MAS 演講稿 · 2026-07-28
MAS 總裁謝德俊先生在MAS 2025/2026年度報告媒體釋出會上的講話（2026年7月28日）
要點
- • 新加坡經濟在2026年上半年錄得6%的同比增長（較2025年下半年的5%上升），由人工智慧相關行業增長抵消中東衝突對能源密集型產業的影響。
- • 新加坡金融管理局預計核心及綜合通脹率將在2026年平均為1.5%-2.5%，在2027年上半年保持高位，下半年明顯下降，主要由燃料及進口商品價格上升推動。
- • 新加坡金融管理局在2026年初保持升值的貨幣政策立場，並在4月份中東衝突後收緊政策以緩解進口價格上升帶來的通脹壓力。
- • 人工智慧驅動的電子產品出口佔亞洲今年迄今出口增長的70%以上（較2024年的46%上升），而與人工智慧相關的企業佔標普500指數市值約40%，佔MSCI新興市場亞洲指數30%以上。
- • 新加坡金融部門2025年增長4.3%，其中銀行資產增長3.1%，保險資產增長7.6%至4,935億新元，資產管理規模增長10.1%至6.7萬億新元。
- • 新加坡金融管理局宣佈成立「未來金融研究所」，通過「創新工作室」共同創造金融用例及「PathFin.ai」平臺訪問經過驗證的解決方案來推動金融部門創新。
完整譯文（繁體中文）
MAS 英文原文譯文 · 翻譯日期: 2026-09-06
1. 下午好，感謝各位今天參加我們釋出2025/2026財年MAS年度報告。
2. 我將介紹央行業務、金融部門發展與監管方面的最新進展，以及我們財務表現的主要亮點。最近經濟發展 3. 讓我首先談及最近的經濟發展和貨幣政策。
4. 在過去的一年裡，宏觀經濟形勢受到了連續衝擊的塑造。在去年新聞釋出會召開時，全球前景籠罩在以下擔憂中：關稅上升和貿易政策不確定性將對增長造成重大影響。隨後，2026年2月末中東衝突的爆發給全球經濟帶來了新的衝擊。
5. 面對這些反覆的衝擊，全球經濟證明了比預期更強的韌性。更高關稅的影響被供應鏈重新配置所緩衝。全球貿易和工業生產繼續擴張，儘管關稅率仍保持在較高水平。自3月以來能源供應中斷的規模很大，但被庫存以及敏捷的供應和需求調整所緩衝。能源市場實現了重新平衡，價格處於預期情景的較低範圍，限制了對增長的拖累，但提升了總體通脹的路徑。然而，能源價格仍然保持在較高水平，在全球庫存較低的背景下，中東衝突再次升級帶來了上行風險。
6. 全球經濟的增長和韌性受到了強勁全球AI投資的提振。在資料中心、晶片、計算基礎設施和半導體產能方面投資的激增支撐了全球電子產品生產和貿易的持續擴張。這已經使增長前景向上轉變，特別是對於深度融入全球技術供應鏈的經濟體。
7. 新加坡受到了所有這些逆風的影響。中東衝突對新加坡經濟的某些部分產生了明顯影響。能源相關部門，如化工製造業，在第二季度錄得兩位數的收縮。然而，這些部門帶來的拖累被技術相關部門的激增所超過。新加坡經濟在2026年上半年錄得強勁的6%同比增長，相比2025年下半年的5%增長有所提高。
8. 展望未來，新加坡經濟的增長在今年餘下時間應保持堅實。雖然中東局勢仍對前景構成風險，但全球AI相關需求可能繼續提供有意義的提振。大多數非AI相關部門可能保持接近趨勢的增長步伐。
9. 通脹從較低基數反彈，在未來一段時間內將進一步上升，然後在2027年下半年開始緩解。2025年國內總體和核心通脹保持在低於1.0%的低水平。在2026年第一季度，核心通脹為1.4%，在第二季度為1.5%。我們預期從7月開始這將進一步上升，並在未來幾個季度保持高位，由更高的燃料和進口商品價格驅動，抵消緩解的國內成本壓力和一些政府補貼的抑制效應。MAS核心通脹和CPI總項通脹預計在2026年平均為1.5%-2.5%，在2027年上半年保持高位，並在2027年下半年明顯緩解。
10. MAS的貨幣政策立場在年初也是良好定位的。我們在去年下半年維持升值立場，今年1月再次如此。這有助於緩解今年通脹的上升。在4月，隨著中東衝突的爆發，我們收緊了貨幣政策，以應對預期的進口通脹壓力上升。自那以來，新加坡的一系列能源和其他商品投入品的進口價格大幅上升，並開始向國內消費者價格傳導。因此，這個4月的舉措在我們進入更強通脹環境時，將貨幣政策設定置於有利位置。
11. 最近的7月MPS是一個精心設計的政策調整，建立在4月政策決定的基礎上。通脹預計在7月進一步上升，並在未來幾個季度保持高位，然後在2027年下半年緩解。正產出缺口現在預計將略微擴大，而不是如4月政策評估中所預想的那樣縮小。綜合考慮4月和7月的決定，新加坡元更強的升值將更有效地抵禦即將到來的通脹壓力。
12. 在高度不確定的經濟環境中，MAS對前景的風險保持警惕，並處於良好定位以應對，以維持中期價格穩定並控制新加坡元名義有效匯率(S$NEER)的過度波動。金融市場和金融穩定 13. 我現在轉向金融市場的發展和金融穩定的風險。
14. 迄今為止全球金融條件是良好的並支援經濟活動。儘管最近有所回撥，股票市場估值仍然很高，信用利差仍然偏緊。
15. 對迄今為止良好態勢的一個重大不確定性是AI投資熱潮的可持續性。全球增長、投資和金融市場表現已經高度依賴於大規模且不斷增加的資料中心和半導體晶片投資持續到遠期的預期。這在美國和半導體出口亞洲經濟體中尤為如此。
a) AI驅動的電子產品出口佔亞洲年初至今出口增長的70%以上，相比2024年的46%有所上升。
b) AI相關公司目前約佔S&P 500市場資本化的40%和MSCI新興市場亞洲指數的30%以上。它們也主導了在美國資本市場籌集的新融資，代表約50%的投資級債券發行、38%的高收益債券發行和87%的新風險投資資金。
16. 因此，AI投資的可持續性對全球增長和金融穩定具有高度重要性。雖然近期投資由承諾訂單和強勁的超大規模企業現金流支援，但圍繞這些投資在中期的可持續性存在更大的不確定性。在追求模型優勢和擴大采用的競爭中，超大規模企業和模型建設者的預期投資已超出現金流和商業收入。在未來數年內將需要大規模的股權和債務融資。
17. 市場將越來越多地看向商業收入增長來證明融資風險的合理性。收入增長將取決於AI生產率收益在企業層面的早期跡象擴大到整個經濟，以及變革性應用的深化。如果由加速的收入增長和不斷擴大的生產率收益支援，投資熱潮可能是一個延長的。但在AI投資貨幣化的路徑上也存在明顯的風險。這些包括能源和晶片成本的上升、原材料供應瓶頸、監管不確定性、模型提供商之間的激烈競爭（包括來自低成本開源模型的競爭），以及生產率收益的分享範圍有多廣。如果AI投資的回報在中期未能達到預期，超大規模企業將放緩投資步伐，市場將重新評估資產估值。
18. 如果任何一種結果發生，其含義都是重大的。
19. 如果我們處於一個長期的AI投資熱潮中，且伴隨顯著和廣泛的生產率收益，對收入、需求和通脹的更強和更廣泛的溢位效應可能隨之而來。對通脹的影響將是複雜的，取決於對能源和投入品的更高需求與生產率收益的步伐和程度之間的相互作用。這些將對央行對潛在產出和中性利率的評估產生後果。
20. 另一方面，如果AI投資出現重大削減，它可能通過商業投資和半導體需求下降以及負面財富效應而急劇削弱全球增長。金融穩定風險也可能通過股票、信貸和貸款市場對具有惡化現金流和複雜融資結構中薄弱信用條款的不可持續商業模式的敞口而顯現。全球金融條件可能出現急劇收緊。
21. 第二個風險來源是中東衝突的持久再升級。雖然迄今為止全球石油市場通過更大的供應和需求敏捷性以及比預期更低的價格影響做出了回應，但中東曠日持久的更具破壞性的衝突可能在庫存較低的背景下重新引發商品價格波動。雖然這不是我們的基本假設，但我們不能排除石油價格大幅上升和石油及下游產品短缺惡化的風險。
22. MAS強調了在全球壓力和不確定性下評估國內金融脆弱性。我們更新了壓力測試，以捕捉關鍵的下行風險，包括中東衝突的再次升級和金融條件的急劇收緊。
23. 雖然結果證實國內金融系統保持大體韌性，但在高度槓桿化的家庭和企業中存在小範圍的脆弱性。
a) 金融條件收緊帶來的風險對能源密集型部門的企業以及相對支出而言金融緩衝較薄的家庭更為突出。
b) 隨著全球不確定性保持在較高水平，暴露程度更高的部門的企業應為更多流動性做好準備，並採取措施多元化收入來源，而收入不夠穩定且緩衝較薄的家庭應謹慎對待承擔大額新貸款承諾。
金融服務部門發展 24. 我現在轉向金融部門的發展。
25. 儘管全球環境更加不確定，金融部門在2025年仍實現了4.3%的健康增長。
a) 2025年的增長基本保持了2021-2025年期間的勢頭，該期間年均增長為4.6%，每年增加4200個工作崗位（不包括控股公司的工作崗位），就業收益流向本地人士。
26. 增長繼續保持廣泛基礎。
a) 銀行部門資產在2025年增長了穩健的3.1%，相比2024年的異常增長有所放緩。
b) 保險業資產在2025年增長7.6%，達到493.5億新元，較2024年加快增速，與2021年至2025年7.5%的複合年增長率一致。
c) 截至2025年底，管理資產增長10.1%，達到6.7萬億新元。財富管理行業也在更廣泛的資產管理行業的帶動下繼續強勁增長。
d) 作為亞洲領先的外匯交易中心，外匯平均日交易量在2025年繼續擴大，達到1.6萬億新元。
e) 新加坡也在繼續發展成為該地區領先的固定收益中心。2025年企業債務市場總髮行量接近340億新元，較上年增長10%。
f) 可持續融資活動保持強勁，新加坡繼續在ASEAN的綠色、社會、可持續性和可持續性掛鉤（"GSSSL"）債券和貸款市場中處於領先地位，佔區域活動的一半以上。貸款起源超過27億新元，儘管全球出現小幅回撥，債券發行仍保持韌性。
27. 金融中心的增長得到有效且風險相稱的AML/CFT框架的支援，該框架保護新加坡金融體系和經濟的完整性，同時對合法商業擁有者和投資者保持歡迎態度。
a) 新加坡第五輪金融行動特別工作組（FATF）相互評估取得的積極成果驗證了金管局對金融部門的強有力且風險聚焦的AML/CFT監管、強有力的行業接觸和夥伴關係，以及良好的風險意識。
28. MAS繼續推出新措施以增強我們金融部門的競爭力並開發新的支柱。
29. 如副總理顏金勇在2026年6月25日ABS年度晚宴上所分享的，這些包括：
a) 深化新加坡的增長資本生態系統，以更好地調動和中介資本流動，支援亞洲長期增長；
b) 通過擬議的Protected Cell Company框架擴大替代風險轉移解決方案和風險中介能力；以及
c) 加強新加坡在黃金交易、清算和儲存中的作用。
30. 關於黃金，副總理顏金勇在6月亞太貴金屬會議上宣佈，MAS將取消基金稅收激勵計劃下對實物貴金屬投資的5%上限。
a) 我們想提供一個更新，該變化將從2026年8月1日起生效，為符合條件的基金和家族辦公室提供在新加坡投資實物黃金的更大靈活性。
b) MAS將很快釋出通函以闡述進一步的細節。
31. 今天，我將更新我們為維持創新、穩定性、安全和信任而進行的努力，這些是新加坡作為競爭和創新金融部門地位的基礎。我的評論將涉及2個領域：
a) 首先，擴大金融部門的創新規模。
b) 其次，加強我們金融體系對AI驅動威脅的應對能力。
金融部門創新規模擴大 32. 上月，MAS宣佈了建立金融未來研究所（FFI）的計劃。
33. 通過FFI，我們將在整個生態系統中對連線、能力、人才和基礎設施進行重大投資，以推動新加坡金融創新的下一階段。
a) FFI的創新車庫將匯聚金融機構（FIs）、FinTechs和技術提供商，共同建立和驗證新的用例。例如，我們正在形成金融和欺詐檢測中的智慧體AI的合作。
b) FFI將把金融行業與我們大學的研究能力相連，並建立行業相關的研究計劃。
c) 金融機構可以通過利用PathFin.ai來加速其採用之旅，從其他FIs、技術公司和FinTechs中找到行業驗證的解決方案。
d) AI採用必須與治理和護欄並行推進。現有的AI風險管理工具包將不斷更新和擴充套件，納入新的實施手冊、可重複使用的護欄、控制庫，以及在AI使用和風險的具體領域內的實施模板，以支援FIs安全地部署AI。
34. FFI將在接下來的幾個月內開始運營，並在未來一年內擴大其計劃規模。
35. 在新加坡開發FinTech生態系統也繼續是MAS的優先事項。目前有1,900家FinTech公司僱用將近10,000名員工。MAS將更新和增強我們對FinTech發展的支援。通過對能力發展、增長融資和人力資源接入的支援，我們旨在加強動態FinTech部門的增長前景。我們正在審查這些戰略，並與新加坡FinTech協會和該部門進行諮詢，將在今年晚些時候分享更多資訊。加強我們的金融體系應對AI驅動威脅的韌性 36. 現在讓我轉向加強我們的金融體系應對AI驅動威脅的韌性。
37. 隨著金融部門的快速數字化，我們在過去幾年中專注於增強我們的數字金融服務對欺詐、運營中斷和網路安全威脅的韌性。
38. 我們在這些領域中都引入了重大措施，並取得了有意義的進展。
39. 在打擊欺詐方面，我們與銀行合作增強其欺詐監控能力，並在數字交易流程中引入經過精心設定的摩擦，以阻礙欺詐。
a) 使用Money Lock功能從數字轉賬中保護的金額在2026年5月達到約47億新元，幾乎是一年前金額的兩倍。
b) 2025年，新加坡詐騙案件和損失金額同時出現自2021年以來的首次下降。
40. 為加強關鍵服務和職能的運營抗壓能力，我們一直在與金融機構合作，加強其跨四個關鍵支柱的風險管理框架的有效性：運營風險、技術和網路風險、第三方風險以及業務連續性。
a) 2026年3月，新加坡金融管理局就更新的《運營風險管理指引》和《第三方風險管理指引》進行諮詢。
i. 新指引對金融機構在第三方安排中理解和管理風險方面提出了更高的要求。
b) 我們目前正在就新加坡金融管理局《技術風險管理通知》的修訂徵求意見，以提升金融機構的技術抗壓能力。
i. 擬議通知強制要求在IT資產管理、IT容量規劃、持續系統和安全監控、資料備份和恢復等關鍵領域實施健全的做法。
41. 隨著數字支付在日常交易中的核心地位不斷提高，我們也一直在與行業合作，以提升零售支付的抗壓能力。
a) 2025年8月實施了NETS簽帳金融卡支付在銷售點的備用處理，使得在銀行系統中斷期間，無接觸簽帳金融卡支付可在限額內繼續進行。
b) 我們正在與主要銀行合作，為其數字零售支付引入類似的備用功能，以便銀行在系統中斷期間能夠繼續提供支付和資金轉賬等關鍵銀行服務。
42. 前沿人工智慧和量子計算的進步在打擊詐騙和網路威脅的鬥爭中增加了新的挑戰。我將介紹新加坡金融管理局如何與行業合作逐一應對這些挑戰。
43. 首先，人工智慧驅動的詐騙。
a) 藉助人工智慧，可以大規模地進行更個性化和更有說服力的網路釣魚，使用深度偽造冒充、詐騙資訊定製和其他欺騙戰術。
44. 迄今為止在數字銀行交易中引入的防護措施仍然有用。這些包括：
a) 更強大的身份驗證方法，例如通過數字令牌而不是一次性密碼進行應用內授權批准；
b) 為高風險交易設定冷卻期，給予潛在詐騙受害者在資金轉賬前重新考慮的時間。
45. 與此同時，銀行必須利用人工智慧並加強控制措施，以更好地防禦更復雜的人工智慧驅動詐騙戰術。
a) 新加坡金融管理局將審查銀行加強欺詐檢測模型有效性的努力，包括它們使用人工智慧的有效性。
b) 新加坡金融管理局目前也正在與政府科技局、新加坡警察部隊（SPF）和五家銀行合作，進行價值驗證，以測試在跨銀行和公私合作資料基礎上訓練的人工智慧模型是否能夠改進詐騙交易的整體檢測。
i. 價值驗證的發現預計將在明年準備好。
ii. 這些發現將指導進一步測試和增強銀行內部欺詐模型，並有助於確定建立行業級平臺以彙集資料並應用人工智慧的必要性。
46. 新加坡金融管理局也在繼續與銀行合作，研究更強大的身份驗證方法，例如符合FIDO標準的安全金鑰和軟體通行金鑰，以便行業在這些技術需要時做好準備。
47. 我們將繼續推進這些努力，使系統更難被利用，同時繼續與行業和公眾合作，保持對日益複雜的詐騙的警惕。
48. 現在讓我轉向人工智慧驅動的網路威脅。
49. 前沿人工智慧模型發現和利用系統漏洞的能力對金融機構的網路防禦構成威脅。
a) 前沿人工智慧模型的能力有兩個主要影響：
i. 首先，發現的漏洞數量大幅增加，包括在多家金融機構依賴的廣泛使用軟體和共享技術基礎設施中發現的漏洞；
ii. 其次，從發現到利用的時間線更短，從幾周縮短到幾天甚至幾小時。
b) 總體來說，這些顯著壓縮了修補、測試和修復的時間線。
c) 這使得全世界的系統所有者，包括我們的金融機構，難以及時、有力地執行大規模的補救工作。
50. 前沿人工智慧模型的能力必然要求緊迫提升網路防禦能力。
51. MAS於2026年4月首次釋出指引，呼籲金融機構進一步加強網路防禦。
a) 除了縮短應用安全補丁的時間外，金融機構被要求加速在安全編碼、漏洞檢測和安全測試等各種網路安全功能中採用AI。
b) 例如，金融機構應該擴大使用現有先進模型來發現漏洞。
52. MAS隨後在7月1日為關鍵金融機構引入了一項新要求，即對關鍵網際網路面向系統進行AI輔助紅隊測試。
a) 金融機構需要利用先進的AI模型來識別可能使網路犯罪分子中斷關鍵服務或獲取敏感客戶資料的潛在攻擊路徑。
b) 從這些演習中吸取的經驗教訓將與更廣泛的行業分享，以便金融機構能夠加強其集體韌性。
53. MAS將進一步加強我們的要求，並將很快為關鍵金融機構釋出監管期望，要求其制定和提交全面的評估和行動計劃，以加強對AI驅動網路威脅的防禦。這些應該以風險評估框架為基礎，並涉及金融機構的以下能力：
a) 大規模檢測和修補漏洞；
b) 在實施前測試系統變更；以及
c) 在發生中斷時備份、恢復和修復關鍵系統和服務。
54. AI和網路發展正在快速變化，需要不斷評估所造成的威脅並加強韌性措施。我想分享一個訊息，新加坡銀行協會AI驅動網路與技術風險工作小組（ABS-ACT）已經成立。
a) 工作小組由MAS和來自主要金融機構的資深技術和網路領導者組成，共同建立應對AI驅動威脅的戰略和措施。
b) 工作小組將專注於三個關鍵領域：
i. 首先，通過行業分享和與網路安全及AI專家的合作，提高金融機構在網路安全中使用AI的專業知識；
ii. 其次，合作進行先進AI驅動工具的試驗和驗證；以及
iii. 第三，開發行業指南，說明新措施、控制措施和解決方案，以更好地檢測、防止和應對複雜的AI驅動威脅。
55. 量子計算在中期內對金融機構使用的資料和通訊安全構成重大風險，需要提前做好準備。
a) 該技術仍處於發展階段。一些專家估計，量子計算可能在5-10年內破解現有加密技術，儘管也有觀點認為可能不到5年。
b) 然而，向量子安全實踐的過渡將需要時間，現在做好認真的準備肯定不會太早。
56. 在過去幾年，MAS已經開始為量子安全金融部門奠定基礎。
a) 我們在2024年向金融機構釋出了指引，強調了他們應該考慮的量子過渡措施。
b) 我們還與行業和國際同行就量子韌性密碼學解決方案進行了技術工作，以保護通訊安全。
57. MAS將在今年晚些時候釋出一套監管期望，為金融機構向量子韌性的遷移提供明確的期望、里程碑和時間表。將為金融機構設定漸進式時間表，以便：
a) 建立其密碼學資產清單；
b) 制定脆弱資產向量子韌性解決方案遷移的優先順序排序；以及
c) 建立技術能力和治理框架以支援量子安全遷移。
58. 我們的目標是金融機構在本十年末之前實現量子韌性。MAS 2025/26財年財務結果 59. 現在我轉向MAS的財務結果。
60. 截至2026年3月31日的財年，MAS實現淨利潤200億新元，主要原因是：
a) 按成本與市場價值較低值記錄的投資收益為398億新元；
b) 部分被抵消——
i. 新加坡元升值導致的負匯兌影響達16.4億新元；
ii. 金管局貨幣市場操作（或MMO）管理銀行系統流動性和其他支出的淨成本為24億新元；及
iii. 向政府統一基金的繳款為10億新元。
61. 投資收益表現強勁，因為全球經濟對反覆衝擊保持了韌性，金融市場表現良好。債券和股票等所有資產類別在發達市場和新興市場均獲得良好回報。雖然投資收益逐年變化，但今年的結果與去年相似，並高於我們10年曆史平均水平18.3億新元。未來前景存在不確定性，能源市場持續擾動、通脹上升和股票市場估值過高對未來投資回報構成風險。
62. 16.4億新元的負匯兌影響主要是由於新加坡元對美元和日元升值所致。新加坡元升值有助於緩和進口通脹的影響，但同時產生了負匯兌影響。這些影響產生於金管局以外幣形式持有的官方外匯儲備（或OFR）用新元報告時。這些影響不會影響金管局進行貨幣政策或維護金融穩定的能力，後者涉及使用外幣資產和流動性。
63. 與其他央行一樣，金管局通過貨幣市場操作（MMO）來吸收銀行系統過剩流動性而產生利息支出。這些成本反映了MMO的規模和利率水平，兩者在今年都有所下降。儲備管理政府證券（或RMGS）的收入部分抵消了這項支出。
64. 金管局將向統一基金繳款10億新元，並向政府返還淨利潤中的另外25億新元。向政府支付的金額反映了對以往年份累積虧損的抵消。
65. 金管局維持全球多元化、流動性充足的官方外匯儲備（OFR）投資組合，以支援我們作為央行的職能。金管局的財務結果反映了全球資產市場的表現，經計入MMO成本和新加坡元升值帶來的負匯兌影響。由於這些因素在不同年份有不同的影響，金管局年度財務結果也會出現波動。結論 66. 據此，我結束我的發言。感謝各位的關注。
英文原文
MAS 官網原始記錄 · 抓取日期: 2026-09-06
1. Good afternoon and thank you for joining us today for the release of the MAS Annual Report for Financial Year 2025/2026.
2. I will cover updates across central banking, financial sector development and regulation, as well as key highlights of our financial performance. Recent Economic Developments 3. Let me begin by touching on recent economic developments and monetary policy.
4. Over the past year, the macroeconomic landscape has been shaped by successive shocks. At the time of last year’s Press Conference, the global outlook was clouded by concerns that rising tariffs and trade policy uncertainty would weigh materially on growth. The subsequent outbreak of the Middle East conflict in late February 2026 introduced a new shock to the global economy.
5. In the face of these repeated shocks, the global economy has proved more resilient than expected. The impact of higher tariffs was cushioned by supply-chain reconfiguration. Global trade and industrial production continued to expand, even as tariff rates remained elevated. The scale of energy supply disruption since March was large, but cushioned by inventories, and agile supply and demand adjustments. Energy markets rebalanced with elevated prices at the lower range of anticipated scenarios, limiting the drag on growth, but lifting the path of headline inflation. Energy prices nevertheless remain elevated with risks to the upside from the renewed conflict in the context of lower global inventories.
6. The growth and resilience of the global economy was boosted by strong global AI investment. A surge in investment in data centres, chips and computing infrastructure and semiconductor capacity has supported a sustained expansion in global production and trade in electronic products. This has shifted growth outturns upwards, especially for economies that are deeply embedded in global technology supply chains.
7. Singapore has been exposed to all these crosswinds. The Middle East conflict has had a discernible impact on some pockets of the Singapore economy. Energy-related sectors such as chemicals manufacturing recorded double-digit contractions in Q2. However, the drag from these sectors was more than offset by the surge in the technology-related sectors. The Singapore economy recorded a strong 6% year-on-year growth in H1 2026, a step-up from the 5% growth in H2 2025.
8. Looking ahead, growth of the Singapore economy should stay firm for the rest of the year. While the Middle East situation still poses risks to the outlook, global AI-related demand is likely to continue to provide a meaningful boost. Most non-AI related sectors are likely to maintain a pace of growth that is close to trend.
9. Inflation has picked up from a low base and will step up further in the period ahead before easing in the second half of 2027. Domestic headline and core inflation was low at below 1.0% in 2025. In 1Q2026, core inflation was 1.4% and in Q2, this was 1.5%. We expect this to step up further from July and stay elevated for the next few quarters, driven by higher fuel and imported goods prices offsetting moderating domestic cost pressures and dampening effects of some government subsidies. MAS Core and CPI-All Items Inflation are projected to average 1.5-2.5% in 2026, stay elevated in the first half of 2027 and ease discernibly in the second half of 2027.
10. MAS’ monetary policy stance was also well-positioned at the start of the year. We maintained an appreciating stance in the second half of last year and again in January this year. This has helped to moderate the pick-up in inflation this year. In April, following the onset of the Middle East conflict, we tightened monetary policy in anticipation of rising imported inflationary pressures. Since then, Singapore’s import prices for a range of energy and other commodity inputs have risen sharply and have started to pass through to domestic consumer prices. This April move has therefore placed monetary policy settings in a favourable position as we head into a stronger inflationary environment.
11. The latest July MPS was a carefully calibrated policy adjustment that builds on April's policy decision. Inflation is forecast to step up further in July and stay elevated over the next few quarters before easing in the second half of 2027. The positive output gap is now expected to widen slightly, rather than narrow as envisaged in the April policy review. Taking the April and July decisions together, the stronger appreciation of the Singapore dollar will lean more effectively against the incoming inflationary pressures.
12. Amid the highly uncertain economic environment, MAS remains vigilant to risks to the outlook and is well-positioned to respond to maintain medium-term price stability and curb excess volatility in the Singapore Dollar Nominal Effective Exchange Rate (S$NEER). Financial markets and financial stability 13. I will turn now to developments in the financial markets and risks to financial stability.
14. Global financial conditions have thus far been benign and supportive of economic activity. Despite recent pullbacks, equity market valuations are high and credit spreads remain tight.
15. One major uncertainty to the benign picture thus far is the sustainability of the AI investment boom. Global growth, investment and financial market performance have become highly dependent on projections of large and increasing investment in data centres and semiconductor chips continuing well into the future. This is particularly so in the US and semiconductor-exporting Asian economies.
a) AI-driven electronics exports account for more than 70% of Asia's export growth year-to-date, up from 46% in 2024.
b) AI-connected firms now account for around 40% of S&P 500 market capitalisation and more than 30% of the MSCI EM Asia Index. They also dominate new financing raised in US capital markets, representing around half of investment-grade bond issuance, 38% of high-yield issuance, and 87% of new venture capital funding.
16. The sustainability of AI investments is thus highly consequential for global growth and financial stability. While near-term investments are supported by committed orders and strong hyperscaler cashflows, there is greater uncertainty around the sustainability of these investments in the medium term. In the race for model advantage and to scale adoption, projected investments by hyperscalers and model builders have expanded beyond cashflows and commercial revenues. Large equity and debt financing will be needed in the years ahead.
17. Markets will increasingly be looking to commercial revenue growth to justify the financing risks. Revenue growth will in turn depend on early signs of AI productivity gains at the firm level broadening across the economy and a deepening of transformative applications. The investment boom could be an extended one if supported by accelerating revenue growth and broadening productivity gains. But there are also clear risks on the path of AI investment monetisation. These include the escalating costs of energy and chips, supply bottlenecks of raw materials, regulatory uncertainty, intense competition among model providers, including from lower-cost open-weight models, as well as how widely shared the benefits of productivity gains are. If the payoff of AI investments falls short of expectations over the medium term, hyperscalers will moderate the pace of investment and markets will reassess asset valuations.
18. The implications are significant if either outcome comes to pass.
19. If we are in a long AI investment boom with significant and broad productivity gains, stronger and broader spillovers to income, demand and inflation could ensue. The impact on inflation will be complex and depends on the interplay of higher demand for energy and inputs against the pace and extent of productivity gains. These will have consequences for central banks’ assessment of potential output and neutral interest rates.
20. If on the other hand, there is a major retrenchment in AI investment, it could sharply weaken global growth through a fall in business investment and semiconductor demand and negative wealth effects. Financial stability risks could also materialise through equity, credit and loan markets’ exposures to unsustainable business models with deteriorating cashflows and weak credit terms in complex financing structures. A sharp tightening of global financial conditions could result.
21. A second source of risk is a prolonged re-escalation of the conflict in the Middle East. While the global oil market has responded so far with greater supply and demand agility and lower price impact than feared, a prolonged and more damaging conflict in the Middle East could reignite commodity price volatility amid lower inventories. While not our base case, we cannot discount the risk of oil prices moving sharply higher and shortages of oil and downstream products worsening.
22. MAS has placed emphasis on assessing domestic financial vulnerabilities given global stresses and uncertainties. We have updated our stress tests to capture key downside risks, including a resurgence of the Middle East conflict and a sharp tightening in financial conditions.
23. While the results affirm that the domestic financial system remains broadly resilient, there are small pockets of vulnerability among highly leveraged households and corporates.
a) Risks from a tightening in financial conditions are more pronounced for firms in energy-intensive sectors and for households with thinner financial buffers relative to expenditure.
b) As global uncertainty remains elevated, firms in more exposed sectors should provision for more liquidity and take steps to diversify revenue sources, while households with less stable incomes and thinner buffers should be cautious about taking on large new loan commitments.
Developments in the Financial Services Sector 24. I now turn to developments in the financial sector.
25. The financial sector saw healthy growth of 4.3% in 2025 despite a more uncertain global environment.
a) The growth in 2025 broadly maintains the momentum over the 2021-2025 period, which averaged 4.6% in annual growth and added 4,200 jobs Excludes jobs in holding companies. annually, with the gains in employment going to locals.
26. Growth continues to be broad-based.
a) Banking sector assets grew a steady 3.1% in 2025, easing from exceptional growth in 2024.
b) Assets in the insurance industry grew 7.6% in 2025, to S$493.5 billion, accelerating from 2024 and in line with the 7.5% CAGR from 2021 to 2025.
c) Assets under management grew 10.1% to reach S$6.7 trillion as at end-2025. The wealth management industry has also continued to grow strongly alongside the broader asset management industry.
d) As Asia’s leading FX hub, FX average daily traded volumes continued to expand to reach S$1.6 trillion in 2025.
e) Singapore also continues to grow as a leading regional fixed income hub. The corporate debt market saw total issuances of close to S$340 billion in 2025, a 10% increase from the previous year.
f) Sustainable finance activity also remained strong, with Singapore continuing to lead ASEAN’s market for green, social, sustainability and sustainability-linked (‘GSSSL’) bonds and loans, accounting for more than half of regional activity. Loan origination exceeded S$27 billion, and bond issuance remained resilient despite a modest global pullback.
27. The financial centre’s growth is supported by an effective and risk-proportionate AML/CFT framework that protects the integrity of Singapore’s financial system and economy, while remaining welcoming to legitimate business owners and investors.
a) The positive outcomes achieved in Singapore’s 5th round Financial Action Task Force (FATF) Mutual Evaluation validated MAS’ robust and risk-focused AML/CFT supervision of the financial sector, strong industry engagements and partnerships, and good risk awareness.
28. MAS continues to introduce new measures to sharpen the competitiveness of our financial sector and develop new pillars.
29. As shared by Deputy Prime Minister Gan Kim Yong at the ABS Annual Dinner on 25 June 2026, these include:
a) Deepening Singapore’s growth capital ecosystem to better mobilise and intermediate capital flows to support Asia’s long-term growth;
b) Scaling alternative risk-transfer solutions and risk intermediation capacity through the proposed Protected Cell Company framework; and
c) Strengthening Singapore’s role in gold trading, clearing and storage.
30. On gold, DPM Gan announced at the Asia Pacific Precious Metals Conference in June that MAS would be removing the 5% cap on physical investment precious metals under the tax incentive schemes for funds.
a) We would like to provide an update that this change will take effect from 1 August 2026, giving eligible funds and family offices greater flexibility to invest in physical gold in Singapore.
b) MAS will publish a circular soon to set out further details.
31. Today, I will update on our efforts to sustain the innovation, stability, security and trust that underpin Singapore’s position as a competitive and innovative financial sector. My comments will be in 2 areas:
a) First, scaling up innovation in the financial sector.
b) Second, strengthening resilience of our financial system to AI-enabled threats.
Scaling up innovation in the financial sector 32. Last month, MAS announced plans to establish the Future of Finance Institute (FFI).
33. Through the FFI, we will make significant investments in connections, capability, talent and infrastructure across the ecosystem to power the next phase of Singapore’s financial innovation.
a) FFI’s Innovation Garage will bring together financial institutions (FIs), FinTechs, and technology providers to co-create and validate new use cases. For example, we are forming collaborations in agentic AI in finance and fraud detection.
b) FFI will connect the financial industry with research capability in our universities and establish industry-relevant research programmes.
c) Financial institutions can accelerate their adoption journey by tapping on PathFin.ai to find industry-validated solutions from other FIs, technology companies and FinTechs.
d) AI adoption must proceed in tandem with governance and guardrails. Existing AI risk management toolkits will be continually updated and expanded with new playbooks, reusable guardrails, control libraries, and implementation templates in specific areas of AI use and risk to support FIs in deploying AI safely.
34. FFI will be operational in the next few months and scale up its programmes over the next year.
35. Developing the FinTech ecosystem in Singapore also continues to be a priority for MAS. There are today 1,900 FinTech companies employing close to 10,000 individuals. MAS will be renewing and enhancing our support for FinTech development. Through support for capability development, growth funding and manpower access, we aim to strengthen the growth prospects of the dynamic FinTech sector. We are reviewing these strategies and consulting with the Singapore FinTech Association and the sector and will share more later this year. Strengthening resilience of our financial system against AI-enabled threats 36. Let me turn now to strengthening resilience of our financial system against AI-enabled threats.
37. With rapid digitalisation in the financial sector, we have focused in the last few years on enhancing the resilience of our digital financial services against scams, operational disruptions and cybersecurity threats.
38. We have introduced significant measures and achieved meaningful progress in each of these areas.
39. On combatting scams, we have worked with banks to enhance their fraud surveillance capabilities and introduced calibrated frictions in digital transaction journeys to frustrate scams.
a) The amount of money that is protected from digital transfers using the Money Lock function has reached about S$47 billion as at May 2026, almost double the amount a year ago.
b) In 2025, Singapore saw a decrease in both scam cases and loss amounts for the first time since 2021.
40. To strengthen operational resilience for the delivery of critical services and functions, we have been working with FIs to step up the effectiveness of their risk management frameworks across four key pillars: operational risk; technology and cyber risk; third-party risk; and business continuity.
a) In March 2026, MAS consulted on updated Guidelines on Operational Risk Management as well as Third-Party Risk Management.
i. The new Guidelines set higher expectations for FIs to understand and manage risks in third-party arrangements.
b) We are currently consulting on amendments to the MAS Notices on Technology Risk Management which uplift technology resilience in FIs.
i. The proposed Notices mandate sound practices across key areas such as IT asset management, IT capacity planning, continuous system and security monitoring, as well as data backup and recovery.
41. As digital payments continue to be increasingly central to daily transactions, we have also worked with the industry to enhance retail payments resilience.
a) Stand-in processing for NETS debit payments at point of sales was implemented in August 2025, enabling contactless debit payments to continue up to a limit during a bank system disruption.
b) We are engaging with major banks to introduce similar stand-in capabilities for their digital retail payments so that the banks can continue to provide critical banking services such as payments and fund transfers during a system disruption.
42. Advances in frontier AI and quantum-computing add new challenges in the fight against scams and cyber threats. I will update on how MAS is working with the industry to address each in turn.
43. First, AI-enabled scams.
a) With AI, phishing can be made more personalised and persuasive at scale, using deep-fake impersonation, customisation of scam messages and other deception tactics.
44. The frictions introduced so far to digital banking transactions remain useful. These include:
a) More robust authentication methods such as in-app authorisation approvals via digital tokens instead of one-time passwords; and
b) Cooling periods for higher-risk transactions to give potential scam victims time to reconsider before funds are transferred.
45. At the same time, banks must harness AI and strengthen their controls to better defend against more sophisticated AI-enabled scam tactics.
a) MAS will review banks’ efforts to enhance the effectiveness of their fraud detection models, including how well they use AI.
b) MAS is also currently working with GovTech, the Singapore Police Force (SPF) and five banks on a proof-of-value to test whether AI models trained on cross-bank and public-private data can improve overall detection of scam transactions.
i. Findings from the proof-of-value are expected to be ready next year.
ii. The findings will guide further testing and enhancements to banks’ internal fraud models, as well as help determine the case for setting up an industry-level utility to pool data and employ AI.
46. MAS is also continuing to work with banks to study stronger authentication methods, such as FIDO-compliant security keys and software passkeys, so that the industry is ready when these are needed.
47. We will press ahead with these efforts to make the system harder to exploit, while continuing to work with industry and the public to sustain vigilance against increasingly sophisticated scams.
48. Let me turn now to AI-enabled cyber threats.
49. The capabilities of frontier AI models to find and exploit system vulnerabilities pose threats to FIs’ cyber defences.
a) Frontier AI model capabilities have two major implications:
i. Firstly, a higher volume of vulnerability findings, including in widely used software and shared technology and infrastructure relied on by multiple FIs; and
ii. Secondly, shorter discovery-to-exploit timelines, from weeks to days or even hours.
b) Together, these significantly compress the timelines for patching, testing, and remediation.
c) This has made it very challenging for system owners all over the world, including our FIs, to execute the volume of remediation in a timely and robust manner.
50. Frontier AI model capabilities necessitate an urgent step up in cyber-defence capabilities.
51. MAS first issued an advisory in April 2026 calling on FIs to further strengthen their cyber defences.
a) Apart from shortening the time to apply security patches, FIs were asked to accelerate the adoption of AI in various cybersecurity functions such as secure coding, vulnerability detection and security testing.
b) For example, FIs should widen the use of available advanced models to discover vulnerabilities.
52. MAS next introduced a new requirement on 1 July for key FIs to conduct AI-assisted red teaming on critical internet-facing systems.
a) FIs would need to leverage advanced AI models to identify potential attack paths that could enable cyber criminals to disrupt critical services or gain access to sensitive customer data.
b) The lessons learnt from these exercises will be shared with the broader industry, so that FIs can toughen their collective resilience.
53. MAS will step up our requirements further and will soon issue our supervisory expectations for key FIs to develop and submit comprehensive assessments and action plans to strengthen their defence against AI-enabled cyber threats. These should be underpinned by a risk assessment framework and address FIs’ ability to:
a) detect and patch vulnerabilities at scale;
b) test system changes before implementation; and
c) back up, restore and recover critical systems and services when disruptions occur.
54. AI and cyber developments are changing rapidly and there is a need to continually assess the threats posed and strengthen resilience measures. I would like to share that an ABS AI-Driven Cyber and Tech Risk Taskforce, or ABS-ACT, has been formed.
a) The Taskforce comprises MAS and senior technology and cyber leaders from major FIs to co-create strategies and measures against AI-driven threats.
b) The taskforce will focus on three key areas:
i. First, enhance FIs’ expertise in the use of AI in cybersecurity through industry sharing and engagement with cybersecurity and AI experts;
ii. Second, collaborate to conduct trials and validation of advanced AI-enabled tools; and
iii. Third, develop industry guidance on new measures, controls and solutions to better detect, prevent and respond to sophisticated AI-enabled threats.
55. Quantum-computing poses significant risks to the security of data and communications used by financial institutions in the medium term and requires early preparation.
a) The technology is still at a development stage. Some experts have given a 5-10 year estimated timeframe for when quantum-computing could break existing encryption techniques, although there are also some views that it could be less than 5 years.
b) Transition to quantum-safe practices however will take time and it is certainly not too early to make serious preparations.
56. Over the past few years, MAS had begun laying the groundwork for a quantum-safe financial sector.
a) We issued an advisory to FIs in 2024 highlighting quantum transition measures they should consider.
b) We have also done technical work with the industry and international counterparts on quantum-resilient cryptography solutions to secure communications.
57. MAS will next issue a set of supervisory expectations later this year with clear expectations, milestones and timelines for FIs’ migration towards quantum resilience. Progressive timelines will be set for FIs to:
a) Establish an inventory of their cryptographic assets;
b) Develop a prioritisation of the migration of vulnerable assets to quantum-resilient solutions; and
c) Build technical capabilities and governance frameworks to support a quantum-safe migration.
58. Our aim is for FIs to achieve quantum resilience before the end of this decade. MAS Financial Results for FY2025/26 59. I now turn to MAS’ financial results.
60. For the financial year ended 31 March 2026, MAS recorded a net profit of S$20.0 billion, driven by:
a) investment gains of S$39.8 billion, recorded on a lower of cost and market value basis;
b) partially offset by -
i. negative currency translation effects of S$16.4 billion from a stronger Singapore Dollar;
ii. net cost from MAS’ money market operations (or MMO) to manage banking system liquidity and other expenses of S$2.4 billion; and
iii. a contribution to the Government’s Consolidated Fund of S$1.0 billion.
61. Investment gains were strong, as the global economy remained resilient to repeated shocks and financial markets performed well. All asset classes across bonds and equities, developed and emerging markets posted good returns. While investment gains vary from year to year, this year’s outturn was similar to last year’s and above our 10-year historical average of S$18.3 billion. The outlook ahead is uncertain with continued energy market disruption, rising inflation and high equity market valuations posing risks to future investment returns.
62. The negative currency translation effects of S$16.4 billion were due mainly to the strengthening of the Singapore Dollar against the US Dollar and the Japanese Yen. The stronger Singapore Dollar has helped to dampen the effects of imported inflation, while resulting in negative currency translation effects. Such effects arise when MAS’ Official Foreign Reserves (OFR), which are held in foreign currencies, are reported in Singapore Dollars. These effects do not affect MAS’ ability to conduct monetary policy or support financial stability, which involve the use of foreign currency assets and liquidity.
63. Like other central banks, MAS incurs interest expense from MMO to absorb excess banking system liquidity. These costs reflect the volume of MMO and interest rate levels, both of which moderated over the year. Income from the Reserves Management Government Securities (or RMGS) partially offset this expense.
64. MAS will contribute S$1.0 billion to the Consolidated Fund, and return a further S$2.5 billion of its net profit to the Government. The amount to be paid to the Government reflects the offsetting of losses accumulated from previous years.
65. MAS maintains a globally diversified, liquid OFR portfolio to support our functions as a central bank. MAS’ financial results reflect the performance of global asset markets after accounting for the cost of MMO and negative currency translation effects from a stronger Singapore Dollar. As these factors have different effects from year to year, there will also be variability in MAS’ annual financial results. Conclusion 66. With that, I conclude my remarks. Thank you for your attention.