Uber's Monthly Anthropic Bill Reaches $500 Million, Entrepreneurs Shift to Open Source Models
Uber's monthly Anthropic bill has reached $500 million, with entrepreneur Ben Cera's personal bill exceeding $1 million.
Cera has turned to renting GPUs to use open source models, reducing costs by 100 times.
He believes the future of AI belongs to open source.
In market mechanisms, cost-sensitive enterprises and developers are becoming the main buyers shifting to open source. This event-driven funding flow is directed towards GPU leasing and open source model infrastructure, benefiting open source AI projects and hardware providers, while closed-source API providers face pressure.
Source: Public Information
ABAB AI Insight
Uber previously adopted AI on a large scale to enhance operational efficiency, and such high bills reflect the cost pressures of the commercialization path of closed-source models. Earlier, many companies similarly transitioned from APIs to self-built or open source solutions, demonstrating cost-driven optimization of their technology stacks.
In terms of capital pathways, companies are shifting budgets from closed-source subscriptions to GPU leasing and open source deployments, with strategic motives to control long-term expenditures and gain customization flexibility, reallocating resources from API vendors to hardware suppliers and the open source community.
Similar to other high-consumption companies transitioning to open source cases like Llama or Mistral, AI applications are currently in an expansion phase of shifting from closed-source dependency to open source autonomy, further accelerating the adoption of open source models at the enterprise level.
Essentially, this represents a technological substitution, where open source models and proprietary computing replace high-priced closed-source APIs. The mechanism is based on near-zero marginal costs and customization advantages, leading to a shift in pricing power from a few API giants to a distributed open source ecosystem, and driving the AI industry chain towards decentralized deployment reconstruction.
ABAB News · Cognitive Law
AI Cost = Closed-source Pricing × Usage Scale × Open Source Substitution
Large companies sell APIs, small businesses rent GPUs; whoever controls computing power defines the future stack.
The higher the bill, the faster the transition; the counterintuitive aspect is that high prices accelerate the victory of open source.