跳到主要內容

DeepSeek漲價評測:API漲3倍、營收破10億美元 | DeepSeek Price Hike Review: Revenue Hits $1B

By Kit 小克 | AI Tool Observer | 2026-09-27

🇹🇼 DeepSeek漲價評測:API漲3倍、營收破10億美元

DeepSeek本週傳出一個很難忽略的數字:根據《The Information》報導,DeepSeek的年化營收(annualized revenue run rate)短短數月內翻倍,衝上10億美元,這個數字是執行長梁文鋒親口分享給投資人的。撐起這波成長的關鍵不是新產品,而是一次惹議的API漲價——漲幅落在2.3倍到4.5倍之間。

DeepSeek為什麼突然漲價?

簡單說:需求撐得住,公司就敢漲。DeepSeek去年靠「效能接近一線模型、價格只要零頭」打開市場,但訓練與推論成本從沒真的變便宜過。這次API漲價後,梁文鋒對投資人表示客戶需求沒有明顯縮手,公司依然自認是主流模型裡價格數一數二有競爭力的選項。

漲價前後差多少?

  • 漲價前:年化營收約在5億美元上下
  • 漲價後:年化營收翻倍衝上10億美元
  • 漲幅區間:2.3倍到4.5倍(依模型與用量而定)

漲價後,DeepSeek對開發者還划算嗎?

老實說沒有統一答案,要看你原本用在哪個量級。如果你是拿DeepSeek當Claude、GPT或Gemini的平價替代品,漲價後的價差可能已經沒有一年前那麼誇張,值得重新拉一次報價單比較,而不是憑印象繼續用下去。官方說法是「仍然相對便宜」,但這種話每家都會講,實際數字要自己查、自己算才準。

這波營收消息背後在盤算什麼?

DeepSeek同時在敲定第二輪外部募資,目標約75億美元(人民幣500億),估值上看750億美元,預計10月底前完成,並準備登陸上海證交所。營收數字選在這個時間點「分享給投資人」,很難不讓人聯想是替募資與上市故事加分。對用戶來說,這代表DeepSeek短期內漲價的誘因只會更強,不會更弱。

好不好用,試了才知道。


🇺🇸 DeepSeek Price Hike Review: Revenue Hits $1B

DeepSeek just posted a number that's hard to ignore: according to The Information, its annualized revenue run rate more than doubled in a few months to hit $1 billion, a figure CEO Liang Wenfeng shared directly with investors. The growth wasn't driven by a new product — it came from a controversial API price hike of 2.3x to 4.5x.

Why did DeepSeek raise prices?

Short answer: demand held, so the company pushed. DeepSeek won market share last year by offering near-frontier performance at a fraction of the cost, but training and inference costs never actually got cheaper. After the API price hike, Liang told investors that customer demand hasn't meaningfully dropped, and the company still considers itself one of the more competitively priced options among major models.

How much did prices actually change?

  • Before: annualized revenue run rate was around $500 million
  • After: run rate doubled to $1 billion
  • Price increase range: 2.3x to 4.5x, depending on model and usage tier

Is DeepSeek still worth it for developers?

Honestly, there's no single answer — it depends on what tier you were on before. If you were using DeepSeek as a budget alternative to Claude, GPT, or Gemini, the price gap may not be as dramatic as it was a year ago, so it's worth re-running your own cost comparison instead of assuming the old numbers still hold. The company says it's "still relatively cheap," but every vendor says that — check the actual current pricing yourself.

What's really behind this revenue news?

DeepSeek is simultaneously finalizing a second funding round targeting roughly $7.5 billion at a $75 billion valuation, aiming to close by end of October, ahead of a planned Shanghai Stock Exchange listing. The timing of sharing this revenue figure with investors right now is hard to separate from the fundraising and IPO narrative. For users, that means the incentive to keep pushing prices up is only getting stronger, not weaker.

好不好用,試了才知道 — you won't know until you try it.

Sources / 資料來源

常見問題 FAQ

DeepSeek為什麼漲價? Why did DeepSeek raise prices?

因為需求沒有因漲價而萎縮,DeepSeek在2026年將API價格調漲2.3倍到4.5倍,同時年化營收衝上10億美元。Demand held up despite a 2.3x-4.5x API price hike, helping DeepSeek's annualized revenue run rate double to $1 billion.

DeepSeek現在還比Claude、GPT便宜嗎? Is DeepSeek still cheaper than Claude or GPT?

漲價後價差確實縮小,但DeepSeek仍自稱是主流模型裡價格最有競爭力的選項之一,建議依自己的用量重新比價。The gap narrowed after the hike, but DeepSeek still claims to be among the most competitively priced major models — recheck pricing for your own usage tier.

DeepSeek為什麼現在公布營收數字? Why is DeepSeek sharing revenue numbers now?

DeepSeek正在敲定約75億美元的新一輪募資,並準備在上海證交所掛牌,亮眼營收數字有助於募資與上市敘事。DeepSeek is finalizing a roughly $7.5B funding round and preparing for a Shanghai IPO, so strong revenue numbers help the fundraising and listing story.

這對一般開發者有什麼影響? What does this mean for developers?

如果你是把DeepSeek當低價替代方案在用,現在該重新試算成本,而不是沿用一年前的價格印象做決策。If you use DeepSeek as a cheap alternative, re-run your cost math now instead of relying on last year's pricing assumptions.

延伸閱讀 / Related Articles


AI 工具觀察站 — 每日精選 AI Agent 與工具趨勢
AI Tool Observer — Daily curated AI Agent & tool trends

留言

這個網誌中的熱門文章

Google Ironwood TPU v7 推理專用晶片解析:效能追平 NVIDIA、成本低 44%,AI 晶片戰爭正式開打 | Google Ironwood TPU v7 Explained: Matching NVIDIA Performance at 44% Lower Cost — The AI Chip War Heats Up

Claude Code 實測:AI 幫你寫程式到底行不行? | Claude Code Review: Can AI Really Code for You?

Cursor vs GitHub Copilot vs Claude Code:AI 程式助手大比拼 | AI Coding Assistants Compared: Cursor vs GitHub Copilot vs Claude Code