Alphabet Stock Sheds $700B as AI Bills Climb
Alphabet's stock has dropped more than 15% from its May peak, wiping out roughly $700 billion in market value. The decline comes as investors grow concerned over mounting AI infrastructure bills and the company's failure to deliver a decisive technological breakthrough in the AI race.
Key Details
- Stock decline: Down 15%+ from May peak, losing ~$700B in market cap.
- AI costs: Rising expenses for AI compute and data centers are pressuring margins.
- Talent exodus: Top scientist departed, and its 'brightest AI mind' moved to a different position.
- Gemini delays: The newest Gemini model has been delayed, missing expected release timelines.
- Leadership reshuffle: Co-founder Sergey Brin has returned to day-to-day operations to help steer the company.
Context
Alphabet had been leading the Magnificent 7 stocks over the last 16 months, but now faces a 'shakier footing' according to analysts. However, a Wall Street Journal columnist argues that Google retains a 'superlative product suite,' a 'deep bench,' and a 'long history of investing in AI.' The reshuffle, including Brin's return, could be exactly what the company needs to refocus.
For developers and AI engineers, this signals potential instability in the Google AI ecosystem, but also highlights the intense cost pressures facing all major AI providers.
📖 Read the full source: HN AI Agents
👀 See Also

2,000 Hours with Claude Code: The Real Shift Is from Coding to Judgment
A developer spent 2,000 hours directing Claude Code since January. The surprise: the better the agent gets at execution, the more the human role collapses into judgment — defining problems, verifying results, and stopping wrong paths.

Claude Code v2.1.187: Structured Output Fixes, Sandbox Security, and Org Model Restrictions
Claude Code v2.1.187 adds sandbox.credentials setting, org model restrictions, and fixes for structured output loops, remote MCP hangs, and subagent depth tracking.

Claude.ai, API, and Claude Code Experiencing Elevated Errors
Claude.ai, the Claude API, and Claude Code are experiencing elevated errors with the web interface and developer console down. Claude Code login via Claude.ai is broken, though logged-in users can still use it.

Opus 4.7 Token Efficiency: German Prompts Burn Up to 2x Tokens vs English
A Claude Pro subscriber reports that using German with Opus 4.7 consumed 100% of session tokens in seconds, while English used 37%. The tokenizer inefficiency stems from compound nouns and umlauts, causing 1.5–2x token usage.