Timeline
Published analysis of 432k coding agent requests showing median 96k input tokens
Launched DeepSeek-R1 large language model, causing $1 trillion market valuation shift
SemiAnalysis conducted internal bake-off of AI tools, finding Perplexity Slack bot beats Claude
DeepSeek valuation jumps from $20B to $45B in first VC round led by China state fund.
Flagged 8x SRAM understatement on Cerebras website, a transparency issue for IPO due diligence.
Released V4-Pro and V4-Flash open-weight models with up to 1.6 trillion parameters
Released DeepSeek-V4 with 1M-token context at 10% KV cache cost
Initiates first external fundraising round, targeting at least $300M at a $10B+ valuation to set a benchmark for employee stock options.
Published analysis claiming NVIDIA's customer data is driving shift toward disaggregated inference architectures
Delays to DeepSeek-V4 model due to Huawei Ascend chip compatibility work