AI News June 11 2026 - SpaceX Prices in Hours, Microsoft Chips for Claude, Fable 5 Proving Itself

1. SpaceX Prices Tonight - SPCX Trades Tomorrow

SpaceX prices its IPO tonight after market close at a fixed $135/share - $1.77 trillion valuation, $75 billion raise, largest IPO in history. The deal is reportedly twice oversubscribed with over $10 billion in institutional orders verified. MSCI announced June 9 it will apply its usual early inclusion treatment for large-sized IPOs, meaning every passive fund tracking MSCI indexes will need to buy SPCX after listing - structural demand largely independent of fundamental valuation. The 366-day insider lockup (longer than standard) limits initial supply, which can amplify day-one price moves in either direction.

For retail investors: today is the last day to ensure your interest is registered via Robinhood, Fidelity, or Schwab. Allocation is not guaranteed. Morningstar's formal valuation is $780 billion - about 44% of the IPO price. The AI infrastructure argument ($2.17B/month from Anthropic and Google compute contracts at Colossus) is the primary bull case that justifies the premium over fundamental value. Both contracts are terminable with 90 days' notice after December 31, 2026 - disclosed risk.

Full breakdown: MSCI inclusion impact, three day-one trading scenarios, and the numbers behind the valuation ->

2. Anthropic-Microsoft Maia 200 - What the Chip Deal Would Actually Change

CNBC and Reuters confirmed Anthropic and Microsoft are in early-stage talks for Anthropic to run Claude inference workloads on Microsoft's Maia 200 custom AI accelerator chips via Azure. No deal has closed. The Maia 200 - launched January 2026 on TSMC's 3nm process with 216GB HBM3e memory and 10+ petaflops FP4 performance - claims 30%+ better inference performance per dollar than competing silicon. It is currently running OpenAI's GPT-5.2 internally. Getting Claude on Maia 200 would make it the first externally validated frontier model on Microsoft's homegrown chip.

For Anthropic: a fourth custom silicon option (alongside Nvidia GPUs, AWS Trainium, and Google TPUs) that could reduce per-token inference costs at scale - crucial heading into the October 2026 IPO when unit economics will be scrutinized by public market investors. For Microsoft: validation of Maia 200 as a commercially viable chip for external customers, and a stronger Azure-specific Claude story for enterprise sales. Risks: FP8 quantization may reduce output quality by 0.5-1.5% on some tasks; no deal yet after three weeks of talks.

Full breakdown: What Maia 200 is, why both sides want the deal, and the risks that could kill it ->

3. Claude Fable 5 - Two Days In, What Developers Are Reporting

Claude Fable 5 launched June 9 and has now been in developer hands for two days. The 80.3% SWE-Bench Pro benchmark is holding up in real-world coding tasks - developers report significant improvements on hard multi-file refactors, autonomous debugging sessions, and complex codebase analysis compared to Opus 4.8. The extended autonomous work capability (Fable 5 can work longer than any prior Claude model) is showing in longer uninterrupted coding sessions before needing human input.

The safeguard routing (cybersecurity and biology queries going to Opus 4.8) is triggering in under 5% of sessions as Anthropic predicted. API developers should handle stop_reason: "refusal" gracefully - it returns HTTP 200 and is not billed if no output was generated. Remember: Fable 5 is free on Pro, Max, Team, and Enterprise plans until June 22 only. After June 23 it requires usage credits. Model string: claude-fable-5. Full launch details: Claude Fable 5 launch guide - pricing, benchmarks, safety architecture, and the June 22 deadline ->