Hyperdrive AI economy, the $200B Anthropic-Google deal, and the AI 'vibe shift' on jobs

Big Tech earnings reveal an AI economy operating at maximum capacity. Anthropic's deal with Google is bigger than anyone realized. And the dominant 'job apocalypse' narrative is getting a serious challenge.

The financial scale of the AI economy is now genuinely unprecedented, but the more interesting story this week is that the ‘AI will destroy all the jobs’ narrative is starting to fracture.

Google Cloud grows 63% YoY, $460B backlog

My take: Driven entirely by AI demand. AWS grew 28%, Azure 39%. The combined hyperscaler capex forecast for 2026 is over $800B, and customer backlogs are growing even faster. BlackRock CEO Larry Fink's quote captures it: 'There is not an AI bubble. There is the opposite. We're short power. We're short compute. We're short chips.' The implication: build AI into your roadmap as if compute scarcity persists for years.

Anthropic's $200B deal with Google Cloud over five years

My take: New details on the size of the deal, which alone accounts for a major chunk of Google Cloud's backlog. The capital concentration in frontier model training is reaching levels that would have seemed absurd two years ago. Practically: don't expect significant pricing competition between the top labs. They're each spending too much to discount.

OpenAI and Anthropic both launch consulting arms

My take: Multi-billion-dollar ventures to help enterprises actually deploy AI. Both companies acknowledged the biggest blocker to adoption isn't the tech, it's the integration. If you're a consultancy, expect serious new competition. If you're a buyer, evaluate these in-house options carefully. The labs know their products better than anyone.

The 'AI vibe shift' on jobs

My take: Despite the job-apocalypse narrative, demand for software engineers (the most AI-exposed profession) is accelerating. Commentators are invoking Jevons Paradox: AI efficiency creates more demand, not less. The narrative is shifting. If you've been holding off on hiring engineers because 'AI will replace them anyway,' you're going to lose the talent race.

Agents make every job a startup

My take: The new dynamic: instead of saving time, people work more intensely managing their 'infinite backlog' of AI-enabled tasks. Skills like judgment, planning, and coordination become more critical. There's even a new term for it: 'judgment burnout.' If you manage knowledge workers, this is the workload pattern to watch for and design against.

GPT 5.5 Instant becomes the default for free ChatGPT users

My take: Significant quality jump for free users. The free experience is now comparable to frontier models from late 2025. The implication: when your customers and employees encounter AI for the first time, the bar is now much higher. Any AI-powered product you ship needs to clear that benchmark.