本研究以 Ohlson (1995) 評價模型為基礎，探討台灣 AI 供應鏈指標企業在 2020 年至 2025 年間之價值攸關性與評價位移現象。針對 AI 產業兼具高資本支出與高研發比重之財務特質，本研究在傳統財務報表指標（每股帳面淨值與每股盈餘）之外，進一步擴充納入自由現金流 (Free Cash Flow， 以下簡稱 FCF) 與研發費用 (Research and Development， 以下簡稱 R&D) 兩項變數，旨在檢驗擴充模型能否有效捕捉企業之真實經濟價值，並提升對股價之解釋力。 實證結果顯示，於 AI 實質成長期（2023–2025 年），市場評價基準已由歷史資產規模轉向實質獲利動能。其中，每股淨值 (BV) 之評價係數由 1.0 轉為不顯著之負值 (-1.4)，顯示傳統資產已失去定價權；迴歸模型之截距項則由轉型期之 -23.4 顯著上升至 125.2，反映資本市場對 AI 產業賦予龐大之基礎估值溢價。同時，盈餘評價乘數由 11.1 倍大幅躍升至 24.9 倍，然傳統財報指標之整體解釋力卻呈現遞減（Adjusted R² 降至 80.3%）。 然而，在納入 FCF 與 R&D 後，擴充模型之整體解釋力顯著提升至 84.1%，創造了 3.8% 之增額解釋力。進一步實證指出，資本市場對龐大之研發投入給予短期風險折價，反映對其回收期不確定性之考量；反之，對具備穩健自由現金流之企業則給予顯著之評價溢價。本研究證實，在 AI 產業面臨龐大資本擴張之際，FCF 與 R&D 具備顯著之增額解釋力，乃評估企業獲利品質與財務防禦力之核心導航指標。
關鍵詞： 評價位移、價值攸關性、Ohlson 評價模型、自由現金流 (FCF)、人工智慧 (AI) 供應鏈、研發費用 (R&D);Based on the Ohlson (1995) valuation model， this study investigates the value relevance and valuation shift of leading firms in Taiwan′s AI supply chain from 2020 to 2025. Given the financial characteristics of high capital expenditure and high R&D intensity in the AI industry， this paper incorporates Free Cash Flow (FCF) and Research and Development (R&D) expenses into the traditional financial metrics (Book Value and Earnings Per Share). The objective is to examine whether the expanded model effectively captures the true economic value of these firms and enhances the explanatory power for stock prices. Empirical results indicate that during the AI growth phase (2023–2025)， the market′s valuation benchmark shifted from historical asset scale to actual earnings momentum. The valuation coefficient of Book Value (BV) turned from 1.0 to an insignificant negative value (-1.4)， indicating that traditional assets have lost their pricing power. Meanwhile， the regression model′s intercept surged from -23.4 in the transition phase to 125.2， reflecting a massive baseline valuation premium assigned to the AI industry. Concurrently， the earnings multiple leaped from 11.1x to 24.9x， yet the overall explanatory power of traditional financial metrics declined (Adjusted R² dropped to 80.3%). However， after incorporating FCF and R&D， the expanded model′s explanatory power significantly increased to 84.1%， generating an incremental explanatory power of 3.8%. Further empirical evidence reveals that the market applies a short-term risk discount to massive R&D investments， reflecting concerns over the uncertainty of payback periods; whereas firms with robust FCF receive a significant valuation premium. This study confirms that amid massive capital expansion in the AI sector， FCF and R&D possess significant incremental explanatory power， serving as core indicators for evaluating earnings quality and financial resilience.
Keywords: Valuation Shift， Value Relevance， Ohlson Model， Free Cash Flow (FCF)， Artificial Intelligence (AI) Supply Chain， Research and Development (R&D)
