OpenAI just closed a $110 billion funding round at an $840 billion valuation. Amazon, NVIDIA, and SoftBank led the raise. That's not a typo. It's the largest amount of capital ever deployed into a single startup in one go.
To put that in perspective: OpenAI is now valued higher than most of the S&P 500. It's worth more than Coca-Cola, Pfizer, or Boeing. And it's still a private company that hasn't turned a profit yet.
Why this much capital
The reasoning from investors is straightforward: AI infrastructure requires scale that only a handful of companies can afford. Training frontier models costs hundreds of millions of dollars per run. Operating them at global scale costs billions annually. OpenAI is burning through cash faster than almost any company in history - and investors are betting that dominance in AI is worth the burn rate.
Amazon and NVIDIA aren't neutral investors here. Amazon provides cloud infrastructure through AWS. NVIDIA supplies the GPUs that train and run these models. This is vertical integration dressed up as venture capital. They're not just betting on OpenAI's success - they're ensuring their own infrastructure stays critical to it.
SoftBank's involvement is more interesting. After backing a string of failed or struggling bets in tech, they're going all-in on AI as the next platform shift. This is a bet that AI becomes as foundational as the internet - and that OpenAI becomes something like Google was in search.
What else is moving
While OpenAI dominated headlines, Anthropic quietly stood its ground with the Pentagon. The company refused to allow its Claude models to be used for certain military applications, citing ethical concerns around autonomous weapons systems. That's a significant stance - and it sets Anthropic apart in a market where most AI labs are racing to secure government contracts.
The same week, vision-language-action models emerged as a technical focus across robotics labs. These unified systems - which combine perception, language understanding, and motor control - represent a shift from hand-coded robot behaviours to learned, generalised ones. Google's RT-2 and NVIDIA's GR00T N1 are early examples, but the pattern is clear: robotics is becoming an AI problem, not just a mechanical one.
On the infrastructure side, inference optimisation is now a priority. As models grow larger, the cost of running them at scale becomes unsustainable without compression, quantisation, and better hardware. Companies like Groq and Cerebras are positioning themselves as the infrastructure layer for real-time AI - and their performance numbers are impressive enough that hyperscalers are paying attention.
The bigger pattern
This funding round isn't just about OpenAI. It's a signal that the AI market has entered a consolidation phase. The companies with access to capital, compute, and data are pulling ahead. Everyone else is either getting acquired or becoming niche players.
For developers and builders, this means two things. First, the cost of competing at the frontier is now beyond reach for most startups. You're not training GPT-5 in your garage. But second, the application layer is wide open. These models are tools - and the interesting work is in what people build with them, not the models themselves.
The infrastructure is being built by a handful of giants. The products are still up for grabs.
OpenAI's raise is historic - but it's also a reminder that scale matters in AI. Not just model scale, but financial and operational scale. The question now is whether that scale translates into products people actually need, or whether we're watching the most expensive research project in history.