AI IPO Wave and Model Routing Threat
Key Questions
What new features did OpenAI launch with GPT-5.6?
OpenAI launched GPT-5.6 Sol/Terra/Luna on July 8, including the ChatGPT Work agent and confirmed it as the preferred model in Microsoft 365 Copilot.
What is the status of OpenAI's IPO plans?
OpenAI is considering delaying its IPO to 2027 and is in early talks to give a 5% stake to the US government.
How is Anthropic performing financially?
Anthropic has reached a $965 billion valuation with over $1 billion in quarterly profit.
What impact did Kimi K3 have on the AI market?
Moonshot AI's Kimi K3 launch, a 2.8T MoE model with open weights, triggered a tech sell-off and intensified pricing competition from Chinese models.
What is the model routing threat in AI?
Model routing allows cheaper alternatives like Grok 4.5 to match top models at a fraction of the cost, accelerating commoditization of the model layer.
Why was Google's Gemini 3.5 Pro delayed?
Gemini 3.5 Pro was delayed due to internal coding performance gaps that fell short of targets.
How is China influencing global AI competition?
China is considering banning its AI models globally while companies like Alibaba and DeepSeek release cheaper, competitive models, escalating decoupling.
What is Apple's lawsuit against OpenAI about?
Apple's lawsuit against OpenAI threatens to further delay OpenAI's IPO timeline amid ongoing regulatory and competitive pressures.
GPT-5.6 Sol/Terra/Luna launched July 8 with ChatGPT Work agent—major product release. GPT-5.6 Sol Ultra proved a 50-year-old math conjecture in under an hour. Sam Altman confirmed GPT-5.6 is now the preferred model in Microsoft 365 Copilot. OpenAI considering delaying IPO to 2027; in early talks to give 5% stake to US government. Anthropic at $965B valuation with >$1B quarterly profit. DeepSeek closes $7.4B funding; R2 never shipped, V4 with reasoning instead; open-sources DSpark for 85% inference speedup. Alibaba surges on cheaper AI models, undermining US dominance. Model routing threat persists: Grok 4.5 matches Fable 5 Max at 1/10th cost; muse spark 1.1 beats Opus 4.8 at 20% cost. Agentic AI dominates VC funding. Model layer commoditization accelerating. China considering banning its AI models globally—major decoupling escalation. Microsoft shifting to in-house models for Office 365 reduces demand for third-party AI. New: Google's Gemini 3.5 Pro delayed due to coding gap; Modal's DFlash speculator claims 67% higher throughput; Asia startup funding hit $42.8B in Q2 with DeepSeek $7.4B. Kimi K3 (Moonshot AI) launched with 2.8T MoE, 1M context, open weights July 27, beating Claude Fable 5 on coding at half price—intensifying pricing war and Chinese competition. New today: Kimi K3 launch triggered a broad tech sell-off, reinforcing Chinese AI competition and pricing pressure as a market-moving event. Nuance: Kimi K3 outperforms in frontend code but lags in complex math. Moonshot AI suspended new subscriptions due to demand. Apple's lawsuit against OpenAI threatens OpenAI's IPO timeline.