After years of broken promises, false advertising, embarrassing delays, and a Siri that has lagged far behind competitors, Apple is finally preparing to deliver a meaningful upgrade to its voice assistant. According to The Information, the long-awaited overhauled Siri will launch in September 2026 alongside iOS 27, relying on a hybrid approach that taps Google Cloud’s Gemini models and Nvidia’s Blackwell GPUs for the heavy lifting.
This is the same Siri overhaul that was teased at WWDC 2024, delayed multiple times throughout 2025, and pushed into 2026 amid internal missteps, executive shakeups, and what critics have called a lack of urgency in Apple’s AI strategy.
Tim Cook Bears Responsibility
Tim Cook, Apple’s outgoing CEO, ultimately owns these repeated delays. In fact, this whole debacle may have been what finally precipitated Apple’s CEO transition. Apple publicly hyped advanced, personalized Siri capabilities as part of Apple Intelligence at WWWC 2024, only to admit they needed “more time to meet our high-quality bar.” That phrase became a recurring corporate euphemism while rivals like Google and OpenAI surged ahead. Cook addressed the setbacks in earnings calls and internal meetings, but the pattern points to leadership-level complacency if not utter blindness.
Under Cook’s tenure, Apple has excelled at operational execution and services growth, but it has repeatedly found itself playing catch-up in transformative technologies. The decision to lean on Google and Nvidia for this Siri reboot — after heavily marketing on-device privacy and Apple Silicon prowess — underscores how far behind the company was in scaling generative AI. Internal reshuffles (including changes to AI leadership) and reports of “ugly, embarrassing” delays further highlight execution shortfalls on his watch.
What the New Siri is Expected to Deliver
• Hybrid AI Architecture: Simple tasks stay on-device for speed and privacy. Complex queries route to Google’s Gemini via Nvidia-powered cloud infrastructure with confidential computing encryption.
• September Launch: Expected with iOS 27, iPadOS 27, macOS 27, tvOS 27, etc. More previews likely at WWDC 2026 (starting June 8th).
• Potential Impact: If it finally works as promised — handling context, interruptions, multi-step tasks, and personalization — LLM Siri could quiet critics and help drive iPhone, iPad, and Mac upgrades and sales.
This hybrid pivot represents a pragmatic (if humbling) admission that Apple’s late, go-it-alone approach wasn’t scaling fast enough. For loyal users tired of Siri’s limitations, it’s welcome news. But the multi-year wait leaves a sour taste — one that lands squarely at the top of the org chart.
MacDailyNews Take: With the right leader, Apple has the resources, talent, and ecosystem to lead. The question now is whether WWDC26 begins to deliver a Siri that actually works or if it’s just the latest patch on a decade-plus Siri underperformance. On that point, we’ve been hearing very hopeful murmurs for months now, so things could be looking up very soon!
Please help support MacDailyNews — and enjoy subscriber-only articles, comments, chat, and more — by subscribing to our Substack: macdailynews.substack.com. Thank you!
Support MacDailyNews at no extra cost to you by using this link to shop at Amazon.
Has there been any indication of what devices this will work on? As someone with OG HomePods, an M1 iPad, an 8 gig m2 MacBook air, and an iPhone 14, are ANY of these devices likely to run the new siri?:?
Everything I’ve read is that it will require a 2025 or later variant of an iPhone or an M4 Mac or later to be able to take advantage of the vast majority of this new functionality. But, this is just based upon “anonymous sources” from within Apple. So, of course, this could be way off.
Hopefully we’ll know more details next week.
“In theory,” it should be able to run anywhere ( TV, Watch, etc.). As they described it a couple years ago, Apple Intelligence is supposed to do Whatever It Can on the local device, and do Everything Else on a secure iCloud server. On the other hand, at this point, there’s no reason to believe that their half-baked plan years ago is the same one they have now. And maybe they’ll announce something equally impressive, and equally unfounded in reality.
We’ll know more on Monday.
Thanks. Yes, they’ve been so full of shit about AI so far, gonna have to wait and see.
One would hope Ternus will be Steve Jobs-ing this! That is. In parallel, a team will be working with Apple’s own M variant Ultra or beyond processors to build its own servers and have its own generative LLM AI, and keep working and working and working and in 5 years have it equal to or better than the fragmented market solutions of today.
Apple did this with their OS and own chipset and there is zero reason – other than next quarter short-sightedness, this should not be being done. 100% this should be being done.
Apple should have also had its low earth orbit Attlee liable system up by now to have its own in house carrier solution and doesn’t have jack.
So much financial shortsightedness was brought to the forefront by Tim Cook, while long-term highly profitable. Highly controllable solutions were shoved off into the distance or abandoned altogether…
Meanwhile, people with patients like Elon Musk, and vision, have taken the Raines in several areas where Apple should be leading the pack.
Here’s to hoping John Tennis knows what the heck he is going to do and has a vision for the company beyond next quarter.
“With the right leader, Apple has the resources, talent, and ecosystem to lead.”
That leader is not Craig Federighi
Good riddance Tim Cook hope he does not try and stop further innovation as Chairman otherwise they should move him on soonest.
From my vantage point there is little to be excited about in this article. It’s more evidence that Apple has lost its mojo. The fact that they are dependent upon Google and Nvidia for their innovation speaks volumes.
Apple needs to catch up soon. Apple cannot do that 100% on its own. That’s fact whether Apple fans like it or not.
If Apple uses Google and Nvidia as a temporary means of catching up over the next couple of years then I’m 100% OK with that.
If Apple uses Google and Nvidia over the next dozen years and irreversibly ties Apple to Google and Nvidia that may be the dumbest move Apple has ever done — even surpassing Sculley’s licensing of the MacOS source code to Microsoft back in the day.