DeepSeek targets $12 billion mega round led by Tencent and CATL for AI expansion

If the reported numbers hold, DeepSeek is about to close one of the largest private AI financings ever, and it is doing so explicitly to chase AGI scale. An 800 billion to 1 trillion yuan war chest backed by Tencent and CATL would give the lab the capital to both train next generation models and build its own power hungry, Huawei based compute clusters in Inner Mongolia. That is China constructing a homegrown alternative to the OpenAI plus NVIDIA plus US hyperscaler stack, with its own capital markets and energy system behind it.

Strategically, this deal would lock in alignment between a leading model lab, a dominant internet platform and the world’s biggest battery maker. For Tencent, it is a hedge against relying only on in house models. For CATL, it is a way to climb up the AI power stack from batteries into end to end “compute plus energy” systems. For DeepSeek, it is a chance to move from scrappy disruptor to national scale infrastructure provider. If it pulls this off and executes the planned gigawatt data center, the lab could sustain model training and inference at scales previously limited to US giants.

In the race to AGI, such a large round dramatically strengthens the non US frontier ecosystem. It makes it more likely that the next few frontier scale models will emerge from a genuinely multipolar set of actors, not just a handful of US labs.