News Heat Trend
DeepSeek Announces Full Open-Source Release of Internal Inference Framework DeepInfer
DeepSeek announced that it will fully open-source its internal inference framework DeepInfer, while simultaneously releasing the production-grade token encoder-decoder component deepseek-recipe, to help developers and enterprises optimize the throughput, latency, and deployment costs of large-scale model inference. DeepInfer integrates high-performance operators and a deeply optimized inference engine, supporting capabilities such as distributed inference, speculative decoding, SWA Bounded Replay, and structured output, while maintaining compatibility with NVIDIA and Huawei Ascend platforms. DeepSeek noted that it has already realized high-throughput, low-latency large-scale inference services for DeepSeek V4.1 Flash based on DeepInfer, and future inference code for new models along with subsequent optimizations will continue to be updated within this framework.
DeepSeek Self-Develops AI Inference Chip, Plans to Break Dependence on NVIDIA and Huawei
According to Reuters, Chinese AI startup DeepSeek is developing its own AI chips, three informed sources revealed. The chip is designed specifically for inference scenarios, rather than for model training. The project was launched approximately one year ago and remains in the early stages. The company has engaged with chip design, wafer foundry, and storage enterprises, and has quietly increased recruitment of chip design engineers without publicly posting job listings. If successfully developed, DeepSeek will reduce its reliance on Nvidia and Huawei Ascend chips, following the trend of global AI giants such as OpenAI and Anthropic developing their own hardware. Affected by U.S. export controls, DeepSeek previously shifted from Nvidia H800 to Huawei chips. This self-developed chip is regarded as a significant strategic transformation. Meanwhile, DeepSeek also plans to complete its first round of external financing, with a fundraising scale of approximately $7 billion, and a valuation reaching $52 billion to $59 billion.
Nvidia's market cap breaks $5 trillion again
as of the close of US stocks on April 24, Nvidia's stock price rose over 4%, pushing its market cap above $5 trillion. This marks the first time in nearly six months, since late October last year, that Nvidia's market cap has surpassed $5 trillion again.A report released by Stanford University last week mentioned the rapid adoption of AI. The report stated that AI is being adopted at an unprecedented rate, with the adoption rate of generative AI reaching 53%, faster than the adoption speed of personal computers and the internet. However, DeepSeek released its new model, DeepSeek-V4, which features a million-character ultra-long context and achieves leading performance in China and the open-source field in terms of agent capabilities, world knowledge, and reasoning ability. This model was developed in collaboration with Huawei Ascend. The cooperation and adaptation of cutting-edge models with domestic computing power providers are expected to promote the use of domestic computing power, which objectively also poses a challenge to Nvidia. (Yicai Global)
Related news
DeepSeek Announces Full Open-Source Release of Internal Inference Framework DeepInfer
DeepSeek announced that it will fully open-source its internal inference framework DeepInfer, while simultaneously releasing the production-grade token encoder-decoder component deepseek-recipe, to help developers and enterprises optimize the throughput, latency, and deployment costs of large-scale model inference. DeepInfer integrates high-performance operators and a deeply optimized inference engine, supporting capabilities such as distributed inference, speculative decoding, SWA Bounded Replay, and structured output, while maintaining compatibility with NVIDIA and Huawei Ascend platforms. DeepSeek noted that it has already realized high-throughput, low-latency large-scale inference services for DeepSeek V4.1 Flash based on DeepInfer, and future inference code for new models along with subsequent optimizations will continue to be updated within this framework.
TAC Launches GAP Framework to Enable DeFi Composability for Tokenized Securities
Odaily News: The Tokenized Asset Coalition (TAC) has launched the Global Access Protocol (GAP), designed to enable composability of tokenized securities within DeFi. TAC stated that GAP was developed over the course of a year by TAC and the Digital Securities Initiative, during which multiple discussions were held with the U.S. SEC, CFTC, and the Treasury Department. More details will be announced next week. TAC also announced the addition of 12 new members: Ascend, Control, Cryptio, Digital Securities Initiative, Enterprise Ethereum Alliance, Euler, Lava, OpenZeppelin, Overmatch Capital, Protocol Wealth, Reserve, and T-RIZE.
Huawei bids for Egypt AI data center project, US plans countermeasures
According to Bloomberg, Huawei Technologies Co., Ltd. has submitted a proposal to the Egyptian government to build an AI data center, planning to provide 1,408 Ascend 950 series chips for an AI training cloud, as well as 600 Ascend 950 series or earlier 910B chips for two inference clusters.
DeepSeek Self-Develops AI Inference Chip, Plans to Break Dependence on NVIDIA and Huawei
According to Reuters, Chinese AI startup DeepSeek is developing its own AI chips, three informed sources revealed. The chip is designed specifically for inference scenarios, rather than for model training. The project was launched approximately one year ago and remains in the early stages. The company has engaged with chip design, wafer foundry, and storage enterprises, and has quietly increased recruitment of chip design engineers without publicly posting job listings. If successfully developed, DeepSeek will reduce its reliance on Nvidia and Huawei Ascend chips, following the trend of global AI giants such as OpenAI and Anthropic developing their own hardware. Affected by U.S. export controls, DeepSeek previously shifted from Nvidia H800 to Huawei chips. This self-developed chip is regarded as a significant strategic transformation. Meanwhile, DeepSeek also plans to complete its first round of external financing, with a fundraising scale of approximately $7 billion, and a valuation reaching $52 billion to $59 billion.
Nvidia's market cap breaks $5 trillion again
as of the close of US stocks on April 24, Nvidia's stock price rose over 4%, pushing its market cap above $5 trillion. This marks the first time in nearly six months, since late October last year, that Nvidia's market cap has surpassed $5 trillion again.A report released by Stanford University last week mentioned the rapid adoption of AI. The report stated that AI is being adopted at an unprecedented rate, with the adoption rate of generative AI reaching 53%, faster than the adoption speed of personal computers and the internet. However, DeepSeek released its new model, DeepSeek-V4, which features a million-character ultra-long context and achieves leading performance in China and the open-source field in terms of agent capabilities, world knowledge, and reasoning ability. This model was developed in collaboration with Huawei Ascend. The cooperation and adaptation of cutting-edge models with domestic computing power providers are expected to promote the use of domestic computing power, which objectively also poses a challenge to Nvidia. (Yicai Global)