Tech Brew Ride Home - Gemini To Power The New Siri?

Episode Date: November 6, 2025

Apple is probably going to let Gemini power the new Siri, at least for a while. Does OpenAI want the government to backstop its AI buildout? And two new AI products. How about an AI smart-ring to remember your shower-thoughts, and what if Foursquare, but for the AI era? Apple Nears $1 Billion-a Year Deal to Use Google AI for Siri (Bloomberg) OpenAI Isn’t Yet Working Toward an IPO, CFO Says (WSJ) Microsoft Lays Out Ambitious AI Vision, Free From OpenAI (WSJ) Google’s rolling out its most powerful AI chip, taking aim at Nvidia with custom silicon (CNBC) Whisper Into This AI-Powered Smart Ring to Organize Your Thoughts (Wired) The Foursquare founder's new app is an AI-powered 'DJ' for neighborhood updates (Engadget) Learn more about your ad choices. Visit megaphone.fm/adchoices

It's peak pollination season, and my business is scaling fast. To keep the nectar flowing, I need a phone plan with top priority data speed. That's why I chose GoogleFi Wireless. My connections stay strong even when the hive is buzzing. Plus, unlimited plans start at $35 a month. Now, that's a deal that doesn't stay. Explore GoogleFi Wireless plans today. Plus taxes and government fees.

GoogleFi Wireless is not subject to data traffic deprioritization during times of high network usage. Welcome to the TechBrew right home for Thursday, November 6, 2025. I'm Brian McCullough today. Apple is probably going to let Gemini power the new Syria at least for a while. Does OpenAI want the government to backstop its AI buildout and two new AI products? How about an AI smart ring to remember your shower thoughts? And what if four square, but for the AI era. Here's what you missed today in the world of tech. If you, yes, you are looking for enterprise-grade identity automation minus the enterprise grade baggage, aka having your users log-on 500 times, yesh ID delivers advanced IAM automation without moving teams onto a legacy identity provider. Whether you use Google Workspace, Microsoft 365 or Octa, Yesh ID integrates directly. No rebuilds or rip-and-replaces are required. Yesh ID helps IT and security teams reduce risk, not just tickets. And so IT teams everywhere are breathing a collective sigh of relief.

Every access, change, review, and approval is tracked and exportable, helping security teams effortlessly demonstrate compliance with SOC2, ISO, or HIPAA. IT and security teams can spot risk before it becomes a finding. Learn more at yesh ID.com slash tech brew. That's YESHID.com slash tech brew. Mark German is reporting that Apple is finalizing a deal to pay Google around a billion a year for a $1.2 trillion parameter Gemini model to help power the new Siri, which is on track for next spring. Quoting Bloomberg, the iPhone maker is banking on Google's help to rebuild

series underlying technology setting the stage for a new slate of features next year. The Google models 1.2 trillion parameters, a measure of the AI software's complexity, would dwarf the level of Apple's current models. Apple had previously mold using other third-party models to handle the tasks, but after testing Gemini, OpenAI's chat GPT, and Anthropics Clod, Apple zeroed in on Google earlier this year. The hope is to use the technology as an interim solution until Apple's own models are powerful enough. The new series on track for next spring, Bloomberg has reported, given the launch is still months away, the plans and partnership could still evolve. Apple and Google spokespeople declined to comment. The custom Gemini system represents a major advance from

the 150 billion parameter model used today for the cloud-based version of Apple Intelligence. The move would vastly expand the system's power and its ability to process complex data and understand context. Known internally as Glenwood, the effort to fix Siri with a third-party model has been led by Vision Pro headset creator Mike Rockwell and Software Engineering Chief Craig Federigi. The new voice assistant itself planned for iOS 26.4 is codenamed Linwood. Under the arrangement, Google's Gemini model will handle series Summarizer and planner functions, the components that help the voice assistant synthesize information and decide how to execute complex tasks. Some Siri features will continue to use Apple's in-house models.

The model will run on Apple's own private cloud compute servers, ensuring that user data remains walled off from Google's infrastructure. Apple has already allocated AI server hardware to help power the model. While the partnership is substantial, it's unlikely to be promoted publicly. Apple will treat Google as a behind-the-scenes technology supplier instead. That would be make the pact different than the company's Safari browser deal, which made Google the default search engine. The agreement is also separate from earlier talks about integrating Gemini directly into Siri as a chatbot. Those discussions came close to fruition in both 2024 and earlier this year, but ultimately didn't materialize into a feature.

The partnership also doesn't weave Google AI search into Apple's operating system. Apple still doesn't want to use Gemini as a long-term solution, despite the company bleeding AI talent, including the head of its models team, management intends to keep developing new AI technology and hopes to eventually replace Gemini with an in-house solution, the people said. To that end, the company's models team is working on a one trillion parameter cloud-based model that it hopes to have ready for consumer applications as early as next year. Apple executives believe it can reach a similar quality level as the custom Gemini offering, but Google continues to enhance Gemini and so catching up won't be easy, end quote.

OpenAI chief financial officer Sarah Fryer says an IPO is not on the cards right now, but that OpenAI hopes the U.S. government will quote, backstop financing of its data center deals, quoting the journal. Speaking at the Wall Street Journal's Tech Live conference, Fryer threw a dose of cold water on what could become one of the largest public listings in history. She said the AI giant's conversion to a new structure doesn't portend an imminent public offering as the company prioritizes growth and R&D over profitability. IPO is not on the cards right now, Friar said. We are continuing to get the company into a state of

constantly stepping up into the scale we are at, so I don't want to get wrapped around an IPO. The company has discussed a public listing as soon as 2027, the Wall Street Journal reported. As Open AI ramps up its spending on data center capacity to unheard of levels, the company is hoping the federal government will support its efforts by helping to guarantee the financing for chips behind its deals, Friar said. The depreciation rates of AI chips remain uncertain, making it more expensive for companies to raise the debt needed to buy them. This is where we're looking for an ecosystem of banks, private equity, maybe even governmental, the ways governments can come to bear, she said.

Any such guarantee can really drop the cost of the financing, but also increase the loan to value, so the amount of debt you can take on top of an equity portion. Friar said Open AI could reach profitability on very healthy gross margins in its enterprise and consumer businesses quickly if it weren't seeking to invest so aggressively. I'm not overly focused on a break-even moment today, she said, I know if I had to get to break even, I have a healthy enough margin structure that I could do that by pulling back on investment. Friar said the company is trying to find new ways to grow sales beyond its chat GPT subscription business. She emphasized the growth in OpenAIs enterprise sales saying that they now account for roughly 40% of revenue up from 30%

at the beginning of the year. Many corporate customers are now moving from pilot to full production, she added, citing sectors such as financial services and health care. Farris said the company would end the year with about two gigawatts of computing power to help train and power its AI models up from 200 megawatts two years ago. It rents much of that capacity from Microsoft Oracle and CoreWeave. OpenAI is also building its own data centers and recently signed a giant deal with NVIDIA to lease the chips needed for the effort. As part of the deal, NVIDIA is discussing, and guaranteeing some of the loans OpenAI plans to take out to build new data centers of the journal reported. We are talking about country-sized deployments, she said, about how much computing power OpenAI intends to bring online, end quote.

Also possibly related to that same conference, Microsoft AI CEO Mustafa Soleiman, has laid out that company's plans to develop AI self-sufficiency beyond Open AI, like releasing its own voice image and text models, quoting the journal. While he praised Open AI and the work the company of have done together, he offered a criticism of treating AI systems as though they have human-like feelings or rights. AI chatbots shouldn't trick people into thinking they are having conversations with sentient beings, he said. Sullyma noted Microsoft's focus on powerful software tools that can help people accomplish their work, improve medical diagnoses, and play a role in scientific breakthroughs that will offer the world plentiful clean, renewable energy. AI is going to become more human-like,

but it won't have the property of experiencing suffering or pain itself, and therefore we should over-emopathize with it, Suleiman said in an interview. We want to create types of systems that are aligned to human values by default. That means they are not designed to exceed and escape human control. Microsoft's co-pilot chatbot relies heavily on OpenAI, but Microsoft is building testing and releasing its own voice image and text models. Microsoft has access to OpenAI's models until 2032, a schedule Suleiman said, gives his team time to make its models into leading technology. Healthcare is a priority for Microsoft AI and one of the first industries Suleiman expects to be touched by superintelligence. The company recently struck a partnership with Harvard Health

to provide trustworthy responses in co-pilot and rolled out other features to find doctors based on location language and other preferences. It also developed an AI tool that it said diagnosed disease in a test at a rate four times more accurate than a group of doctors at a fraction of the cost. It is these sorts of instruments that prove AI can be powerful in non-conversational ways, Suleiman said. AI diagnostic tools are very close to being market-ready, he said. Microsoft's models will be built with containment in mind, including probing and testing the models to ensure they only communicate in a language that humans understand and designing systems to avoid appearing as if they are conscious, he said, end quote.

Ready to soundtrack your summer? With Red Bull Summer All Day Play, you choose a playlist that fits your vibe the best. Are you a festival fanatic, a deep end DJ, a road dog, or a trail mixer? Just add a song to your chosen playlist and put your summer on track. Red Bull Summer All Day Play. Red Bull gives you wings. Visit Red Bull.com slash bright summer ahead to learn more. See you this summer. Study and play. Come together on a Windows 11 PC. And for a limited time, college students get the best of both worlds. Get the Unreal. college deal. Everything you need to study and play with select Windows 11 PCs. Eligible students get a year of Microsoft 365 premium and a year of Xbox GamePass Ultimate with a custom color Xbox wireless controller. Learn more at Windows.com slash student offer.

While supplies last, ends June 30th, terms at AKA.m.m.m.m.com slash college PC. Own it all. Pay off your home, travel for life, drive a Ferrari. In celebration of the world premiere of the Monopoly Big Board Buckslot Machine by Aristocrat Gaming, Yamava Resort and Casino at San Manuel is giving. giving one person a $1.6 million dream package. The biggest prize in Yamava's history. Club Serrano members can earn daily instant prizes and secure a spot in the finale May 29. Don't pass go and own it all. Only at Yamava, celebrating its 40th anniversary. You win? Details at yamava.com must be 21-20. Please gamble responsibly.

Monopoly is a trademark of Hasbro. Hasbro is not a sponsor of this promotion. Google says Ironwood, its seventh generation, TPU will launch in the coming weeks and is more than four times faster than its sixth-gen TPU. That's something. Quoting CNBC, The chip built in-house is designed to handle everything from the training of large language models to powering real-time chatbots and AI agents. In connecting up to 9,216 chips and a single pod,

Google says the new Ironwood TPUs eliminate data bottlenecks for the most demanding models and give customers the ability to run and scale the largest most data-intensive models in existence. Google is in the midst of an ultra-high-stakes race alongside rivals Microsoft, Amazon, and Meta to build out the AI infrastructure of the future, while the majority of large language models and AI workloads have relied on Nvidia's graphics processing units, Google's TPUs, fall in the category of custom silicon, which can offer advantages on price, performance, and efficiency. TPUs have been in the works for a decade. Ironwood, according to Google, is more than four times faster than its predecessor, and major customers are already lining up. AI startup Anthropic plans to use

up to one million of the new TPUs to run its clawed model, Google said. Alongside the new chip, Google is rolling out a suite of upgrades meant to make its cloud cheaper, faster, and more flexible, as it vise with larger cloud players, Amazon Web Services, and Microsoft Azure. In its earnings report last week, Google reported third quarter cloud revenue of $15.15 billion, a 34% increase from the same period a year earlier. Azure revenue jumped 40% while Amazon reported 20% growth for AWS. Google said it signed more billion-dollar cloud deals in the first nine months of 2025 than in the previous two years combined. To meet soaring demand, Google upped the high end of its forecast for capital spending this year to $93 billion from $85 billion. We are seeing substantial demand for

our AI infrastructure products, including TPU-based and GPU-based solutions, CEO Sundar Pichai said on the earnings call, it is one of the key drivers of our growth over the past year, and I think on an going forward basis, I think we continue to see very strong demand, and we are investing to meet that, end quote. Finally, today, two new AI-focused, products for you. First, Sandbar has unveiled a $249 plus stream ring, an AI-powered smart ring for transcribing audio notes into text via an app available to pre-order shipping in the summer of 2006. Sandbar wants to capture the thoughts that usually vanish between, you know, subway stops and showers and the like. The stream ring is an AI-powered smart ring designed to turn fleeting

inner monologues into organized notes and conversations, founded by alumni of control labs and later meta, the company has raised $13 million and is emerging after two years in stealth. Worn on the index finger, stream ring activates with a tap and hold on a capacitive edge. Speak or whisper into it, the audio isn't stored. Instead, speech is transcribed to text in the stream app. A quick tap can mute any AI response, and the hardware is slated to be waterproof at launch. The ring also doubles as a media controller, tap to play or pause, double tap to skip, swipe for volume, ensuring basic utility even if back-end services ever go dark. Health tracking isn't currently included. Pre-orders are open now. There's a silver version at $249 and a gold one at $299 with a

sizing kit. Purchases include three months of Stream Pro for unlimited notes and chats. Afterward, the plan is $10 per month with a limited free tier available. Shipping is targeted for summer of 26. Access to the app currently requires the ring, though that may change. Functionally, stream apparently sits between a notes app and an AI chatbot. Users can ask for, say, recipe ingredients, auto-build shopping lists, get reminders in store with haptic configurations and snapshot ideas into organized notes. It draws on multiple large language models with a blend of on-device, phone, and cloud processing. Responses can speak back in a near clone of the user's voice via 11 labs. Alternate voices are available. Apparently, early usage skews toward longer back and

force at around 60% of back-on-force, with the remainder split between note creation and single queries. Battery life is reported to be all day, with desktop access sharing and time-based reminders planned. Sandbar frames the product as a low-friction self-extension that prioritizes user control and safety. And For Square founder, Dennis Crowley's new startup Hopscotch Labs, has released B-Bot, which combines AI audio and location-based social features on the App Store, quoting and gadget. Yes, it's another location-based social app, but rather than the check-ins, Crowley first popularized more than 15 years ago, BeBot has a very 20-25 take on the concept. Instead, the app is an AI-powered DJ that can deliver contextual audio updates to your ears as you move

throughout your day. Crowley describes B-Bot as an app for AirPods, though it will work with any type of headphones as well as smart glasses with audio capabilities like metas. Whenever you put your AirPods in, it turns on, Crowley explains in a post on medium. Whenever you take your AirPods out, it turns off. And when it's on, it'll push you snippets of audio about the people, places, and events that are nearby. To do this, you'll need to give the app access to your location and share a handful of keywords about your interests. You can also share your contacts to get updates from friends who are using the app. The BeBot DJ, which of course has an AI voice, will then be able to talk to you as you go throughout your day and alert you to interesting events, landmarks, or updates from friends who happen to be nearby.

In some ways, it sounds like Crowley is trying to recreate some of the serendipitous IRL social interactions enabled by the original version of Foursquare. B-Bot doesn't have mayorships, badges, or any of the gamification features that helped popularize Foursquare, but it's meant to have some of the same playful spirit of OG Foursquare, according to Crowley. Foursquare shut down its city guide app of the same name earlier this year, though it's check-in app swarm lives on. And because it's 2025, there's also a whole bunch of AI thrown in, including a mix of different LLMs and synthetic voices. The app is powered by a TikTok-style algorithm, Crowley says, but one that's focused on what's happening nearby and in real life, end quote.

Speaking of Dennis and IRL, I only met Dennis in real life this week. That's despite me having interviewed him for the internet history podcast, which you'll hear soon, and him being an LP in the Right Home AI Fund for years now. Thank you, BetoWorks, for putting that IRL connection together. Talk to you tomorrow. Ambition comes in all shapes and sizes. At First Citizens Bank, we roll with your goals because we're built for what you're building. Fit for your ambition for Citizens Bank.