专题
返回时间轴
2025-01-20

DeepSeek releases R1, an open-source reasoning model that shocks global markets

Capability Breakthrough Commerce & Industry

事件摘要

DeepSeek R1: an open-source reasoning model matching OpenAI’s best at a fraction of the cost. How a Chinese AI lab shocked global markets in 2025.

影响评估

  • Economic Disruption -3 · Medium-term

    Triggered $593B Nvidia single-day market cap loss — the largest in US history. Global AI semiconductor selloff. Nasdaq fell 3%. Forced revaluation of $500B AI infrastructure buildout. Raised fundamental questions about the compute-centric AI investment thesis.

    Affected Groups: investors, semiconductor industry, AI companies, global markets

  • Paradigm Shift +3 · Long-term

    Proved open-source AI can match closed-source frontier reasoning at a fraction of the cost. Democratized reasoning capabilities globally. Rendered US chip export controls less effective. Established DeepSeek as a permanent force in AI. Triggered global policy and regulatory responses.

    Affected Groups: AI researchers, developers, policymakers, global AI community

共识度与来源

重要度 L2
分类 Capability Breakthrough / Commerce & Industry
共识度 Broad Consensus
影响指数 9/10

关联事件

← Origins
2024-07-23 L2

Meta releases Llama 3.1 405B, the first frontier-level open-source AI model

On July 23, 2024, Meta released Llama 3.1 405B — the first open-weight model to match closed-source frontier models like GPT-4 and Claude 3.5 Sonnet on key benchmarks. With 405 billion dense parameters and a 128K token context window, it proved that open-source AI could compete at the highest level. The release reshaped the AI industry by demonstrating that open models could match proprietary ones, accelerating the open-weight ecosystem that would define 2025.

2024-09-12 L2

OpenAI releases o1, the first reasoning model that 'thinks before it answers'

On September 12, 2024, OpenAI released o1 (codenamed 'Strawberry'), a large language model trained to spend more time 'thinking' before generating responses — using a chain-of-thought reasoning process internally. o1 scored in the 89th percentile on the International Mathematics Olympiad qualifier and reached PhD-level accuracy on physics, chemistry, and biology benchmarks. It demonstrated that scaling compute at inference time — not just training time — could unlock new capabilities, opening a second axis of AI progress.

→ Successors
2025-07-09 L1

Nvidia becomes the first company in history to reach $4 trillion market capitalization

In July 2025, Nvidia became the first company in history to reach a $4 trillion market capitalization, driven by insatiable demand for its AI training and inference chips. The valuation had quadrupled since Nvidia surpassed $1 trillion in 2023 and made up approximately 7.3% of the S&P 500. By October 2025, Nvidia reached $5 trillion. The milestone symbolized AI's transformation from a technology trend into the dominant economic force of the 2020s.

2026-04-24 L1

DeepSeek releases V4, an open-weight 1.6 trillion parameter model that undercuts all competitors on price

On April 24, 2026, DeepSeek released V4, its most powerful open-weight model family — with V4-Pro at 1.6 trillion total parameters (49B active) and V4-Flash at 284B (13B active), both under the MIT license with a 1-million token context window. Priced at $0.87 per million output tokens, V4 undercut GPT-5.5, Claude Opus 4.7, and Gemini 3.1 Pro by 5-10x. The release demonstrated that the open-weight ecosystem had permanently reshaped AI economics.

Local Lineage DeepSeek releases R1... 2025-01-20 · L2 Meta releases Llam... 2024-07-23 · L2 OpenAI releases o1... 2024-09-12 · L2 Nvidia becomes the... 2025-07-09 · L1 DeepSeek releases ... 2026-04-24 · L1