Mark Gurman - Apple Plans to Use Google's Gemini Model to Power New Siri
When a tech giant known for its walled garden pays a rival a billion dollars to power its flagship AI, is it a pragmatic shortcut or a sign of deeper trouble?
Apple Plans to Use 1.2 Trillion Parameter Google Gemini Model to Power New Siri - Bloomberg
When Siri Starts Speaking Google's Language
According to the latest report from Bloomberg's renowned journalist Mark Gurman, the tech landscape appears to be undergoing a subtle yet profound restructuring. Apple, the giant known for its closed ecosystem and vertical integration, is finalizing a deal with its arch-rival, Alphabet. The core of the agreement: Apple plans to pay approximately $1 billion annually to use Google's staggering 1.2 trillion-parameter Gemini AI model to power its next-generation Siri.
The heart of this deal points to an open secret in the AI field: Apple has fallen behind. As the industry was swept by the wave of Large Language Models (LLMs), Apple's response—Apple Intelligence—while well-integrated on-device, saw its 15 billion-parameter cloud model pale in comparison to Google's trillion-parameter behemoth. This collaboration is, without a doubt, Apple's most candid admission of this technological gap. It's not just a technology procurement; it's a significant strategic compromise, designed to inject the much-needed modern AI capabilities into its smart assistant, Siri, to fend off intensifying market competition.
A Meticulously Designed 'Behind-the-Scenes' Partnership
Despite this reliance on a competitor, Apple is striving to ensure the collaboration remains distinctly Apple. According to the report, Google's Gemini model will run on Apple's own Private Cloud Compute servers. This is a crucial detail, meaning user data will remain isolated from Google's infrastructure, allowing Apple to continue upholding its celebrated privacy moat.
On the technical integration front, Gemini will primarily handle Siri's summarizer and planner functions—the core brain for understanding complex commands and planning execution paths. Other Siri features will continue to be powered by Apple's in-house models. This hybrid approach showcases Apple's strategy: selectively adopting top-tier external technology to fill its own gaps without completely surrendering autonomy. Interestingly, Apple apparently does not plan to publicly promote this partnership, instead positioning Google as a hidden, backstage technology provider. This is a stark contrast to the collaboration that made Google the default search engine for Safari, revealing Apple's cautious and conflicted mindset in the core domain of AI.
A Stopgap Measure or Long-Term Dependence?
So, does this mean Apple will rely on Google's AI technology for the long term? The answer appears to be no. Apple views this collaboration as a "transitional solution," a stopgap measure to buy time while its proprietary technology matures.
Reportedly, Apple's internal AI model team is in a full-throttle effort to develop its own 1 trillion-parameter cloud model, aiming for a consumer product launch as early as next year. However, catching up is no easy feat. Google's Gemini, particularly its 2.5 Pro version, consistently ranks at the top of most LLM leaderboards and is continuously iterating. Apple must not only catch up to today's Gemini but tomorrow's as well—an undeniably arduous race.
The Unique Game in the Chinese Market
The complexity of this partnership is vividly illustrated in the Chinese market. Due to Google's absence in China, a Siri overhaul reliant on Gemini is not viable there. Apple must find a localized solution.
To this end, Apple Intelligence in the Chinese market will use Apple's internal models combined with a filtering layer developed by Alibaba to ensure content complies with local regulatory requirements. Meanwhile, Apple is also in negotiations with Baidu, considering its AI products as a potential alternative. This once again proves that in the global tech competition, the Chinese market will always be a separate battlefield requiring a unique strategy.
Returning to the initial question: for Apple, is this deal a pragmatic shortcut or a sign of deeper trouble? Perhaps it's both. It is an exceptionally pragmatic tactical choice, allowing Apple to quickly enhance its product competitiveness and avoid being marginalized in the AI era. At the same time, it clearly exposes the anxiety and struggle of a company that once defined the mobile internet era as it confronts a new technological revolution. Apple has bought precious time with its $1 billion, but the real challenge lies in whether it can use this time to reclaim its leadership in technological innovation. That is the key that will determine its fate for the next decade.
