Moonshot AI has reportedly closed an F-round financing worth more than $3.5 billion at a $35 billion post-money valuation, and has begun preparing an earlier-than-expected pre-IPO round that could value the Chinese artificial intelligence company at $50 billion before new capital is added.
People familiar with the transaction said the F round was oversubscribed by more than three times its initial target and closed ahead of schedule. The company’s rapid move toward another financing suggests demand for exposure to large Chinese foundation-model developers remains strong despite the high cost of training and operating advanced AI systems.
A $50 billion pre-money valuation would mark a sharp revision from market estimates a week earlier, which placed Moonshot’s prospective valuation nearer $31.5 billion before a final private round expected to begin in August. The reported change would place the company among the most highly valued privately held AI businesses globally, though its valuation will ultimately depend on the terms and size of the next financing.
Valuation has risen from $300 million in three years
Moonshot AI, founded in April 2023, has moved from an estimated $300 million valuation to $35 billion in a little over three years, according to the financing history provided by people familiar with the company’s fundraising. A pre-IPO round priced at a $50 billion pre-money valuation would represent an increase of more than 160 times from its early angel-round valuation.
The company raised more than $200 million roughly two months after it was established, the same sources said. Its A+ round in February 2024 raised more than $1 billion and valued Moonshot at about $2.5 billion. A B round around six months later brought in more than $300 million at a valuation of approximately $3.3 billion.
Fundraising accelerated through late 2025 and 2026. Moonshot reportedly raised $500 million in a C round at the end of 2025, taking its post-money valuation to $4.3 billion. Additional financings in the first two months of 2026 lifted that figure first to $10 billion and then to $18 billion.
In May, the company completed a roughly $2 billion D round at a $20 billion post-money valuation. China Mobile, Guozhitou, CPE Yuanfeng and several state-backed funds participated in that financing, according to the supplied account. The involvement of telecom and state-linked capital gives Moonshot access to backers with an interest in domestic AI infrastructure, enterprise deployment and strategic computing capacity.
The latest round began taking shape in June, when market estimates placed Moonshot’s pre-money value at about $31.5 billion. The reported $35 billion post-money result indicates the company attracted substantially more capital than originally anticipated, while its decision to pursue a pre-IPO round earlier than planned could reduce the time it needs to remain dependent on successive private raises.
Kimi K3 puts open weights at the center of the launch
Moonshot’s fundraising push coincided with the launch of Kimi K3, its latest large language model. The company introduced K3 on July 16 and released its full model weights, a technical report and portions of its infrastructure tooling on July 27.
K3 uses a mixture-of-experts architecture, a design that divides a model into specialized components and activates only part of the system for each request. Moonshot said K3 contains 2.8 trillion total parameters, with 104 billion activated during an inference, or a single model response. The model supports a context window of up to 1 million tokens and is designed for vision tasks, coding, reasoning and extended tool use.
The open-weights release gives companies and developers the ability to download the model, modify it, run it in their own environments and continue training it for specialized use cases. That can appeal to organizations that want greater control over data handling or need models tailored to internal software, documents and workflows.
The practical barrier is substantial. K3’s model files exceed 1.5 terabytes, creating high storage, memory and computing requirements for anyone seeking to load and deploy the full system. Open access does not eliminate infrastructure costs; it transfers more responsibility for hardware, model serving and optimization to the organization using it.
That distinction is relevant for cryptocurrency-linked computing projects that promote decentralized access to graphics processing units. Moonshot’s financing and K3 release demonstrate demand for AI capacity, but they do not establish demand for any particular token, decentralized computing protocol or hardware-rental network. A model’s hardware requirements can create opportunities for cloud providers and distributed-compute operators, yet revenue depends on actual workloads, pricing, uptime and the availability of compatible chips.
US debate focuses on model access and chip controls
The K3 weights release arrived during a renewed policy debate in the United States over downloadable AI models. On July 24, OpenAI, Google, Microsoft, Nvidia, AMD, Meta and Hugging Face publicly backed open-weight models and opposed rapid restrictions on systems that can be downloaded and deployed.
Anthropic and Amazon did not join that statement. On July 27, Anthropic chief executive Dario Amodei said access to open models without dangerous capabilities should remain available, while arguing for mandatory safety testing before sufficiently capable models are released, whether they are open or closed. Amodei also called for continuing controls on advanced chip exports and chipmaking equipment to China.
Those competing positions frame Moonshot’s K3 release as part of a larger contest over how advanced AI should be distributed. Open weights can speed experimentation and local deployment, while export controls and safety rules could shape which organizations can train, host or improve the largest models.
Moonshot’s reported $35 billion valuation rests on the expectation that it can turn advanced models and infrastructure into a durable commercial business. Its next financing will test whether private backers continue to support that expectation at a $50 billion valuation as model development becomes more capital-intensive and the policy environment around chips and open releases grows more restrictive.
Explore how AI is reshaping finance and trading in Web3, AI, and Crypto: Breaking the Internet for deeper market context.
Disclaimer: The content on this page is provided for general informational purposes only and does not represent the views or financial advice of Toobit. We make no guarantees regarding the accuracy or completeness of this information and shall not be held liable for any errors, omissions, or outcomes resulting from its use. Investing in digital assets involves risk; users should independently evaluate their financial situation and the risks involved. For further details, please consult our Terms of Service and Risk Disclosure.
