The AI Gold Rush Is Turning Into a Wall Street Bloodbath
THE SO WHAT
A roughly $500 billion drawdown in Alphabet’s value on AI spend fears is a warning shot—public markets will not underwrite indefinite negative cash flow for speculative model bets. Operators should expect a pivot from “grow at any GPU cost” to disciplined ROI narratives, with internal hurdle rates rising fast.
READ THE SOURCE
MORE FROM THE WIRE
Applied AIApple confirms its AI features reach China in December, with Qwen
Apple is localizing its AI stack for China by wiring Alibaba’s Qwen directly into Siri, Writing Tools, and Visual Intelligence—while holding back its newest Siri AI. If you operate in China, assume a bifurcated Apple Intelligence surface with different partners, capabilities, and compliance constraints than the rest of the world.
Applied AIDatabricks and SAP expand their deal to put AI on SAP data
Databricks is moving more of its AI stack—Genie, Lakeflow, Unity Gateway, Lakebase—directly onto SAP’s new Application Data Foundation. If you run SAP, the path of least resistance for AI on ERP data is consolidating around this pairing, so vendor-neutral architectures will take more deliberate effort.
Applied AIA human subject has been injected with an AI-designed virus aiming to kill antibiotic-resistant bacteria, so let's see how that turns out
An AI-designed virus moving into human trials is a line-crossing moment for bio-AI—therapeutic upside and dual-use risk now coexist in the same workflows. Any organization touching generative biology needs governance, audit trails, and red-team protocols that assume their outputs may eventually enter bodies, not just papers.
Applied AIOllie, a privacy-focused AI assistant, is shutting down
A privacy-first assistant closing shop while general assistants proliferate is a reminder that values differentiation doesn’t guarantee distribution or unit economics. If you’re building “ethical” or “privacy” AI, you still need a wedge on cost, capability, or channel, not just positioning.