Big Tech will spend nearly $700B on AI in 2026. No one knows where buildout ends
12 hours ago
- #Market Trends
- #Infrastructure
- #AI Investment
- SoftBank plans to list a new AI and robotics company in the US, targeting up to $100 billion valuation.
- Big Tech's AI capital expenditures are soaring, with combined spending of over $130 billion last quarter and projected to reach $700 billion this year.
- AI infrastructure spending is driven by high costs for GPUs, data centers, and networking, with McKinsey projecting $6.7 trillion in global AI capex by 2030.
- Market reactions are mixed; investors debate if AI buildout is justified, with concerns about overbuilding and fast-depreciating hardware.
- OpenAI explained that 'goblin' references in its models stem from unintended reinforcement learning rewards during training.
- Google's Gemini AI shows improvement in trip planning but still requires human oversight due to errors and contextual limitations.
- A Deloitte study reveals 75% of tech leaders see need for operational changes to maximize AI value, despite high confidence in deployment capabilities.