{"version":"1.0","type":"rich","provider_name":"Acast","provider_url":"https://acast.com","height":250,"width":700,"html":"<iframe src=\"https://embed.acast.com/$/69f3e0f28beeba531023e35c/6aa0bf393dd40cf71e701314?\" frameBorder=\"0\" width=\"700\" height=\"250\"></iframe>","title":"Week of September 7th 2026","thumbnail_width":200,"thumbnail_height":200,"thumbnail_url":"https://open-images.acast.com/shows/69f3e0f28beeba531023e35c/1788919536437-9639a9f3-2fc1-4a5c-b7b8-b16f4b8ecaf7.jpeg?height=200","description":"<p>For three years AI was sold like a gym membership: one flat number, use it all you want. This week it started being sold like electricity. OpenAI's own rate card now prices a message at forty cents and a research report at two dollars — and then it shipped a brand-new model that burns two and a half times the credits and slid it into the same $20 plan. Underneath all of it, Nvidia posted a record quarter and warned its own costs are rising because of the AI boom itself. The sticker price was fiction. The meter is the truth.</p><p><strong>In this episode:</strong></p><p><br></p><ul><li><strong>GPT-6 Astra</strong> launched Thursday — \"the world's most intelligent and aligned model\" — at $10/$50 per million tokens, 2.5x the model it replaces, \"included within existing subscription allowances.\" Translation: same monthly price, you hit the wall faster, and the wall is the product. OpenAI turned it off by default for enterprises. So should you.</li><li><strong>The rate card, decoded.</strong> One credit = four cents. A regular question: 40¢. A pro question: $2. Agent mode: $1.20 a message. A deep-research report: $2. An image: 20¢. Voice: 20¢ a minute. A PowerPoint: up to $2. Per task, insultingly cheap. The risk is the plumbing — OpenAI's own help page says a blank monthly-reload box \"allows unlimited automatic reload purchases,\" and per-seat limits are \"no limits specified\" by default.</li><li><strong>Nvidia's $96.2B quarter</strong> — and the stock fell, because margins are shrinking on memory costs that the CFO says are \"driven in large part by the AI buildout itself.\" $279B in supply commitments, a paused cloud-financing program, and a $12.9B purchase of Hugging Face. The wholesale cost of thinking is going up; don't expect the top of the rate card to fall.</li><li><strong>Spotlight — two ways to put AI on the office desks:</strong> <strong>ChatGPT Business</strong> ($20/user annual, $25 monthly, 2-seat minimum; the best brain plus a visible meter; $100 Premium seat for the heavy user) vs <strong>Microsoft 365 Business Standard with Copilot</strong> ($23.50/user annual; Copilot inside Word/Excel/Outlook/Teams, flat, no meter on the assistant; Copilot Business add-on $21, $18 promo through September). Honest take: Google shop → skip both, you already have Gemini; Microsoft shop → the bundle; agent-curious → ChatGPT Business with every cap set on day one; two-person office on free accounts → skip both.</li><li><strong>The Operator's Move (tool-free):</strong> <em>Price out one week of your shop's AI — then cap it.</em> List ten weekly AI tasks, attach the unit price, multiply, compare to what you pay flat — then open every AI tool's billing page and put a number in the blank.</li></ul><h2>🔗 Show Notes &amp; Sources</h2><p><strong>Story 1 — GPT-6 Astra</strong></p><p><br></p><ul><li>OpenAI announcement: <a href=\"https://openai.com/index/gpt-6-astra/\" rel=\"noopener noreferrer\" target=\"_blank\">https://openai.com/index/gpt-6-astra/</a></li><li>9to5Mac — rollout details, \"included within existing subscription allowances\": <a href=\"https://9to5mac.com/2026/09/03/openai-releasing-major-upgrade-to-chatgpt-and-codex-with-gpt-6-astra-details-here/\" rel=\"noopener noreferrer\" target=\"_blank\">https://9to5mac.com/2026/09/03/openai-releasing-major-upgrade-to-chatgpt-and-codex-with-gpt-6-astra-details-here/</a></li><li>CNBC: <a href=\"https://www.cnbc.com/2026/09/03/open-ai-astra-gpt-6-cyber.html\" rel=\"noopener noreferrer\" target=\"_blank\">https://www.cnbc.com/2026/09/03/open-ai-astra-gpt-6-cyber.html</a></li><li>API pricing ($10/$50, 2.5x Sol): <a href=\"https://artificialanalysis.ai/models/gpt-6-astra\" rel=\"noopener noreferrer\" target=\"_blank\">https://artificialanalysis.ai/models/gpt-6-astra</a></li></ul><p><strong>Story 2 — The rate card</strong></p><p><br></p><ul><li>OpenAI ChatGPT Rate Card (Business/Enterprise): <a href=\"https://help.openai.com/en/articles/11481834-chatgpt-rate-card-business-enterpriseedu-credit-based-pricing\" rel=\"noopener noreferrer\" target=\"_blank\">https://help.openai.com/en/articles/11481834-chatgpt-rate-card-business-enterpriseedu-credit-based-pricing</a></li><li>Managing credits, auto-reload and spend controls: <a href=\"https://help.openai.com/en/articles/20001155\" rel=\"noopener noreferrer\" target=\"_blank\">https://help.openai.com/en/articles/20001155</a></li><li>Flexible pricing FAQ: <a href=\"https://help.openai.com/en/articles/11487671-flexible-pricing-for-the-enterprise-edu-and-business-plans\" rel=\"noopener noreferrer\" target=\"_blank\">https://help.openai.com/en/articles/11487671-flexible-pricing-for-the-enterprise-edu-and-business-plans</a></li><li>Premium seats announcement ($100 = 2,500 credits): <a href=\"https://openai.com/index/premium-seats-chatgpt-business/\" rel=\"noopener noreferrer\" target=\"_blank\">https://openai.com/index/premium-seats-chatgpt-business/</a></li><li>The Decoder on Premium seats: <a href=\"https://the-decoder.com/openai-introduces-125-premium-seats-for-chatgpt-business-as-agentic-ai-burns-through-more-tokens/\" rel=\"noopener noreferrer\" target=\"_blank\">https://the-decoder.com/openai-introduces-125-premium-seats-for-chatgpt-business-as-agentic-ai-burns-through-more-tokens/</a></li><li>CNBC — July 30 GPT-5.6 price cuts: <a href=\"https://www.cnbc.com/2026/07/30/open-ai-price-cut-gpt.html\" rel=\"noopener noreferrer\" target=\"_blank\">https://www.cnbc.com/2026/07/30/open-ai-price-cut-gpt.html</a></li></ul><p><br></p>","author_name":"Shaun Gehring"}