🗞️In this edition

• Alphabet targets Meta with Gemini AI eyewear

• Sponsored: Kosmik – Create moodboards in seconds with AI

• ChatGPT enterprise usage explodes amid Google threat

• Trump: US gets 25% of Nvidia's China chip sales

• In other AI news –

  • Google Doppl gets shoppable discovery feed

  • Instacart groceries now available inside ChatGPT 

  • Claude code can now read Slack messages and write code

  • 4 must-try AI tools

Google plans first AI-powered glasses launch in 2026

What's happening:

Google said Monday it plans to launch the first of its AI-powered glasses in 2026, as the tech company ramps up efforts to compete against Meta in heating the consumer market for AI devices.

The Alphabet-owned company is collaborating on hardware design with Samsung, Gentle Monster, and Warby Parker, with whom Google agreed to a $150M commitment in May.

Google plans to release audio-only glasses allowing users to speak with Gemini AI assistant. Google also said there will be glasses with an in-lens display showing users information such as navigation directions and language translations. The company said the first glasses will arrive next year but didn't specify which styles that will include.

The glasses will be built on Android XR, Google's operating system for headsets.

AI wearables space has been gaining traction with Meta leading pack. Social media company's Ray-Ban Meta glasses were met with surprising success. Meta also released its own display glasses in September allowing users to see features like messages, photo previews, and live captions through a small display built into one lens.

Other companies like Snap and Alibaba have also been churning out their own AI glasses offerings as the small but competitive market continues to grow.

Why this is important:

Google returning to smart glasses after Google Glass failure shows confidence AI assistant capabilities solve previous problems.

The $150M Warby Parker commitment plus Samsung and Gentle Monster partnerships shows Google learned from past mistakes of controlling the entire value chain. Partnering with established eyewear brands addresses fashion and distribution.

Meta's Ray-Ban glasses success validated the market exists for AI eyewear. Google's entering a proven category, not creating a new one.

Our personal take on it at OpenTools:

Google Glass failed because it was creepy, expensive, and useless. AI assistants fix the useless part but not the other two.

Leadership saying "less advanced AI" was a problem with the original Google Glass is revisionist. Real problems were $1,500 price, terrible design, and recording people without consent making wearers social pariahs.

Partnering with Warby Parker, Samsung, and Gentle Monster addresses design and price issues. Warby Parker knows affordable eyewear. Gentle Monster knows fashion. That's learning from past mistakes.

But AI assistants in glasses still face a fundamental question: why? What task is better done through glasses than pulling out a phone? Navigation while walking, maybe. Real-time translation, possibly. Beyond that, unclear.

Meta's Ray-Ban glasses succeeded because they look normal and focus on audio (music, calls) plus camera for content creation. If Google's audio-only glasses match that with better AI assistant, there's a market.

This could work if glasses are affordable ($200-400 range), look normal, and AI assistant is meaningfully better than pulling out a phone. Big ifs.

Read More

Kosmik — Build moodboards fast with AI

Bring your ideas to life - without tabs, folders, or file names. Kosmik’s AI pulls the right images, videos, and notes into one clean workspace with just a few words.

✨ Work faster. - Find the perfect assets in seconds.
 🧠 Stay organized. - Everything you drop in is auto-tagged and categorized.
 🤝 Collaborate easily. - Share boards, gather feedback, and move projects forward together.

For Creatives, Designers, and Visual Thinkers

Subscribe to Kosmik Pro today with the code “Opentools40” and get a 40% discount on your first year of subscription.

Try it free

OpenAI highlights soaring enterprise adoption amid Google threat

What's happening:

OpenAI released data Monday showing ChatGPT message volume grew 8x since November 2024, with workers reporting they save up to an hour daily. The findings arrive a week after CEO Sam Altman's internal "code red" memo about Google's competitive threat.

Close to 36% of U.S. businesses are ChatGPT Enterprise customers versus 14.3% for Anthropic, per Ramp AI Index. However, the majority of OpenAI's revenue still comes from consumer subscriptions, a base threatened by Google's Gemini.

Organizations using OpenAI's API are consuming 320 times more "reasoning tokens" than a year ago, suggesting companies use AI for more complex problem-solving or are experimenting heavily and burning through tokens without necessarily getting long-term value.

Custom GPTs, which companies use to codify institutional knowledge, jumped 19x this year, now accounting for 20% of enterprise messages. OpenAI says BBVA regularly uses over 4,000 custom GPTs.

Participants reported saving 40 to 60 minutes per day with OpenAI's enterprise products, though that may not include time spent learning systems, prompting, or correcting AI output.

OpenAI reported a 36% increase in coding-related messages outside of engineering, IT, and research teams. Three-quarters of those surveyed say AI enables them to do technical tasks they couldn't do before, though more "vibe coding" could lead to security vulnerabilities.

Why this is important:

8x message growth and 320x reasoning token consumption in one year shows rapid enterprise adoption, but also raises sustainability concerns about energy usage and costs.

Majority of OpenAI revenue still coming from consumer subscriptions while facing the Google Gemini threat makes enterprise growth essential to justify $1.4T infrastructure commitments.

Custom GPTs up 19x and accounting for 20% of enterprise messages shows companies are customizing AI for specific workflows, not just using general chatbot.

Our personal take on it at OpenTools:

This is a marketing response to Altman's "code red" memo, not organic transparency.

Releasing enterprise usage data one week after leaked "code red" about Google competition is damage control. OpenAI's reassuring investors and customers that enterprise adoption is strong despite consumer market pressure from Gemini.

8x message growth sounds impressive but lacks context. What's the baseline? If it's a low starting point, 8x could still be modest absolute numbers. Without revenue disclosure or customer count growth, it is hard to assess significance.

320x reasoning token consumption is alarming, not just impressive. Reasoning tokens correlate with energy usage. If companies are burning 320x more tokens without proportional value creation, that's an unsustainable cost structure. OpenAI didn't address this concern when asked.

The $1.4T infrastructure commitment context is critical. OpenAI needs enterprise revenue growth to justify that spending. This report is evidence they're presenting to investors that the bet is paying off. But sustainability questions (energy costs, token burn rates, shallow feature adoption) remain unanswered.

Read More

Trump approves H200 exports with 25% cut

What's happening:

President Trump said Monday Nvidia will be allowed to ship H200 AI chips to "approved customers" in China on condition the US gets a 25% cut. Chinese President Xi "responded positively" to the proposal, Trump wrote on Truth Social.

The Commerce Department is finalizing details, and "the same approach will apply to AMD, Intel, and other GREAT American Companies," Trump added.

Nvidia and AMD agreed in August to share 15% of revenue from China chip sales with the US government. But China reportedly warned companies against using Nvidia's H20 chip, which was designed specifically for the country.

The H200 is a higher-grade chip than the H20 but not Nvidia's top-of-the-line product.

Nvidia shares rose about 2% after hours. "Offering H200 to approved commercial customers, vetted by the Department of Commerce, strikes a thoughtful balance that is great for America," a Nvidia spokesman said.

Why this is important:

25% revenue share versus August's 15% deal shows Trump extracting a higher price for China access. That's a 67% increase in government take from chip sales.

"Approved customers vetted by Department of Commerce" means US controls who in China gets advanced chips. That's strategic leverage over Chinese AI development, not open market access.

H200 being higher-grade than H20 (which China rejected) tests whether Beijing will accept better chips with strings attached or continue preferring domestic alternatives.

Xi "responding positively" to the revenue share proposal is a significant diplomatic concession if true. China paying the US government 25% of chip sales is an unusual arrangement in global trade.

Our personal take at OpenTools:

25% revenue share means the US government takes a quarter of Nvidia's China sales without providing product, service, or infrastructure. That's taxation on foreign customers for permission to buy American products.

China rejected H20 last year by telling domestic companies to use Huawei instead. Now Trump's offering higher-grade H200 with 25% tax and approval process. Why would China accept a worse deal for a better chip when they've already committed to domestic alternatives?

The policy applying to AMD and Intel creates consistency but also means all US chipmakers face the same constraints. If China rejects H200 with 25% tax, the entire US chip industry loses access, not just Nvidia.

August's 15% deal is already unprecedented. Jumping to 25% in four months shows Trump extracting maximum short-term revenue without considering long-term market dynamics. China's building Huawei, Cambricon, and other domestic alternatives specifically to avoid this dependency.

Semiconductor trade being "at the center of the AI race between US and China" means this isn't just a commercial transaction. It's a strategic competition where both sides are willing to sacrifice economic efficiency for control.

This could backfire spectacularly. China's already invested billions in the domestic semiconductor industry. Every punitive US policy accelerates that investment. 25% tax on imports makes domestic alternatives more competitive, even if technically inferior.

Read More

📢 What else is happening?

• Google’s AI try-on app Doppl adds a shoppable discovery feed – The tech giant says the idea behind the new feed is to display recommendations so users can discover and virtually try on new items. Nearly everything in the feed is shoppable, with direct links to merchants.

•  You can buy your Instacart groceries without leaving ChatGPT – OpenAI and Instacart are launching a grocery shopping experience inside of ChatGPT, allowing customers to brainstorm meal ideas, make a grocery list, and check out, all without leaving the chat interface.

• Anthropic's Claude Code can now read your Slack messages and write code for you – Allowing software engineers to delegate coding tasks without leaving the workplace messaging platform where much of their daily communication already happens.

👩🏼‍🚒Discover mind-blowing AI tools

1. GetMax - An AI-powered content marketing assistant that helps businesses plan, create, and optimize their content strategies

2. Any Summary - An AI-powered tool that quickly summarizes long interview audio or video files

3. PlayHT - An AI-powered text-to-speech (TTS) tool that can generate realistic audio using synthetic voices

4. Civitai - An online platform that makes it easy for people to share and discover resources for creating AI art