Data centers, chips, energy contracts, and cloud deals became the backbone of the AI race
At the start of 2025, the AI industry ran on bold optimism and even bolder spending. Investors poured in cash at record levels.
OpenAI raised $40 billion at a $300 billion valuation, while Thinking Machine Labs and Safe Superintelligence secured $2 billion seed rounds without releasing products.
Startups like Lovable and Mercor also attracted huge funding, showing how quickly AI valuations had inflated. xAI raised at least $10 billion after acquiring X.
Big money, bigger spending
Meta spent nearly $15 billion to secure top talent and accelerate data center growth. Alphabet increased its infrastructure push, including a $4.75 billion acquisition of Intersect, and plans to spend up to $93 billion on compute in 2026.
Across the industry, companies pledged nearly $1.3 trillion toward future AI infrastructure.
Data centers, advanced chips, long-term energy contracts, and cloud deals became the backbone of the AI arms race.
By midyear, however, the tone began to change. Optimism remained, but scrutiny grew. Investors started worrying about an AI bubble, limited enterprise adoption, and whether demand justified the scale of spending.
Infrastructure challenges, like rising power costs, grid limits, and delayed projects, added to the pressure.
New model releases from OpenAI and Google’s Gemini impressed on paper but felt incremental rather than transformative.
By the end of 2025, attention shifted from flashy breakthroughs to real businesses. Investors demanded proof of profits, safety, and sustainable growth. AI still promises to reshape the world, but now it must earn trust, deliver value, and justify its enormous price tag.