Instruct model was released, with a total of 10 trillion parameters, becoming the first domestic open-source foundation model to break the trillion-parameter barrier. Apple’s 150 billion parameters are clearly lagging in this scale of competition.

At WWDC 2024 in June, Apple showcased a new Siri powered by Apple Intelligence, including stronger context understanding, screen perception, cross-application operations, and other new features. According to Apple, the new Siri features were initially scheduled to be gradually released within the iOS 18 update cycle, but many key AI features were delayed early on, with the earliest unified release now expected in spring this year. This prolonged delay exposes Apple’s shortcomings in large model technology.

Even more seriously, Apple’s AI team is facing severe talent loss. Since July last year, dozens of core members have left. Ruo-Ming Pang, head of Apple’s foundational model team, was poached by Meta for 200 million USD; Ke Yang, responsible for Siri’s intelligent search project, recently left to join Meta; several key researchers who published AI papers last year have also left for OpenAI, Cohere, and xAI. This small team of over 100 people lost key talent at a critical moment.

Apple’s secretive culture was once its moat, but in the AI era, this approach has failed. When researchers cannot freely publish papers, cannot build reputation in academia, and lack open-source exchanges, they miss out on the rapid iteration of the AI community. More critically, Apple started late in computing resources, and training data is relatively scarce due to privacy policies. As a result, Apple has no choice but to seek external help.

OpenAI as the Biggest Loser with a Disparity in Default Status

After the collaboration between Apple and Google materialized, the most awkward party is actually OpenAI. Although Apple told CNBC that there are currently no plans to change agreements with OpenAI, the disparity in status is obvious. Equisights Research analyst pointed out sharply: “Google’s AI will become the ‘default intelligence layer’ on Apple devices, while OpenAI is forced into a ‘supporting role.’”

This status difference is commercially significant. Default options enjoy huge traffic advantages; most users will not proactively change default settings. When Siri’s core reasoning is powered by AI, most user interactions will be captured by Google. Although OpenAI retains some integration, its traffic share will be greatly reduced.

According to the latest data released by SimilarWeb in January, global web traffic share of AI has exceeded 20% for the first time, and ChatGPT’s market share plummeted from 86% in January 2025 to 64.5%. The rise and fall of these figures demonstrate user support for AI. Apple’s choices further accelerate this trend, potentially becoming a turning point in reshaping the AI market landscape.

In November last year, Google launched its most powerful AI series model to date, with multiple core metrics surpassing ChatGPT, which also prompted OpenAI CEO Sam Altman to urgently sound the “red alert,” personally ordering improvements to ChatGPT. Last year, Google AI team member Logan Kilpatrick revealed on social media that Google processes 1.3 trillion tokens per month, setting a record for industry-wide computational power consumption.

Integrating Google AI significantly improves the likelihood of Apple delivering on time. If Apple insists on purely self-developed solutions, considering talent loss and technical gaps, it remains uncertain whether they can meet the spring deadline. However, by adopting Google’s well-developed models, Apple gains a ready-made “shortcut.” According to previous leaks, Gemini is expected to handle Siri’s summarizer and task planning functions—core capabilities for information integration and executing complex tasks.

Of course, the Chinese domestic version of Siri AI is unlikely to use Google AI. Apple must prepare different AI solutions for different markets, such as collaborating with domestic vendors or using specially tailored self-developed models. In the AI era, even Apple cannot remain an isolated island. Letting go of pride and returning to pragmatism may be the most rational survival instinct for tech giants in the AI age.