{"slug":"microsoft","name":"마이크로소프트","name_en":"Microsoft","ticker":"MSFT","updated_at":null,"canonical":"https://valuechain.wiki/microsoft","kicker":"클라우드 · AI 인프라 · 자체 실리콘","kicker_en":"Cloud · AI infrastructure · In-house silicon","lead":"윈도우·오피스·Azure 클라우드를 운영하는 빅테크. FY2025(2025년 6월 종료) 매출 $281.7B. AI 데이터센터 확충을 위해 현금 설비투자를 1년 만에 약 45% 늘려 $64.6B를 썼다. 엔비디아 GPU 최대 구매자이자 AMD·TSMC·자체 칩 마이아(Maia)로 공급선을 넓히는 AI 칩 수요의 핵심 종착지.","lead_en":"Big-tech operator of Windows, Office and the Azure cloud. FY2025 (ended June 2025) revenue was $281.7B. To expand AI data centers it raised cash capex about 45% in one year to $64.6B. As Nvidia’s largest GPU buyer, plus AMD, TSMC and its own Maia silicon, it is a key demand sink for AI chips.","segments":{"fy":2025,"parts":[{"desc":"Azure·서버 (매출 $106.3B)","name":"Intelligent Cloud"},{"desc":"Microsoft 365·Office·LinkedIn (매출 $120.8B)","name":"Productivity and Business Processes"},{"desc":"Windows·검색·게이밍 (매출 $54.6B)","name":"More Personal Computing"}],"total":"FY2025(2025년 6월 종료) 매출 $281.7B (+15%)","source":"https://www.sec.gov/Archives/edgar/data/789019/000095017025100235/msft-20250630.htm","concentration":"AI 인프라 수요의 종착지. 현금 capex $64.6B(FY2025, 전년 $44.5B에서 +45%). GPU는 엔비디아(2024년 호퍼 48.5만대, 최대 구매자)·AMD, 파운드리는 TSMC(자체칩 마이아), 연결은 Astera Labs."},"segments_en":{"parts":[{"desc":"Azure and servers ($106.3B revenue)","name":"Intelligent Cloud"},{"desc":"Microsoft 365, Office, LinkedIn ($120.8B revenue)","name":"Productivity and Business Processes"},{"desc":"Windows, Search, Gaming ($54.6B revenue)","name":"More Personal Computing"}],"total":"FY2025 (ended June 2025) revenue $281.7B (+15%)","concentration":"Demand sink for AI infrastructure. Cash capex $64.6B (FY2025, up 45% from $44.5B). GPUs from Nvidia (485k Hopper units in 2024, its largest buyer) and AMD, foundry at TSMC (its own Maia chip), connectivity from Astera Labs."},"edges":[{"direction":"revenue","counterparty":"기업·소비자 (Azure·M365·Windows)","counterparty_slug":null,"confidence":"high","metric":{"type":"revenue","unit":"USD","as_of":"2025-06-30","basis":"10-K 연간 총매출","value":281700000000,"period":"FY2025","basis_en":"10-K annual total revenue"},"note":"밸류체인의 수요 끝단. 기업·소비자에게 클라우드(Azure)·생산성 SW(M365)·Windows를 판매. FY2025 총매출 $281.7B.","note_en":"The demand end of the chain: sells cloud (Azure), productivity software (M365) and Windows to enterprises and consumers. FY2025 total revenue $281.7B.","sources":[{"stance":"confirms","publisher":"미국 증권거래위원회(SEC EDGAR)","publisher_en":"SEC EDGAR","kind":"filing","lang":"en","published_at":"2025-07-30","url":"https://www.sec.gov/Archives/edgar/data/789019/000095017025100235/msft-20250630.htm","title":"Microsoft FY2025 Form 10-K (회계연도 2025, 2025-06-30 종료)","title_en":"Microsoft FY2025 Form 10-K (year ended June 30, 2025)","excerpt":"FY2025 총매출 $281.7B. 부문별 Intelligent Cloud $106.3B, Productivity and Business Processes $120.8B, More Personal Computing $54.6B.","excerpt_en":"FY2025 total revenue $281.7B: Intelligent Cloud $106.3B, Productivity and Business Processes $120.8B, More Personal Computing $54.6B."}]},{"direction":"cost","counterparty":"NVIDIA","counterparty_slug":"nvidia","confidence":"high","metric":{"type":"units","unit":null,"as_of":"2024-12-19","basis":"옴디아 추정","value":485000,"period":"CY2024","basis_en":"Omdia estimate","qualifier":"estimate"},"note":"엔비디아 호퍼 GPU 최대 구매자. 2024년 약 48.5만대(H100·H200)로 차순위 구매자의 약 2배(옴디아 추정).","note_en":"Nvidia’s largest Hopper GPU buyer: about 485k units (H100/H200) in 2024, roughly double the next biggest (Omdia estimate).","sources":[{"stance":"mentions","publisher":"미국 증권거래위원회(SEC EDGAR)","publisher_en":"SEC EDGAR","kind":"filing","lang":"en","published_at":"2025-07-30","url":"https://www.sec.gov/Archives/edgar/data/789019/000095017025100235/msft-20250630.htm","title":"Microsoft FY2025 Form 10-K (회계연도 2025, 2025-06-30 종료)","title_en":"Microsoft FY2025 Form 10-K (year ended June 30, 2025)","excerpt":"10-K: FY2025 현금 capex $64.6B(전년 $44.5B에서 +45%)로 AI 데이터센터·가속기 투자가 급증(벤더명 비공개).","excerpt_en":"10-K: FY2025 cash capex of $64.6B (up 45% from $44.5B) reflects a surge in AI data-center and accelerator investment (vendors not named)."},{"stance":"confirms","publisher":"하이제(heise online)","publisher_en":"heise online","kind":"news","lang":"en","published_at":"2024-12-19","url":"https://www.heise.de/en/news/Microsoft-buys-most-Nvidia-accelerators-10215630.html","title":"마이크로소프트, 엔비디아 가속기 최다 구매 (옴디아 추정)","title_en":"Microsoft buys most Nvidia accelerators (Omdia estimate)","excerpt":"\"Microsoft purchased 485,000 Hopper GPUs, i.e. H100 and H200, in 2024\" — 옴디아 추정상 엔비디아 AI 가속기 최대 고객.","excerpt_en":"\"Microsoft purchased 485,000 Hopper GPUs, i.e. H100 and H200, in 2024,\" making it \"by far Nvidia’s largest customer for AI accelerators.\""},{"stance":"mentions","publisher":"엠투데이","publisher_en":"M Today","kind":"news","lang":"ko","published_at":null,"url":"https://www.autodaily.co.kr/news/articleView.html?idxno=537428","title":"MS, 엔비디아 칩 대체 선언…2026년 자체 AI 칩 마이아 200 본격 도입","title_en":"Microsoft to phase out Nvidia chips, deploys its own Maia 200 AI chip in 2026","excerpt":"마이크로소프트가 데이터센터에서 쓰는 AMD·엔비디아 칩을 자사 칩으로 단계적으로 대체하겠다고 밝혀, 현재 엔비디아 GPU 대량 사용을 확인.","excerpt_en":"Microsoft says it will gradually replace the AMD and Nvidia chips it uses in data centers with its own silicon, confirming heavy current Nvidia GPU use."}]},{"direction":"cost","counterparty":"AMD","counterparty_slug":"amd","confidence":"high","metric":null,"note":"Azure에서 AMD 인스팅트 MI300X로 Azure OpenAI 서비스(GPT-3.5/4)와 Copilot을 운영. ND MI300X v5 VM 제공.","note_en":"Runs Azure OpenAI Service (GPT-3.5/4) and Copilot on AMD Instinct MI300X; offers ND MI300X v5 VMs.","sources":[{"stance":"confirms","publisher":"AMD 투자자관계","publisher_en":"AMD Investor Relations","kind":"ir","lang":"en","published_at":"2024-05-21","url":"https://ir.amd.com/news-events/press-releases/detail/1198/amd-instinct-mi300x-accelerators-power-microsoft-azure-openai-service-workloads-and-new-azure-nd-mi300x-v5-vms","title":"AMD 인스팅트 MI300X, 마이크로소프트 Azure OpenAI 서비스 구동","title_en":"AMD Instinct MI300X Accelerators Power Microsoft Azure OpenAI Service Workloads","excerpt":"AMD 인스팅트 MI300X와 ROCm이 Azure OpenAI ChatGPT 3.5/4 서비스를 구동.","excerpt_en":"\"The AMD Instinct MI300X and ROCm software stack is powering the Azure OpenAI Chat GPT 3.5 and 4 services.\""},{"stance":"confirms","publisher":"서울경제","publisher_en":"Seoul Economic Daily","kind":"news","lang":"ko","published_at":null,"url":"https://www.sedaily.com/article/13803160","title":"AMD \"AI칩 MI300X, 오픈AI·메타·MS도 고객\"","title_en":"AMD: AI chip MI300X counts OpenAI, Meta, Microsoft among customers","excerpt":"AMD가 MI300X 고객으로 오픈AI·메타와 함께 마이크로소프트(MS)를 명시.","excerpt_en":"AMD names Microsoft, alongside OpenAI and Meta, among MI300X customers."}]},{"direction":"cost","counterparty":"TSMC","counterparty_slug":"tsmc","confidence":"high","metric":{"type":"process","unit":null,"as_of":"2024-08-26","basis":"TSMC N5 + CoWoS-S","value":null,"period":"2024","basis_en":"TSMC N5 + CoWoS-S","qualifier":"confirmed_no_disclosed_value"},"note":"자체 AI칩 마이아(Maia 100/200)를 TSMC N5 공정·CoWoS-S 패키징으로 위탁생산. 마이아 100은 HBM2E 64GB 탑재.","note_en":"Has its own Maia AI chips (Maia 100/200) fabricated at TSMC on N5 with CoWoS-S packaging; Maia 100 carries 64GB HBM2E.","sources":[{"stance":"confirms","publisher":"서브더홈(ServeTheHome)","publisher_en":"ServeTheHome","kind":"analysis","lang":"en","published_at":"2024-08-26","url":"https://www.servethehome.com/microsoft-maia-100-ai-accelerator-for-azure/","title":"Azure용 마이크로소프트 마이아 100 AI 가속기","title_en":"Microsoft Maia 100 AI Accelerator for Azure","excerpt":"마이아 100은 \"TSMC 5nm\" 공정에 \"TSMC CoWoS-S\" 패키징, \"64GB of HBM2E\" 메모리.","excerpt_en":"Maia 100 is a \"TSMC 5nm part\" using \"TSMC CoWoS-S\" packaging with \"64GB of HBM2E.\""},{"stance":"confirms","publisher":"엠투데이","publisher_en":"M Today","kind":"news","lang":"ko","published_at":null,"url":"https://www.autodaily.co.kr/news/articleView.html?idxno=537428","title":"MS, 엔비디아 칩 대체 선언…2026년 자체 AI 칩 마이아 200 본격 도입","title_en":"Microsoft to phase out Nvidia chips, deploys its own Maia 200 AI chip in 2026","excerpt":"MS가 자체 추론 칩 마이아 200을 2026년 본격 도입한다고 보도.","excerpt_en":"Reports Microsoft will deploy its own inference chip, Maia 200, in 2026."}]},{"direction":"cost","counterparty":"Astera Labs","counterparty_slug":"astera-labs","confidence":"mid","metric":null,"note":"PCIe·이더넷 리타이머와 패브릭 등 연결 반도체를 직접 구매하는 핵심 하이퍼스케일러 고객(소수 집중 구조).","note_en":"A key hyperscaler customer buying PCIe/Ethernet retimers and fabric connectivity directly (concentrated customer base).","sources":[{"stance":"confirms","publisher":"미국 증권거래위원회(SEC EDGAR)","publisher_en":"SEC EDGAR","kind":"filing","lang":"en","published_at":"2026-02-10","url":"https://www.sec.gov/Archives/edgar/data/1736297/000173629726000010/alab-20251231.htm","title":"Astera Labs 10-K (2025회계연도)","title_en":"Astera Labs Form 10-K (FY2025)","excerpt":"Astera Labs 10-K가 리타이머·패브릭을 직접 구매하는 소수 집중 하이퍼스케일러 고객으로 마이크로소프트를 포함.","excerpt_en":"Astera Labs’ 10-K lists Microsoft among the concentrated hyperscaler customers buying its retimers and fabric directly."}]},{"direction":"cost","counterparty":"Intel","counterparty_slug":"intel","confidence":"mid","metric":null,"note":"Azure 서버용 인텔 제온(Xeon) CPU 매입. ND MI300X v5 VM은 4세대 제온을 호스트 CPU로 사용.","note_en":"Buys Intel Xeon CPUs for Azure servers; the ND MI300X v5 VM uses 4th-gen Xeon as host CPU.","sources":[{"stance":"mentions","publisher":"AMD 투자자관계","publisher_en":"AMD Investor Relations","kind":"ir","lang":"en","published_at":"2024-05-21","url":"https://ir.amd.com/news-events/press-releases/detail/1198/amd-instinct-mi300x-accelerators-power-microsoft-azure-openai-service-workloads-and-new-azure-nd-mi300x-v5-vms","title":"AMD 인스팅트 MI300X, 마이크로소프트 Azure OpenAI 서비스 구동","title_en":"AMD Instinct MI300X Accelerators Power Microsoft Azure OpenAI Service Workloads","excerpt":"ND MI300X v5 시리즈는 8개 MI300X GPU와 함께 \"2개의 4세대 인텔 제온\" 프로세서로 구성.","excerpt_en":"The ND MI300X v5 series pairs 8 MI300X GPUs with \"two 4th Gen Intel Xeon Scalable processors.\""}]},{"direction":"cost","counterparty":"Marvell","counterparty_slug":"marvell","confidence":"mid","metric":{"type":"qualifier","unit":null,"as_of":"2026-04-14","basis":null,"value":null,"period":"2026","basis_en":null,"qualifier":"sole_source"},"note":"하이퍼스케일러 맞춤형 실리콘 파트너로 거론(구체 프로그램·물량은 비공개). 과대귀속 주의.","note_en":"Cited as a hyperscaler custom-silicon partner (specific program and volume undisclosed); attribution kept cautious.","sources":[{"stance":"mentions","publisher":"한국경제","publisher_en":"Korea Economic Daily","kind":"news","lang":"ko","published_at":"2026-04-14","url":"https://www.hankyung.com/article/2026041405651","title":"마벨, 하이퍼스케일러 맞춤형 AI 반도체 공급","title_en":"Marvell supplies custom AI silicon to hyperscalers","excerpt":"마벨이 하이퍼스케일러에 맞춤형 AI 반도체를 공급한다는 보도(특정 마이아 프로그램 귀속은 비공개).","excerpt_en":"Reports Marvell supplies custom AI silicon to hyperscalers (no public attribution to a specific Maia program)."}]},{"direction":"revenue","counterparty":"OpenAI","counterparty_slug":null,"confidence":"mid","metric":null,"note":"Azure가 OpenAI의 핵심 클라우드 컴퓨트 제공자. 마이아는 OpenAI 모델 워크로드에 맞춰 협업 설계.","note_en":"Azure is OpenAI’s primary cloud compute provider; Maia was co-designed with OpenAI for its model workloads.","sources":[{"stance":"confirms","publisher":"마이크로소프트 Azure","publisher_en":"Microsoft Azure","kind":"ir","lang":"en","published_at":"2023-11-15","url":"https://azure.microsoft.com/en-us/blog/azure-maia-for-the-era-of-ai-from-silicon-to-software-to-systems/","title":"AI 시대를 위한 Azure 마이아: 실리콘에서 소프트웨어, 시스템까지","title_en":"Azure Maia for the era of AI: From silicon to software to systems","excerpt":"마이아는 Azure OpenAI 서비스 등 워크로드를 위해 설계되었고 OpenAI와 협력했다고 마이크로소프트가 설명.","excerpt_en":"Microsoft says Maia was designed for workloads such as Azure OpenAI Service, developed in collaboration with OpenAI."}]},{"direction":"cost","counterparty":"메모리·HBM (SK하이닉스·삼성·마이크론)","counterparty_slug":null,"confidence":"low","metric":{"type":"qualifier","unit":null,"as_of":"2024-08-26","basis":null,"value":null,"period":"2024","basis_en":null,"qualifier":"confirmed_no_disclosed_value"},"note":"Azure 서버 DRAM·SSD와 마이아 HBM2E를 메모리 3사에서 매입(구체 공급사·물량 비공개).","note_en":"Buys server DRAM/SSD and Maia HBM2E from the three memory makers (specific supplier and volume undisclosed).","sources":[{"stance":"mentions","publisher":"서브더홈(ServeTheHome)","publisher_en":"ServeTheHome","kind":"analysis","lang":"en","published_at":"2024-08-26","url":"https://www.servethehome.com/microsoft-maia-100-ai-accelerator-for-azure/","title":"Azure용 마이크로소프트 마이아 100 AI 가속기","title_en":"Microsoft Maia 100 AI Accelerator for Azure","excerpt":"마이아 100이 HBM2E 64GB를 사용한다고 명시(HBM은 SK·삼성·마이크론 공급).","excerpt_en":"Maia 100 uses 64GB of HBM2E (HBM supplied by SK hynix, Samsung and Micron)."}]}]}