Get tomorrow's brief in your inbox
Today: Anthropic overtook OpenAI in business spending for the first time while ChatGPT's market share slipped below 50%. SpaceX paid $60 billion for AI coding tool Cursor, betting big on enterprise development workflows. Zapier made its pricing simpler and doubled down on AI automation that actually works for small teams.
For the first time since launch, ChatGPT's market share dropped below 50%, hitting 46.4% by May while Gemini climbed to 27.7% and Claude reached 10.3%. More importantly, Anthropic crossed a bigger milestone: it now captures 41% of business spending on AI subscriptions compared to OpenAI's 39.5%, according to payment data from Ramp.
Source: TechCrunch | Anthropic's growth
What this means: Users are no longer locked into a single AI assistant. They're testing multiple tools and picking the one that fits their work. Claude is winning on coding and productivity tasks, with 13% of users paying for subscriptions (the highest conversion rate in the industry). Gemini benefits from Google's ecosystem integration, making it the default for many Android and Google Workspace users.
If you're a business owner, this is good news. Competition is driving prices down and features up. Try multiple assistants for a month and see which one actually saves you time. Don't assume ChatGPT is still the best just because it was first.
Zapier simplified its pricing and went all-in on unlimited workflows. Every plan now includes unlimited Zaps (automations), access to Copilot (AI assistant for building workflows), and Zapier MCP, which lets you trigger automations directly from ChatGPT or Claude. The free tier gives you 100 tasks per month. Professional starts at $19.99/month (750 tasks), and importantly, Zapier doesn't count logic steps like Filters and Paths toward your task limit.
Pricing details | What counts as a task
If you're already using Zapier, check whether you're hitting task limits on wasted steps. Moving to logic-based routing instead of multiple action steps can cut your bill significantly. If you haven't tried automation yet, the free tier is generous enough to handle real workflows, not just demos.
Zapier also published a complete guide to employee onboarding automation. The key move: connect your HR system (BambooHR, Rippling, HiBob) to the rest of your tech stack so that when someone's status flips to "hired," it triggers everything else automatically (Slack account, Google Workspace, software licenses, team assignments).
If you're still manually creating accounts for new hires, start here. The time savings on your next three hires will pay for the automation setup.
Google released Android 17 with deeper Gemini integration. The update includes Gemini Omni for video editing in conversations, Lyria 3 for music generation from text or images, and improved AudioLM speech translation on Pixel 10a devices. The "bubble bar" feature lets you organize recent apps as bubbles at the bottom of your screen for faster multitasking.
For small businesses, the real value is in how Gemini ties into Google Workspace and your existing Android workflow. If your team is already on Gmail and Google Drive, Gemini integration is basically free productivity gains.
Databricks hit $6.9 billion in annualized revenue (up 80% year-over-year), with $1.7 billion coming from AI products. The company's tools let you query business data using natural language (Genie) and build custom AI apps (Agent Bricks). The catch: as AI agents make more queries, Databricks' costs rise and margins shrink.
If you're using Databricks or similar data platforms, their new Unity AI Gateway helps you track token usage and avoid runaway AI bills. Set budgets before you deploy, not after the invoice arrives.
Plaud, the AI notetaker hardware company, crossed $100 million in annual recurring revenue after shipping over 2 million devices. Its Plaud Pin and credit-card-sized gadgets record and transcribe meetings without screens, then generate summaries and action items. Nearly 50% of device buyers upgrade to paid plans for extra transcription minutes.
If your team takes a lot of client meetings or site visits, a $179 device that captures everything and generates follow-ups might be cheaper than the time lost to manual notes. The free tier gives you 300 minutes per month, which is enough to test whether it's useful.
AI is driving measurable results in retail and ecommerce, according to multiple surveys. Nearly half of Americans (49%) now rely on AI recommendations when making purchases. AI shopping assistants have driven a 1,247% increase in conversions, though only 6% of consumers are ready to let AI make purchases autonomously.
Ecommerce trends | AI in retail
For online stores, focus on product discovery tools and personalized recommendations rather than fully automated checkout. Customers want AI to help them find the right product faster, but they still want to click "buy" themselves.
Hyper-personalization is delivering real ROI. Slazenger achieved a 49x return by implementing personalized shopping experiences using AI to analyze customer behavior and deliver targeted product recommendations when customers are most likely to engage.
If you run an online store, start with Send-Time Optimization (delivering emails when customers actually open them) and dynamic product recommendations based on browsing history. Both are available through most ecommerce platforms without custom development.
KPMG pulled one of its own AI reports after alleged hallucinations, creating an awkward situation for a consulting firm that sells "AI Trust" services to clients. The report contained claims that several major organizations said were incorrect.
This is a reminder that even big firms with expensive AI tools can ship bad outputs. If you're using AI to generate reports or client deliverables, have a human review the facts before it goes out the door. AI is excellent at writing, terrible at verifying.
SpaceX is acquiring Cursor for $60 billion in stock just days after its record-breaking IPO. Cursor is an AI coding tool that helps developers generate, edit, and review code. It crossed $1 billion in annualized revenue in November but saw its market share drop from 41% to 26% as Anthropic's Claude Code took over.
CNBC report | The Verge coverage
For business owners, this signals that AI-assisted coding is becoming table stakes for software development. If you're hiring developers or building internal tools, ask whether they're using AI coding assistants. Teams using these tools ship features faster and spend less time on repetitive code.
Robinhood cut 10% of its workforce (290 people) without mentioning AI, a notable shift from the typical tech layoff script. CEO Vlad Tenev framed it as moving toward "lean, hyper-focused teams" and "flatter organizational structures" rather than blaming automation.
The "lean team" narrative is now the default playbook across tech. Companies are cutting middle management and assuming AI tools will fill the productivity gap. If you're planning headcount for 2026, factor in how much work your existing team can absorb with better tools before you hire.
60% of U.S. consumers say brands that use "AI" in their messaging are a turnoff, according to a WordPress VIP survey. Even though 60% of enterprise companies saw increased traffic from AI search engines, 86% of consumers don't fully trust AI and still want to see original sources.
If you're marketing AI tools or features, focus on what the tool does, not the fact that it uses AI. "Automatically categorize customer emails" beats "AI-powered inbox assistant." Let the value speak for itself.
Set AI spending limits before you deploy. Tools like Databricks' Unity AI Gateway let you set token budgets and get alerts before you burn through your allocation. If you're using OpenAI, Anthropic, or any API-based AI, set up billing alerts today. AI costs scale faster than you think, especially if you're running agents or processing large datasets.
Try local AI models for repetitive coding tasks. Georgi Gerganov (creator of llama.cpp) confirmed that Qwen3.6-27B is handling daily coding maintenance tasks reliably on local hardware. If you have developers doing small, repetitive code updates, a local model could cut your API costs to zero for those tasks.
The AI assistant market is fragmenting fast. ChatGPT still leads in total users but no longer dominates business spending or user retention. Companies are realizing they need multiple tools: Claude for coding and productivity, Gemini for Google Workspace integration, ChatGPT for general tasks.
This is good for buyers. Competition is forcing better pricing, more features, and clearer differentiation. The tools are also getting more reliable. Startups like Probably are building validation layers that catch hallucinations before they reach users, achieving 99.99% accuracy by running smaller models with tighter guardrails.
Watch for more acquisitions like SpaceX buying Cursor. AI coding tools are becoming infrastructure, not novelties. If your business builds software (even internal tools), expect your developers to demand access to these assistants.