According to Bloomberg, Apple is working on a Siri chatbot intended to compete with ChatGPT, Gemini, and Claude. It is expected to be unveiled in less than six months, likely alongside iOS 27 at WWDC in June. While many of the technical details appear concrete, one key question remains: Will Apple charge for the Siri chatbot, and if so, how much?
A common pattern among large AI chatbots is that a free version is available, but limited. Full functionality requires a subscription. Anthropic, Google, OpenAI, and others operate this way. Apple is in a unique position – Siri has been free for years and is so deeply integrated into Apple products that a sudden paywall would be difficult to justify. At the same time, a Siri chatbot running on billions of devices would be expensive to operate, as each request requires computing power and therefore costs money. This tension is precisely what drives the debate.
What Apple plans to do with the Siri chatbot
According to Bloomberg, Apple is not planning a standalone app. Instead, the Siri chatbot will be deeply integrated into iOS, iPadOS, and macOS. This would make it available on a vast number of devices, as part of the system and across all of Apple's apps.
What the Siri chatbot should be able to do
Bloomberg describes the planned features as follows:
- can search for information on the internet
- can create content
- can generate images
- summarizing information
- can analyze uploaded files
Additionally, it should be able to control Apple devices and use personal data and screen information for searches and task completion. This sounds like almost everything existing chatbots like ChatGPT already offer, only with particularly tight integration into the Apple ecosystem.
On-device or cloud: Where AI actually runs
Some of these tasks can be performed directly on the device, especially with Apple's powerful A- and M-series chips. Nevertheless, according to Bloomberg, Apple is relying on a special AI model developed in collaboration with the Google Gemini team.
Model basis and performance limits
The model is intended to be roughly comparable to Gemini 3. Importantly, the full version of Gemini 3 doesn't even run locally on a high-end Mac, let alone on a mobile device. This implies that Apple needs server capacity for the Siri chatbot.
Apple servers are probably not enough
Although Apple has developed private cloud compute servers for AI functions, Bloomberg believes it is unlikely that this infrastructure will be sufficient for a fully-fledged Siri chatbot of the necessary scale.
Bloomberg also suspects that Apple is indeed discussing running the chatbot on Google servers. And: Google won't do it for free.
Computing costs and infrastructure: Why every Siri request costs money
Whether Apple uses its own private cloud compute servers or relies on Google's Tensor servers, the demand for computing power would be enormous. Every question to Siri and every generated image incurs ongoing costs.
Examples from the industry
- OpenAI is not profitable and spends billions on inference every year.
- OpenAI has committed to spending $1.4 trillion on infrastructure to keep up with demand – a sum OpenAI does not yet have.
- In 2025, Google spent approximately $85 billion on infrastructure to meet AI demand.
- In August, Google announced that an average text request in Gemini apps consumes 0.24 watt-hours of energy. Across all Google devices and products, this amounts to hundreds of millions of dollars annually in electricity costs alone.
These figures provide context for why a widely used Apple chatbot cannot run "on the side".
How Google Gemini is priced and what Apple could learn from it
Google has already integrated Gemini into Pixel smartphones and other Android devices, using a tiered pricing system.
Free vs. subscription at Gemini
- Free version: More cost-effective to operate, can answer questions, summarize texts, write emails, and control apps and smartphone functions.
- Gemini Advanced: $20 per month for an enhanced version with better reasoning, longer context for analyzing larger documents, and improved coding.
Apple could choose a similar model: a basic version of Siri for everyone and more advanced models behind a subscription.
Apple has a well-known freemium role model: iCloud
iCloud already works with a free initial subscription plus a paid upgrade. Apple offers 5 GB of cloud storage for free; anything beyond that costs extra. This logic could be applied relatively seamlessly to AI: free standard storage, paid "more".
Option 1: Temporarily free, to quickly gain market share
One scenario: Apple initially offers the Siri chatbot for free. This could attract users who currently pay for ChatGPT or other services.
The catch: ChatGPT, Claude, and Gemini all cost around $20 per month. It would be difficult for Apple to compete with these providers on running costs for one to two years if its own service were largely free. On the other hand, even undercutting current prices could quickly attract many customers and immediately make Apple a major player in the AI market.
Apple Intelligence is currently completely free, including images created with Image Playground. However, the possibilities are limited, and some functions run on the device.
Option 2: Paid levels directly from the start
It's also possible that Apple doesn't want to or can't permanently bear the AI costs itself. In that case, there could be paid options right from the start.
Price anchor in the market: around 20 dollars
AI providers have effectively settled on entry-level prices of around $20 per month. However, it's unclear whether this price is sustainable in the long run, given the increasing costs of training new models and the need to serve more users. The market currently sees, among others, the following price levels:
- ChatGPT Plus – 20 $/month
- Copilot Pro – 20 $/month
- Gemini Advanced – 19.99 $/month
- Claude Pro – 20 $/month
- Perplexity Pro – 20 $/month
If Apple offers a paid tier, it's likely that it will be positioned at this price point.
Siri and ChatGPT: What a replacement would mean
Apple is currently collaborating with OpenAI to forward complex queries to ChatGPT. Apple isn't paying OpenAI for this service, but is making ChatGPT accessible to millions of Apple users. This ChatGPT integration may be discontinued when Apple introduces its own Siri chatbot.
Possible side effect: Influence on the legal dispute with Elon Musk
The removal of ChatGPT integration could also affect the legal dispute between Apple and Elon Musk. Musk's company xAI has sued Apple and OpenAI for collusion. The allegation: ChatGPT is given preferential treatment over other AI bots like Grok, and Apple should also integrate other chatbots into Siri.
If Apple were to discontinue ChatGPT via Siri in favor of its own Siri chatbot, this would be no more unusual in terms of market logic than if:
- Google Gemini integrated into all Android devices
- Meta has limited its smart glasses to Meta AI
The timeline: iOS 26.4 as an interim step, iOS 27 as a major overhaul
More details are expected to emerge in the coming months. Currently, the following is being discussed:
- With iOS 26.4, Apple plans to introduce the first new LLM Siri features. These will utilize models running in Private Cloud Compute and based on an older generation of Gemini.
- The actual Siri chatbot is expected to arrive in the iOS 27 cycle and will be presented at WWDC in June, along with iOS 27, iPadOS 27 and macOS 27.
Apple's cloud strategy: Possible switch to Google servers
Bloomberg's report contains a particularly striking piece of information: Apple and Google are reportedly discussing running the next generation of Siri models directly on Google's servers, not on Apple's own.
Why this would be a change of course
Apple presented Private Cloud Compute as a concept intended to extend iPhone privacy into the cloud. However, this vision was developed under previous leadership. When Apple unveiled Private Cloud Compute at WWDC 2024, it was not yet foreseeable that the next steps would apparently require licenses for Google's Gemini models.
In this context, it seems plausible that the entire AI plan could still change. While Apple will likely want to keep as much processing as possible in its own cloud, the new Siri leaders – Craig Federighi and former Vision Pro manager Mike Rockwell – reportedly seem to prioritize practicality over idealism. There is clearly pressure to catch up and quickly deliver modern Siri features, even if this means deviating from previously communicated visions.
Chatbot experience: previously downplayed, now a priority
Apple had previously downplayed the premise of a chatbot experience. However, plans change, and Apple is clearly responding to the market success of services like ChatGPT.
Data privacy in the Google Cloud
If Apple is indeed relying on Google servers, it is reasonable to assume that Apple would negotiate the agreement in such a way that sensitive user data is not logged or stored by Google and is sufficiently shielded from Google's advertising and data collection activities.
External infrastructure is not new to Apple
Behind the scenes, parts of iCloud have relied on external providers like Amazon Web Services and Google Cloud Platform since its inception. These partners enable the operation and scaling of features like iCloud Photos, while Apple retains the encryption keys.
This also aligns with an earlier report: In 2021, it was reported that Google's cloud hosts around 8 exabytes of iCloud content.
Apple caught between AI catch-up, cost issues and a break in strategy
Apple plans to revamp Siri with a true chatbot, bringing it technologically on par with ChatGPT, Gemini, and Claude. The planned functionality is extensive, the system integration deep, and the timeline points to a unveiling during the iOS 27 cycle at WWDC in June. At the same time, Apple faces a classic AI dilemma: a service of this scale is expensive because every request incurs computing costs.
This opens up several realistic possibilities: a free basic version with a paid Pro tier, a time-limited free trial for rapid adoption, or a direct subscription model priced around $20 per month, as the market dictates. Additionally, Apple could remove the current ChatGPT forwarding feature from Siri, which would not only affect the product but also the ongoing debates surrounding the preferential treatment of individual AI providers.
And if the Siri chatbot were indeed to run on Google servers, that would represent a significant shift in Apple's cloud and privacy narrative – clearly signaling that Apple is currently prioritizing speed of implementation in its efforts to catch up in AI. The best products for you: Our Amazon storefront offers a wide selection of accessories, including those for HomeKit. (Image: fizkes / DepositPhotos.com)
- iPhone 2026: All new models and Apple's big plans
- Apple has big plans for AI's future: Siri 2.0 is just the beginning
- Apple 2026: These five new products are in focus
- Apple Creator Studio or Adobe Creative Cloud Pro?
- Apple plans a major Health app offensive with iOS 26.4
- Apple & Google: New AI deal fundamentally changes Siri
- Apple will modernize its Mac line in 2026 with two new products
- New Siri 2026: Overview of Apple's planned innovations
- iOS 27 is coming: All iPhone updates and dates for 2026
- iPhone Fold: Everything about the design, technology & possible price in 2026
- Apple will launch four new MacBooks in 2026
- iPhone 17e: More power and modern features at a bargain price
- Siri before relaunch: When will the big AI update be released?
- Apple is preparing to launch AI-based smart glasses
- MacBook innovations 2026: Three models launch at the beginning of the year
- iOS 27 in detail: New services, AI and focus on stability
- Apple in January 2026: These products and updates are coming
- iPad Air 2026: All the information on technology, design and release
- Apple TV 4K: All the rumors about the big 2026 update
- Apple 2026: An overview of all new devices, leaks, and dates
- Apple 2025: How broad and strong the product year was
- iPhone Fold: These features set new standards in the market
- iPad 12: Details on design, hardware, and the planned launch
