"Shelling" Google Gemini, But Apple Hasn't Given Up on Its Own Model

marsbitОпубликовано 2026-01-13Обновлено 2026-01-13

Введение

Apple has reportedly entered into a significant partnership with Google to integrate its Gemini AI model into Apple's ecosystem, as a "foundation" component for its future AI features, including a more personalized Siri expected later this year. The deal, estimated at around $1 billion annually, is not a direct replacement of Apple's own models but rather a strategic collaboration where Gemini will assist in training and enhancing Apple's proprietary on-device AI, with all processing occurring on Apple's private cloud servers to ensure user data privacy and isolation from Google. This move is seen as a tactical, transitional step for Apple to accelerate its AI capabilities and meet product launch timelines, especially after facing delays and setbacks with its in-house Apple Intelligence development and losing key AI talent to competitors like Meta. Despite this partnership, Apple continues its independent research and is reportedly developing its own trillion-parameter model, targeted for around 2027. The collaboration has drawn criticism, notably from Elon Musk, who raised concerns about the concentration of power with Google, which also controls Android and Chrome. For Google, the deal is a major win, boosting its market valuation, while OpenAI's existing partnership with Apple, which positions ChatGPT as a supplementary option for Siri, appears less central by comparison. The arrangement highlights Apple's pragmatic approach to bridging its AI gap while maintaining its l...

Written by: Xiaojinya, Zimu AI

Apple has "bowed its head".

On January 12 local time, Apple and Google issued a joint statement:

"After careful evaluation, Apple believes that Google's artificial intelligence technology provides the most robust foundation for the Apple base model, and is excited about the innovative experiences it will unlock for Apple users."

The two companies stated that these models will help support future Apple Intelligence features, including a more personalized Siri set to launch later this year.

As soon as the news broke, Musk was the first to express displeasure, commenting on X that this kind of "power concentration" is unreasonable, given that Google also owns Android and the Chrome browser.

This is tantamount to saying that Google now controls Apple's AI as well. But Apple would certainly disagree with this statement.

For Apple, this is a temporary bowing of the head, but by no means an admission of defeat or surrender.

The introduction of Google Gemini is more like a temporary worker Apple has found for the transition period; it will not completely replace Apple's self-developed model. Moreover, Apple is still developing a trillion-parameter model.

Gemini Has Not Replaced Apple's Self-Developed Model

The statements released by Apple and Google are limited and deliberately vague, but one thing is clear: Gemini will participate in the construction of the new Apple base model.

This matter was reported by Bloomberg's Mark Gurman back in November last year, when he cited sources familiar with the matter saying that Apple would pay Google about $1 billion annually.

However, this is not a simple "one-for-one" swap of the Gemini model for the Apple base model, but rather a system architecture that allows Apple to continue using its own model while not providing any data to Google.

As the initial rumor pointed out, Gemini will not be directly embedded into Apple's operating system.

Instead, everything will still be presented externally with the Apple base model at its core, but at the underlying architectural level, Gemini will become its "foundation." In other words, Apple's use of Gemini falls somewhere between "shelling" and "distillation," and of course, given the cooperation between the two parties, this is certainly legal.

This is crucial; interpreting this news as "Apple's AI will be powered by Gemini" is inaccurate.

When you buy an iPhone, it internally uses chips manufactured by TSMC, displays from Samsung, and various components from all over the world. But no one would therefore claim that Samsung is "powering" the iPhone.

The on-device Apple Intelligence will still be powered by the Apple base model.

In other words, the Gemini model will serve as a tool to help train the Apple base model, enabling it to better perform the various tasks Apple sets for it.

Last year, Gurman's report mentioned that, according to the agreement between Google and Apple, Google's Gemini model would be responsible for Siri's summarization and planning functions—components that help the voice assistant synthesize information and decide how to execute complex tasks. Some of Siri's functions will continue to use Apple's self-developed model.

The report also stated that the model would run on Apple's own private cloud computing servers, ensuring user data is completely isolated from Google's infrastructure. Apple has allocated AI server hardware to support the operation of this model.

As the current public statements from both sides are ambiguous, the above content has not been refuted.

This deal is fundamentally different from the "Google Search" type of cooperation because it is completely invisible to the end user. When conversing with Siri, there will be no Gemini branding, nor should there be.

Apple Has Not Given Up

Further cooperation with Google is Apple's "transitional" choice.

Apple users have been looking forward to the new Siri since 2024, when Apple showcased it at WWDC and said it would be rolled out gradually. But since then, it has been repeatedly delayed, becoming a major embarrassment for Apple.

Currently, multiple media reports indicate that the new Siri will be released in March with iOS 26.4.

In addition, Apple Intelligence itself has not been progressing smoothly either; the initially launched features like "News Summaries" have been questioned for making errors.

Worse still, when Meta suddenly initiated a "war for AI talent" last year, Apple also suffered heavy losses. The most well-known case was the head of Apple's AI base model team at the time, Ruoming Pang, being poached, reportedly offered a sky-high compensation of $200 million by Meta.

In this situation, if delays continue, it would not only mean falling behind in the AI race for Apple but would also severely damage market trust.

At this point, turning to a third party to push forward the most urgent matters (like the new Siri) according to the schedule and deliver results to users is Apple's top priority.

And choosing Google is not surprising.

Apple and Google have a long-standing cooperative relationship; for example, Google has been paying Apple over $20 billion annually for years to remain the default search engine in Apple's Safari browser.

Their cooperation has attracted antitrust lawsuits, but the latest ruling only required adjustments to specific practices; the $20 billion cooperation between the two is still permitted.

Furthermore, Gemini is currently one of the strongest models. According to previous coverage, Apple tested several mainstream models including Gemini, GPT, and Claude, and ultimately selected Gemini.

As an aside, Apple's emphasis on "user privacy" in this new cooperation with Google Gemini, stressing that everything runs on the "Apple Private Cloud" and that no data will be shared with Google, is also a lesson learned from past collaborations with Google. In 2012, the "Safari tracking cookie scandal" occurred, where Google was accused of exploiting a privacy setting vulnerability in the Safari browser to place tracking cookies, bypassing the policy of blocking third-party cookies by default. Ultimately, Google was fined $22.5 million by the FTC and required to make changes. Additionally, Google reached a $17 million settlement with 37 states and the District of Columbia to resolve a class-action lawsuit.

In summary, this cooperation between Apple and Google can be called a temporary "bowing of the head," but by no means "admitting defeat" or "giving up."

Apple is still developing its own trillion-parameter model, expected to be launched around 2027.

Even when that model is released, Apple will likely still retain Gemini for reinforcement training, comparative analysis, and further optimization. Using it as a "tool" to strengthen itself is precisely what Apple expects from Google Gemini.

The One Who Benefits Most is Still Google

Although we just spent a long篇幅 trying to clarify what this deeper cooperation between Apple and Google is and is not.

But it can be foreseen that the perception that "Apple's AI relies on Gemini" will be difficult to eliminate.

Some can't help but imagine: Imagine if by March or April, Siri suddenly becomes super easy to use......

This isn't hard to imagine. The awkward situation Apple faces now is that if the introduction of Gemini indeed has an immediate effect and significantly accelerates the progress of Apple Intelligence and Siri, then the evaluation that "Apple really isn't up to it" would be confirmed.

In this case, even if Apple successfully develops its own trillion-parameter model and moves from the transition phase to self-reliance, Gemini would still be considered the功臣 (meritorious servant), not Apple itself.

As for Google, the market has already given it positive feedback.

After the joint statement from Apple and Google was released, Alphabet's stock price rose by up to 1.7%, and its market capitalization closed the day by breaking through the $4 trillion mark for the first time, setting a company record, while Apple's stock price rose by less than 1%.

This cooperation also puts other peers under pressure.

Elon Musk, who spoke out to express dissatisfaction, is one of them. He criticized Google, which controls Android and Chrome, for now also penetrating Apple, calling it an embodiment of "excessive concentration of power," hinting at monopolistic tendencies.

Another who might be uncomfortable is OpenAI's Sam Altman. OpenAI previously reached a cooperation agreement with Apple, where ChatGPT would serve as a "supplementary option" when Siri answers user questions. That is, just as Siri sometimes lists web search results now, it would, when necessary, provide answers from ChatGPT in the future.

This cooperation had previously brought glory to OpenAI, but now it pales in comparison to the new cooperation between Apple and Google.

Clearly, this cooperative status of being a "supplementary option" on the front stage is far inferior to Google Gemini's status as a "foundational helper" in the back stage. It's like the Apple club is putting on a show: OpenAI is performing on stage, but Google is working the lights backstage. Altman has not yet commented on this.

And in the Chinese market, whether Apple is also looking for a "transitional helper" before it builds its trillion-parameter self-developed model remains to be seen.

Связанные с этим вопросы

QWhat is the nature of the partnership between Apple and Google regarding AI, as announced on January 12th?

AApple and Google announced a partnership where Google's Gemini AI technology will serve as the foundational base for Apple's on-device models. This is not a direct replacement but a system where Gemini helps train and support Apple's own models, with all processing occurring on Apple's private cloud servers to ensure data privacy and isolation from Google.

QAccording to the article, why did Apple choose to partner with Google's Gemini instead of relying solely on its own AI development?

AApple chose this partnership as a transitional solution to meet its deadlines, particularly for the launch of the new, more personalized Siri. The company has faced delays and setbacks with its own Apple Intelligence and Siri development, including the loss of key AI talent to competitors like Meta.

QHow does the article describe the potential market perception of this Apple-Google deal, and what is the risk for Apple?

AThe article states that the public perception will likely be that 'Apple AI is powered by Gemini,' which could solidify the view that 'Apple is not capable' on its own. Even if Apple later succeeds with its own trillion-parameter model, Gemini could still be seen as the功臣 (key contributor), damaging Apple's reputation.

QWhat was Elon Musk's reaction to the Apple-Google AI partnership, as mentioned in the article?

AElon Musk expressed dissatisfaction on X, criticizing the partnership as an unreasonable 'concentration of power' because Google already controls Android and Chrome, and this deal further expands its influence into Apple's ecosystem.

QHow does the status of OpenAI's partnership with Apple compare to Google's new role, according to the article?

AOpenAI's partnership with Apple is described as significantly less important. OpenAI's ChatGPT is a 'supplementary option' presented to users on the front end, similar to web search results, whereas Google's Gemini is a foundational 'helper' in the backend, akin to a backstage technician compared to a stage performer.

Похожее

Silicon Valley 'Startup Guru' Steve Hoffman: Web3 + AI Could Be a Trap

Silicon Valley investor and "Godfather of Startups" Steve Hoffman warns that combining Web3 with AI is likely a trap, not a promising venture. In an interview, Hoffman argues that while AI is a foundational technology touching all industries, Web3 adds complexity, friction, and regulatory risk without solving mainstream consumer or business needs. He advises founders to focus on deep, specialized applications where startups can out-iterate giants, rather than on generic features easily replicated by large tech companies. Hoffman observes that Silicon Valley will lead foundational AI research, while China excels at rapid, large-scale application and commercialization, particularly in robotics. He stresses that AI-driven autonomous agents capable of collaborative, multi-step tasks are 2-4 years away, which will cause significant job displacement. The solution is not to slow AI but to redesign business models around human-AI collaboration and reform social systems like education and retraining. For startups, Hoffman recommends focusing on vertical, expertise-heavy domains to build defensibility. He sees major opportunities in AI fraud detection and cybersecurity. Key founder mindsets include systemic thinking over feature-focus, relentless customer centricity, building adaptive teams, and deeply understanding AI's capabilities and limits. Hoffman is also leading a non-profit initiative to establish university centers aimed at training future leaders in responsible, human-value-aligned AI innovation.

marsbit1 ч. назад

Silicon Valley 'Startup Guru' Steve Hoffman: Web3 + AI Could Be a Trap

marsbit1 ч. назад

Token Inefficient, Economy Tokenless

The article "Tokens Aren't Economical, Economics Aren't Tokenized" analyzes a pivotal shift in the AI industry from a technology-driven narrative to one dominated by capital efficiency. It highlights two concurrent trends: a severe capital shortage due to the exorbitant and recurring costs of compute (e.g., OpenAI's high burn rate) and a wave of corporate spin-offs where major tech companies are separating their AI units (like Kuaishou's Kling and Baidu's Kunlunxin). The core argument is that AI's "anti-internet" business model, where user growth increases costs rather than profits, has created a disconnect between high valuations and actual cash flow. Spin-offs address this by allowing AI assets to be valued independently. Within a parent company, they are seen as cost centers, but as standalone entities, they are priced based on their growth potential and scarcity in the primary market, leading to massive valuation premiums (e.g., Kling's estimated value tripling post-spin-off). The industry is at an inflection point, moving from "model worship" to "value realization." The competition is evolving from a pure compute (GPU) race to a broader focus on systemic efficiency and full-stack engineering (involving CPUs and orchestration) to achieve viable commercialization. The year 2026 is framed as a critical moment where the industry must definitively answer how to economically translate AI capability into tangible business value, reshaping the sector's future power structure.

marsbit1 ч. назад

Token Inefficient, Economy Tokenless

marsbit1 ч. назад

Crossing the 'Memory Wall': The Wafer-Level Revolution and Computing Power Routes in the AI Inference Era

In 2026, a historic shift occurred in AI as major cloud providers' inference spending surpassed training spending for the first time, signaling a move from "building large models" to "using large models." This shifts the core challenge from computing power to the "memory wall"—the bottleneck of data movement (model weights, activations, KV Cache) between external DRAM and processors, where energy and latency from data transfer far exceed computation itself. Companies like Nvidia face GPU idle time due to bandwidth limits. In contrast, Cerebras Systems adopts a radical "wafer-scale" approach with its Wafer-Scale Engine (WSE). Instead of cutting a silicon wafer into many chips, Cerebras uses almost the entire wafer as one massive chip (WSE-3). This design provides 44GB of on-chip SRAM, delivering memory bandwidth thousands of times higher than traditional HBM (e.g., 21 PB/s vs. Nvidia B200). For LLM inference, weights are streamed layer-by-layer from external MemoryX storage to the chip, avoiding HBM bottlenecks. This results in token generation speeds 1.5–5 times faster than Nvidia's B200 in some models and significant advantages in first-token latency and long-context tasks. Additionally, Cerebras's architecture offers much lower interconnect power consumption (0.15 pJ/bit vs. GPU's ~10 pJ/bit). However, Cerebras faces challenges: SRAM scaling has slowed with advanced nodes, limiting future capacity gains; the chip requires specialized liquid cooling and custom software stacks; and its external I/O bandwidth (150 GB/s) is low compared to NVLink, hindering multi-system scaling for very large models. Competition is intensifying. Major players are pursuing three paths: 1) Developing proprietary inference ASICs (e.g., Google TPU, Microsoft Maia), 2) Leveraging advanced packaging (e.g., TSMC's SoW) to democratize wafer-scale-like integration, potentially eroding Cerebras's process advantage within a few years, and 3) Exploring optical interconnects for ultimate bandwidth. Commercially, Cerebras is transitioning from a hardware vendor to a service provider, facing the immense challenge of building high-power, specialized data centers to meet large contracts (e.g., 250MW/year from 2026–2028). In conclusion, the AI inference era presents a fundamental architectural trade-off. Cerebras opts for extreme physical optimization for low-latency, single-task performance, while Nvidia prioritizes versatility and massive cluster throughput. The path forward remains uncertain, with technology and business models still evolving in the race toward advanced AI.

marsbit1 ч. назад

Crossing the 'Memory Wall': The Wafer-Level Revolution and Computing Power Routes in the AI Inference Era

marsbit1 ч. назад

Has Bitcoin's 'Rebound Ended', Officially Entering the Late Bear Market Phase?

**Title: Has Bitcoin's Rebound Ended, Entering the Late Bear Market Phase?** **Summary:** Bitcoin's price has declined by 13% this week, signaling a potential return to late-stage bear market conditions. The price fell to around $67k, positioned between the Realized Price and Realized Cap Weighted Average. For the first time since early 2022, the Short-Term Holder cost basis has dropped below this key average, confirming a hallmark of late-cycle bear markets. Profitability metrics have collapsed sharply. The 7-day average of the Realized Profit/Loss ratio plummeted from a local high of 3.16 to 0.29, mirroring the February panic sell-off. Critically, the 90-day average never breached the threshold of 2, indicating the recent rally to $82k was a bear market bounce, not a structural shift. Realized losses surged to $1.35 billion daily, with $770 million coming from Long-Term Holders selling at a loss. This accelerating redistribution of supply from weak to strong hands is a necessary but ongoing process for a market bottom. The rally stalled almost precisely at the aggregate cost basis (~$83k) of US spot Bitcoin ETF investors, turning that level into strong resistance and leaving the average ETF holder underwater again. Spot market flows have turned decisively negative, showing sellers are dominating order books despite the price drop. While a significant futures long liquidation event cleared over $400 million in leverage, providing a potential reset, sustained spot demand is yet to materialize. Options markets continue to price in higher future volatility (Implied Volatility) than recent price action (Realized Volatility) has shown, with a persistent skew towards put options, indicating ongoing demand for downside protection. In conclusion, multiple metrics point to a fragile market structure. Resistance at the ETF cost basis, accelerating realized losses, dominant spot selling, and cautious options pricing all suggest the bear market trend persists. A sustainable recovery likely requires a resurgence of spot demand, ETF holders returning to profit, and a clear reduction in selling pressure.

marsbit1 ч. назад

Has Bitcoin's 'Rebound Ended', Officially Entering the Late Bear Market Phase?

marsbit1 ч. назад

Торговля

Спот
Фьючерсы

Популярные статьи

Как купить T

Добро пожаловать на HTX.com! Мы сделали приобретение Threshold Network Token (T) простым и удобным. Следуйте нашему пошаговому руководству и отправляйтесь в свое крипто-путешествие.Шаг 1: Создайте аккаунт на HTXИспользуйте свой адрес электронной почты или номер телефона, чтобы зарегистрироваться и бесплатно создать аккаунт на HTX. Пройдите удобную регистрацию и откройте для себя весь функционал.Создать аккаунтШаг 2: Перейдите в Купить криптовалюту и выберите свой способ оплатыКредитная/Дебетовая Карта: Используйте свою карту Visa или Mastercard для мгновенной покупки Threshold Network Token (T).Баланс: Используйте средства с баланса вашего аккаунта HTX для простой торговли.Третьи Лица: Мы добавили популярные способы оплаты, такие как Google Pay и Apple Pay, для повышения удобства.P2P: Торгуйте напрямую с другими пользователями на HTX.Внебиржевая Торговля (OTC): Мы предлагаем индивидуальные услуги и конкурентоспособные обменные курсы для трейдеров.Шаг 3: Хранение Threshold Network Token (T)После приобретения вами Threshold Network Token (T) храните их в своем аккаунте на HTX. В качестве альтернативы вы можете отправить их куда-либо с помощью перевода в блокчейне или использовать для торговли с другими криптовалютами.Шаг 4: Торговля Threshold Network Token (T)С легкостью торгуйте Threshold Network Token (T) на спотовом рынке HTX. Просто зайдите в свой аккаунт, выберите торговую пару, совершайте сделки и следите за ними в режиме реального времени. Мы предлагаем удобный интерфейс как для начинающих, так и для опытных трейдеров.

836 просмотров всегоОпубликовано 2024.03.29Обновлено 2026.06.02

Как купить T

Обсуждения

Добро пожаловать в Сообщество HTX. Здесь вы сможете быть в курсе последних новостей о развитии платформы и получить доступ к профессиональной аналитической информации о рынке. Мнения пользователей о цене на T (T) представлены ниже.

活动图片