For accredited investors with private market access, SpaceX represents a generational infrastructure play as it scales toward a projected $100 billion in ARR by year-end.

Investors looking for liquid opportunities should watch Alibaba Group Holding Limited ($BABA**)** closely as its aggressively priced Qwen 3.8 Max model accelerates enterprise AI adoption.

The launch of the TeraFab joint venture by SpaceX and Tesla ($TSLA**)** signals a massive incoming disruption to traditional semiconductor manufacturers like TSMC ($TSM**)** and ASML Holding ($ASML**)** over the next few years.

With Alibaba ($BABA**)** undercutting Western AI competitors by up to 88% on cost, international open-weight models are creating a profitable new wave for cost-conscious enterprise software buyers.

SpaceX (Private / Pre-IPO)

• SpaceX expects to hit $100 billion in annual recurring revenue (ARR) by the end of the year.
• The company is aiming for $1 trillion in total revenue by 2030 (or possibly as early as 2029).
• Posted $7.8 billion in revenue for the quarter, representing a 92% year-over-year increase.
• Starlink is outperforming with 12 million subscribers (2x year-over-year) and revenues growing 66% to $4.3 billion.
• The company plans to reach 2 gigawatts of compute by the end of 2026 and 10 gigawatts by the end of 2027.
• SpaceX and Tesla are jointly investing $16.8 billion to build the TeraFab, a massive 100 million square foot facility intended for advanced manufacturing, which could fundamentally disrupt the semiconductor and lithography markets currently dominated by companies like TSMC and ASML.

Takeaways

• SpaceX is scaling at a historic rate, driven by the explosive growth of Starlink and upcoming infrastructure expansion like StarMind and the TeraFab.
• For investors with access to private markets, SpaceX is viewed by some analysts as a generational, civilizational infrastructure play.

Alibaba Group Holding Limited (BABA)

• Alibaba's Qwen team released Qwen 3.8 Max, a multimodal model featuring 2.4 trillion parameters, 95 billion active per request, and a one-million token context window.
• The model processes up to 750,000 words per query, handles 200-page documents, analyzes 100+ hour videos, and creates software applications from screenshots without source code.
• Priced at $2 per million input tokens and $6 per million output tokens, coming in 80% cheaper than GPT 5.6 Sol and 88% cheaper than Claude Fable 5.
• Following the announcement on the Hong Kong exchange, Alibaba's stock went up 7%.

Takeaways

• Open-weight models from international labs are applying heavy competitive pressure on Western frontier AI labs, driving down the cost curve for enterprise AI adoption.