軟體公司的消長時速？

 GCIOSGF Newsletter:會訊：Merry Reading Time 悦讀匯

Host: WGP

Edit: OF

Speaker : NW

2026年2月初，全球科技股市場迎來一場劇烈震盪。一款AI插件的發布，讓傳統軟體產業陷入前所未有的恐慌，從華爾街到中環，從納斯達克到港交所，軟體股集體暴跌的場面令人震驚。

這場風暴的導火線來自美國AI新創公司Anthropic，它在2月3日推出了一款以企業法務為導向的智慧外掛工具。這款以Claude Cowork平台為基礎的產品，可自動化處理合約審查、保密協議分類、簡報撰寫等法律工作。消息一出，涉足法律資訊服務的荷蘭威科集團和英國勵訊集團當天股價重挫超過13%，隨後恐慌情緒迅速蔓延至整個軟體產業。

市場的反應之激烈超越了所有人的預料。 2月3日，追蹤美國軟體、金融數據和交易所股票的兩個標普類股指數，單日市值蒸發約3,000億美元。微軟跌2.87%，甲骨文跌3.37%、賽富時跌6.84%，財捷更是暴跌10.89%。華爾街的交易員甚至創造出一個新詞"SaaSpocalypse"（SaaS末日），來形容這場災難。

恐慌情緒很快傳導至亞洲市場，2月4日港股Wind SaaS指數下跌6.39%，板塊總市值蒸發近150億港元。金蝶國際單日跌幅達12.64%，中國軟體國際跌6.72%。 A股方面，深信服跌4.86%，漢得訊息跌7.41%，整個軟體板塊都籠罩在陰霾之中。

更令人不安的是，根據iShares ETF的追蹤數據，軟體股在過去七天內市值縮水約1兆美元。如果把時間軸拉長到1月14日，也就是Anthropic發表Claude Cowork平台的前後，美股SaaS指數和港股SaaS指數的區間跌幅已經分別達到21%和17%。

AI插件為何威協如此之大

要理解這場恐慌的深層原因，得先搞清楚SaaS的商業邏輯。 SaaS的全名是"軟體即服務"，它的核心獲利模式就是按席位、按月或按年收取訂閱費用。比如說，一家公司有100名員工使用某款CRM軟體，就得購買100個帳號，每年支付對應的訂閱費用。

這種模式聽起來很穩，投資人也喜歡這種"躺著收錢"的生意。但問題是，當AI可以讓一個人做十個人的活時，企業還需要購買那麼多席次嗎？如果AI代理能自動完成大部分標準化工作，那麼傳統SaaS公司的收入不就直接打了骨折？

德勤在2月3日發布的2026年TMT產業預測報告中明確指出，隨著AI智能體在SaaS市場的普及，基於訂閱和按席位授權的模式，很可能被按使用量和按成果計費的混合模式所取代。 Gartner更是預測，到2030年，至少40%的企業SaaS支出將轉向以使用量、代理或成果為基礎的定價模式。

這意味著什麼？意味著傳統SaaS公司如果不轉型，可能真的會被時代拋棄。像Sierra這樣的AI客服公司，已經顛覆了傳統按坐席訂閱收費的方式，改為按成功解決的對話數量計費，沒能解決的問題就不收錢。這種"依結果付費"的RaaS模式，正逐步瓦解SaaS的護城河。

市場反應是否過度恐慌

不過，真的到了"軟體已死"的地步嗎？倒也未必。有科技界人士在2月4日向媒體分析稱，市場反應可能過度恐慌了，SaaS相比AI，在中小企業市場仍有低門檻的優勢。

這個觀點不無道理。 AI雖然強大，但部署成本高、技術門檻高，中小企業很難獨立搭建和維護。相較之下，SaaS產品開箱即用，訂閱就能上手，這種便利性短期內AI還無法完全取代。特別是那些融合了行業最佳實踐的垂直類SaaS，競爭力還是在的。

中信證券在2月4日也發表點評指出，面對歷次針對科技浪潮的質疑，最佳也是唯一的回擊方式就是企業績效。本次AI浪潮亦不例外。換句話說，軟體公司如果能證明自己仍創造價值，市場終將回歸理性。

從長遠來看，未來的SaaS市場很可能會走向分化。那些只是簡單封裝標準化流程、沒有技術深度的老舊SaaS產品，確實可能被AI淘汰。但那些能夠與AI深度融合、提供個人化解決方案的"AI原生SaaS"，反而會迎來新的成長機會。畢竟，AI雖然能處理標準化任務，但企業的個人化需求、產業專屬實踐，仍需要軟體來承載和管理。

軟體公司的生死時速

眼下，留給傳統SaaS公司的時間窗口已經不多了。一方面，它們必須加速AI化轉型，把AI能力深度嵌入產品中，從"賣工具"變成"賣結果"。另一方面，它們也得重新思考商業模式，從單純的訂閱制轉向混合定價，甚至完全擁抱按效果付費。

德勤預測，到2026年透過SaaS應用使用智能體的規模將迅速成長，SaaS應用有望變得更加智慧、個人化和自適應。這既是挑戰，也是機會。那些能夠抓住轉型窗口期的公司，可能會在新一輪競爭中脫穎而出；而那些固守傳統模式的，恐怕真的會成為時代的犧牲品。

資產管理巨頭阿波羅的John Zito在投資者論壇上提出的"軟體是否已死"的質疑，或許不是在宣判軟體產業的死刑，而是在提醒所有從業者：變革已經到來，不進則退。

這場由AI插件引發的股市震盪，表面上看來是科技對商業模式的衝擊，但本質上卻是一次產業大洗牌的前奏。軟體的天沒有塌，但軟體公司必須學會在AI時代重新定義自己的價值。誰能率先完成轉型，誰能在下一個十年繼續領先。

In early February 2026, the global technology stock market experienced a dramatic upheaval. The release of an AI plugin plunged the traditional software industry into unprecedented panic, with software stocks plummeting across Wall Street, Central, Nasdaq, and the Hong Kong Stock Exchange – a truly shocking scene.

The catalyst for this storm came from the US AI startup Anthropic, which launched a corporate legal-oriented intelligent plugin on February 3rd. Based on the Claude Cowork platform, this product automates legal tasks such as contract review, confidentiality agreement classification, and briefing drafting. Upon the announcement, shares of Dutch company Wolverhampton Wanderers and British company Reed Information Services, both involved in legal information services, plummeted by more than 13% that day, and the panic quickly spread throughout the entire software industry.

The market reaction was far more intense than anyone anticipated. On February 3rd, the two S&P 500 sector indices tracking US software, financial data, and exchange-traded stocks lost approximately $300 billion in market capitalization in a single day. Microsoft fell 2.87%, Oracle fell 3.37%, Salesforce fell 6.84%, and Caijie plummeted 10.89%. Wall Street traders even coined a new term, "SaaSpocalypse," to describe the disaster.

The panic quickly spread to Asian markets. On February 4th, the Hong Kong-listed Wind SaaS Index fell 6.39%, wiping out nearly HK$15 billion in market capitalization. Kingdee International fell 12.64% in a single day, and China Software International fell 6.72%. In the A-share market, Sangfor Technologies fell 4.86%, and Hand Enterprise Solutions fell 7.41%, casting a shadow over the entire software sector.

Even more unsettling, according to iShares ETF tracking data, software stocks have lost approximately $1 trillion in market capitalization in the past seven days. If we extend the timeline to January 14th, around the time Anthropic announced its Claude Cowork platform, the US and Hong Kong SaaS indices had already fallen by 21% and 17% respectively.

Why are AI plugins such a significant threat?

To understand the underlying reasons for this panic, we must first understand the business logic of SaaS. SaaS stands for "Software as a Service," and its core profit model is charging subscription fees per account, monthly, or annually. For example, if a company has 100 employees using a certain CRM software, it needs to purchase 100 accounts and pay the corresponding subscription fee annually.

This model sounds very stable, and investors like this "passive income" business. But the question is, when AI allows one person to do the work of ten, do companies still need to purchase so many accounts? If AI agents can automatically complete most standardized tasks, wouldn't the revenue of traditional SaaS companies be drastically reduced?

Deloitte's 2026 TMT Industry Forecast Report, released on February 3rd, explicitly points out that with the increasing prevalence of AI agents in the SaaS market, subscription-based and per-seat licensing models are likely to be replaced by hybrid models based on usage and results. Gartner further predicts that by 2030, at least 40% of enterprise SaaS spending will shift to usage, agency, or results-based pricing models.

What does this mean? It means that traditional SaaS companies that don't transform may truly be left behind by the times. AI customer service companies like Sierra have already disrupted the traditional per-seat subscription model, switching to charging based on the number of successfully resolved conversations; unresolved issues are not charged. This "results-based" RaaS model is gradually eroding the SaaS moat.

Is the Market Reaction Overreacting?

However, has it really reached the point of "software is dead"? Not necessarily. A tech industry insider analyzed to the media on February 4th that the market reaction may be overreacting; compared to AI, SaaS still has a lower barrier to entry in the SME market.

This viewpoint is not without merit. While AI is powerful, its deployment costs and technical barriers are high, making it difficult for SMEs to independently build and maintain. In contrast, SaaS products are ready to use out of the box, and subscriptions allow for immediate adoption; this convenience cannot be completely replaced by AI in the short term. Especially vertical SaaS products that integrate industry best practices remain competitive.

CITIC Securities also commented on February 4th, pointing out that the best, and only, way to respond to past skepticism about technological waves is through corporate performance. This AI wave is no exception. In other words, if software companies can prove they still create value, the market will eventually return to rationality.

In the long run, the future SaaS market is likely to become more differentiated. Outdated SaaS products that simply encapsulate standardized processes and lack technical depth may indeed be eliminated by AI. However, "AI-native SaaS" that can deeply integrate with AI and provide personalized solutions will see new growth opportunities. After all, while AI can handle standardized tasks, the personalized needs of enterprises and industry-specific practices still require software to support and manage them.

The Race Against Time for Software Companies

Currently, the window of opportunity for traditional SaaS companies is closing. On the one hand, these companies must accelerate their AI transformation, deeply embedding AI capabilities into their products, shifting from "selling tools" to "selling results." On the other hand, they must also rethink their business models, moving from simple subscription models to hybrid pricing, or even fully embracing pay-per-performance.

Deloitte predicts that the use of intelligent agents through SaaS applications will grow rapidly by 2026, with SaaS applications expected to become more intelligent, personalized, and adaptive. This presents both challenges and opportunities. Companies that can seize the window of opportunity for transformation may stand out in the new round of competition; while those that cling to traditional models may truly become victims of the times.

John Zito of asset management giant Apollo's question at an investor forum, "Is software dead?", may not be a death sentence for the software industry, but rather a reminder to all practitioners: change has arrived, and stagnation means regression.

This stock market turmoil triggered by AI plugins, on the surface, appears to be a technological impact on business models, but in essence, it is the prelude to a major industry reshuffle. The sky hasn't fallen for software, but software companies must learn to redefine their value in the AI ​​era. Whoever completes the transformation first will continue to lead in the next decade.