SK C&C Pangyo Data Center Fire and the National Kakao/Naver Outage
On 15 October 2022, a lithium-ion UPS battery on the third basement level (B3) of the SK C&C Pangyo Data Center ignited and spread, forcing firefighters and operators to cut power to the entire facility. Because tenant Kakao had concentrated core functions — including its sole authentication service — at this single site, the power cut cascaded into a nationwide outage of KakaoTalk, Daum, Kakao-linked banking, KakaoT and dozens of downstream services. Co-tenant Naver, which kept its own separate data center, was far less affected. KakaoTalk stayed down for more than 11 hours, the longest in its history, and full restoration of all Kakao services took roughly five days. The event triggered a CEO resignation, a three-way blame dispute, and South Korea's "Kakao Outage Prevention Act" mandating geographic redundancy for major platforms.
Failure cascade
Trigger → primary fault → downstream blast radius, derived from the sourced root cause and affected-services record.
Facility & location
- Operator
- SK C&C (tenants: Kakao, Naver)
- Data center
- SK C&C Pangyo Data Center
- Location
- Seongnam (Pangyo), South Korea
- Date
- 2022-10-15
Impact & scale
- Users affected
- Nationwide — the majority of South Korea's population relies on KakaoTalk; ~2M users reported to have left the app in the immediate aftermath (single-source, low confidence)
- Financial
- Over KRW 2 trillion in Kakao-affiliate market-cap loss on Oct 17 (single-source, low confidence); Kakao committed KRW 460B (~US$323M) to new Ansan capacity
- Scope
- National-infrastructure-grade outage
- KakaoTalk (messaging)
- KakaoPage
- Daum portal / Daum News
- Brunch (blogging)
- Kakao Map
- Kakao Bank (partial — Kakao-auth-dependent functions)
- KakaoT (ride-hailing)
- Kakao Pay securities/real-estate functions
Impact data & metrics
| Ignition-to-suppression interval | ~1 minute (spark 15:19 -> gas discharge 15:20) |
| BMS-reported battery temperature during active fire | ~20C (held steady; sensing failed) |
| Clean-agent gas discharged | ~1,500 kg (50 kg x 30 cylinders) |
| Battery packs consumed | ~60 packs |
| Fire-charred area vs room size | ~40 sqm charred of a 3,300 sqm battery room |
| Spark-to-firefighter-arrival | 18 minutes (15:19 -> 15:37) |
| Kakao servers de-energised by whole-building cut | ~32,000 servers (3만2000대) |
| Peak firefighting mobilization | 114 personnel, 46 apparatus |
| Occupants evacuated / casualties | 26 self-evacuated / 0 casualties |
| Total time to full extinguishment | 8 hours 13 minutes (15:33 -> 23:46) |
| Kakao service full-recovery time | 127 hours 30 minutes (>5 days) |
| Ignition cell service age | ~6 years (installed 2016) |
Magnitude profile
Blast radius and user impact are at ceiling: KakaoTalk is de-facto national communications infrastructure, and the outage propagated into banking, ride-hailing, deliveries and grocery. Duration scores high on the 5-day-plus full-restoration tail (7,650 min ≈ 127.5 h) even though the headline KakaoTalk downtime was ~11 hours. Financial score reflects the reported >KRW 2 trillion affiliate market-cap loss (low-confidence, single-source) plus the KRW 460B remediation commitment.
Sequence of events (SOE)
- TRIGGER Lithium-ion UPS batteries installed in the B3F electrical/battery room; per MSIT no Li-ion-specific fire measures were ever added over the ~6 years that followed (latent condition)
- TRIGGER CCTV captures a spark from a single Li-ion cell on a UPS battery rack in the B3F electrical room; internal short from age-related insulation breakdown initiates thermal runaway
- DETECTION BMS fails as a detection layer — captures no abnormal warning and holds battery temperature at a steady ~20C even as the cell burns
- DETECTION Separate automatic fire-detection loop triggers ~1 minute after the spark; automatic suppression commanded to discharge
- MITIGATION Clean-agent/halogen GAS suppression discharges ~1,500 kg (50 kg x 30 cylinders); system activates correctly but does not extinguish the fire
- MITIGATION Gas suppression proven ineffective — sealed Li-ion cells in an enclosed basement room prevent oxygen-displacement agents from penetrating; only water (cooling) can arrest the runaway
- DETECTION Official fire-department report time logged for the B3F electrical-room fire at SK Pangyo Campus Building A
- MITIGATION Firefighters arrive on scene, 18 minutes after the 15:19 spark
- IMPACT All occupants self-evacuate — 26 workers exit under their own power with zero casualties
- MITIGATION Because water needs the power off (electrocution risk), fire services formally request SK C&C cut electricity to the B3F zone
- CASCADE Selective isolation is impossible — the B3F UPS also feeds the whole center — so a WHOLE-BUILDING power cut is executed, de-energising all ~32,000 Kakao servers
- IMPACT Containment holds: the fire stays within the B3F electrical/battery room and never reaches the server halls; the national outage is caused by loss of POWER, not fire damage to servers
- MITIGATION Main flames knocked down roughly two hours after ignition once water is applied to the ~40 sqm affected area (of a 3,300 sqm room); ~60 battery packs consumed
- RECOVERY Full extinguishment declared — an 8h13m operation from the 15:33 report; peak mobilization 114 personnel and 46 apparatus
- RESTORED Kakao services fully normalized after 127 hours 30 minutes of degraded/failed operation — recovery driven by restarting a single-site fleet with no geographic redundancy
- RECOVERY MSIT announces investigation results: BMS detection failure (held ~20C), cabling routed over batteries, no Li-ion-specific measures since 2016, UPS not fully separated; fault assigned to both SK C&C and Kakao
- RESTORED National Forensic Service announces mechanism: internal short from age-related insulation breakdown as probable ignition cause, external factors unlikely, precise trigger still under study
- IMPACT National-scale service disruption begins as Kakao and Naver platforms fail following the power loss at the single Pangyo site
Root cause
Contributing factors
- MAINTENANCE/MANAGEMENT LAPSE: Since installing Li-ion batteries in 2016, SK C&C had never implemented any Li-ion-specific fireproofing or response measures over ~6 years — MSIT: '2016년 리튬이온배터리 도입 이후 현재까지 특화된 방화조치는 한 적이 없었던.'
- DETECTION DESIGN FAILURE: The installed BMS was inadequate/malfunctioning as an early-warning layer — it held ~20°C during an active fire and captured no abnormal signs before ignition, so the primary battery-monitoring system detected nothing.
- SUPPRESSION MISMATCH: The battery room was protected only by clean-agent/halogen gas suppression, which is physically incapable against sealed Li-ion cells in thermal runaway; no large-area water sprinkler system existed in the battery room.
- POWER TOPOLOGY / SINGLE-UPS DESIGN: The UPS feeding the fire zone also fed the entire center, cabling was routed directly above the batteries (burning through immediately), and some UPS units were not fully partitioned from the battery room — making selective isolation impossible and forcing a whole-building shutdown.
- NO EMERGENCY PLAN OR DRILLS: SK C&C had no detailed large-scale emergency-response plan and no drills/simulations for a whole-center power-loss scenario.
- KAKAO SINGLE-SITE DEPENDENCE: Kakao relied almost entirely on the single Pangyo site with no real contingency for the whole data center going dark; MSIT assigned fault to both SK C&C and Kakao.
- REGULATORY GAP: Enhanced data-center sprinkler requirements took effect Feb 2022 but were not applied retroactively to the 2016-built facility, so the battery room had no mandated water suppression.
Correction of errors (COE)
- Replace/supplement gas-only suppression in Li-ion battery rooms with water-based cooling suppression (sprinkler, water-mist, or immersion)
- Zone UPS feeds and re-route power cabling away from and not above battery racks so a single fire zone can be selectively de-energised
- Deploy multi-modal early detection (cell-level thermal + off-gas/VOC + smoke + thermal imaging) beyond pack-temperature BMS
- Establish geographic multi-site redundancy and mandatory full-power-loss disaster-recovery drills for critical services
- Extend enhanced data-center suppression code (effective Feb 2022) to legacy pre-2022 facilities
- Institute Li-ion-specific fire-prevention and battery aging/end-of-life replacement program
Lessons learnt
- A physically small fire can cause a national-scale outage when the failure vector is de-energisation, not combustion: only ~40 sqm burned and servers were never touched by flame, yet ~32,000 servers went dark because power had to be cut to fight it.
- Suppression must match the fuel chemistry: oxygen-displacing gas is physically incapable against a sealed Li-ion cell in thermal runaway, which makes its own heat and oxygen — only water's cooling effect can arrest it, so gas-only battery rooms carry a hidden zero against their worst hazard.
- A monitoring system that reads the wrong signal is worse than none: the BMS reported a steady ~20C through an active fire, giving false assurance while offering no early warning — detection must be validated against the specific failure mode it is meant to catch.
- Power topology is a single point of failure: routing one UPS and its cabling to feed both the fire zone and the whole center, with cabling above the batteries and incomplete separation, guaranteed that suppressing a local fire meant killing everything.
- Single-site dependence turns any facility incident into an existential outage: Kakao's lack of geographic redundancy and of drills for whole-center power loss stretched recovery to 127h30m — the fire lasted hours, the outage lasted days.
- Legacy facilities need retroactive code uplift: enhanced data-center suppression rules effective Feb 2022 did not reach the 2016 building, leaving the highest-risk room governed by the weakest standard.
Improvements & remediation
- SafetyReplace or supplement gas-only suppression in Li-ion battery rooms with water-based systems that provide cooling (sprinkler/immersion/water-mist), since gaseous oxygen-displacement agents cannot arrest thermal runaway in sealed cells — reporting showed ~1,500 kg of gas failed to extinguish a ~40 sqm fire.
- MaintenanceInstitute a mandatory Li-ion-specific fire-prevention and battery-health program (periodic cell inspection, aging/degradation testing, end-of-life replacement) — SK C&C had done no Li-ion-specific measures in the ~6 years since 2016 install, and the ignition cell had aged into an internal short.
- DesignPhysically separate Li-ion battery rooms from UPS units with fire-rated bulkheads (including ceiling voids), route power/data cabling AWAY from and not above battery racks, and zone UPS feeds so a single fire area can be de-energised without cutting the whole center.
- MaintenanceUpgrade battery monitoring from a BMS that only reads pack-level temperature to a multi-modal early-detection stack (cell-level thermal, off-gas/VOC, smoke, thermal imaging) — the existing BMS held ~20C during an active fire and detected nothing.
- ProcessRequire geographic/multi-site redundancy and disaster-recovery for critical services (Kakao ran single-site) and mandate regular full-scale drills simulating total data-center power loss — none existed.
- SafetyEstablish pre-agreed selective power-isolation maps and switch labeling so firefighters can safely apply water to a burning zone in minutes; staff here could not identify isolation switches in time, forcing a whole-building shutdown.
- ProcessClose the regulatory gap by extending enhanced data-center suppression requirements (effective Feb 2022) to existing/legacy facilities rather than exempting pre-2022 buildings like the 2016 Pangyo center.
Comprehensive analysis
A small fire, a national outage
The Pangyo incident is a textbook case of impact decoupled from physical damage. Only ~40 sqm of a 3,300 sqm battery room charred, ~60 packs burned, and flames never reached the server halls — the ~32,000 Kakao servers and Naver gear were intact throughout. The outage came entirely from the whole-building power cut needed to fight the fire safely. The lesson is that a data center's blast radius is defined by its electrical topology and its dependency graph, not by how much of it burns.
Why gas suppression could not work
The clean-agent/halogen system discharged correctly within ~1 minute of the spark, yet the fire burned on. This was physics, not malfunction: gaseous agents extinguish by displacing oxygen, but a Li-ion cell in thermal runaway generates its own heat and oxygen and is sealed in casing, so the agent has no gap to penetrate. In an enclosed basement room, only water — which cools as well as smothers — can arrest the runaway. Reporting underscored the mismatch: ~1,500 kg of gas failed to stop a fire over roughly 40 sqm.
The detection blind spot
The BMS, nominally the early-warning layer, reported a steady ~20C throughout an active fire and flagged no anomaly before ignition. Whether the sensing scheme simply could not see a cell-internal short or the loop had failed silently, the operational result was false assurance. Detection that is not validated against the exact failure mode it must catch — internal short and thermal runaway in an aged cell — is a control that exists on paper only.
Topology as the amplifier
The single most consequential design choice was coupling: the UPS feeding the B3F fire zone also fed the whole center, cabling ran directly above the batteries and burned through at onset, and some UPS units were not fully separated from the battery room. When firefighters needed the power off at 16:52 to apply water, staff could not identify which switches would isolate only the fire zone, so a whole-building shutdown was the only option. Good design would have made selective isolation a labeled, drilled, minutes-long action.
Single-site dependence and the regulatory gap
Kakao's near-total reliance on one site, with no geographic redundancy and no drills for a full power-loss scenario, turned a facility fire into a 127h30m recovery. MSIT assigned fault to both operator and tenant. Compounding it, enhanced data-center suppression requirements effective Feb 2022 were not applied retroactively to the 2016 building, leaving the highest-risk room under the weakest standard — a reminder that grandfathering legacy critical infrastructure can institutionalize the very gap that fails.
Technical deep-dive
References & provenance
- official-postmortem MSIT official press release '디지털서비스 장애 원인 조사결과 및 시정조치 요구' (Ministry of Science and ICT), republished verbatim on the Korean-government policy-briefing portal korea.kr (KOGL type-1)“10.15 에스케이 씨앤시 판교 데이터센터 화재 및 카카오·네이버 등 부가통신 서비스 장애에 대한 조사 결과를 발표하고, 에스케이 씨앤씨, 카카오, 네이버 3사에게 1개월 이내 주요 사고원인에 대한 [시정조치를 요구]”https://www.korea.kr/briefing/pressReleaseView.do?newsId=156540807
- press MSIT investigation results on the SK C&C Pangyo fire (BMS failure, cabling over batteries, no Li-ion measures since 2016)“배터리모니터링시스템(BMS)이 갖춰져 있었으나 화재 발생 직전까지 이상징후를 포착하지 못했다”http://v.daum.net/v/20221206140033953
- press National Forensic Service ignition-mechanism conclusion“배터리 셀 내부 경년열화에 따른 절연파괴로 발생한 단락”http://v.daum.net/v/20230330151845674
- press Fire timeline, gas volume, charred area and firefighting response“1500kg 상당의 가스가 방사됐음에도, 40㎡ 남짓을 태운 불길을 잡는데만 2시간이 소요됐다.”http://v.daum.net/v/20221018172414900
- press BMS held ~20C during fire; Kakao recovery of 127h30m“불이 났음에도 BMS상의 배터리 온도는 20도 수준을 일정하게 유지하고 있었는데”http://v.daum.net/v/20221206155455256
- press Whole-building power cut de-energised all ~32,000 Kakao servers“데이터센터에 보관된 카카오 서버 전체인 3만2000대의 전원 공급이 끊겨 이번 사태가 발생했다.”http://v.daum.net/v/20221018030244168
- press Fire report 15:33, extinguishment 23:46, mobilization 114 personnel/46 apparatus“장비 46대와 소방관 등 인력 114명을 투입해 오후 11시 46분 진화작업을 완료했다.”http://v.daum.net/v/20221016101349346
- press Evacuation of 26 workers, zero casualties“건물 안에 있던 직원 등 26명이 긴급 대피했으며 ... 이번 불로 다행히 인명피해는 발생하지 않았다.”http://v.daum.net/v/20221016103044844
- press National Forensic Service (국립과학수사연구원, NFS) 감정 결과 conveyed to the Gyeonggi Nambu Provincial Police SK C&C fire investigation team (28 Dec 2022) — quoted verbatim by Yonhap“국과수는 … "배터리 셀 내부의 경년열화(經年劣化)에 따른 절연 파괴로 인해 발생한 단락이 발화의 원인으로 작용했을 가능성이 있다" … "외부적 요인에 의한 화재 가능성은 작다"”https://www.yna.co.kr/view/AKR20221229065500061
- press Police + NFS + Gyeonggi Fire HQ + KESCO joint on-site forensic exam (2차 합동감식, 17 Oct 2022) locating the ignition point inside a battery module — quoted verbatim by Yonhap“감식 결과 발화부는 배터리 모듈 내부로 추정됐다. … 경찰 관계자는 "발화부는 전소된 5개 랙에 보관된 배터리 모듈 내부로 추정된다"”https://www.yna.co.kr/view/AKR20221017071351061
- press SK C&C 데이터 센터 화재 (Korean Wikipedia)“지하 3층 무정전전원장치에서 발화 … 8시간 13분 후인 그날 23시 46분에 진화되었다”https://ko.wikipedia.org/wiki/SK_C%26C_%EB%8D%B0%EC%9D%B4%ED%84%B0_%EC%84%BC%ED%84%B0_%ED%99%94%EC%9E%AC
- press 판교 SK C&C 데이터센터 화재 — ZDNet Korea (same-day report)“건물 전체 전원 공급을 차단 … 화재가 서버실과 전산실까지 확산되지 않아”https://zdnet.co.kr/view/?no=20221015231003
- press Kakao (English Wikipedia)“In October, co-CEO Namkoong Whon resigned after a service outage caused by a data center fire.”https://en.wikipedia.org/wiki/Kakao
- news SK's Li-ion batteries blamed for data center fire behind Kakao outage — DatacenterDynamics“SK's lithium-ion batteries were blamed for the data center fire behind the Kakao outage.”https://www.datacenterdynamics.com/en/news/sks-li-ion-batteries-blamed-for-data-center-fire-behind-kakao-outage/
- news All stakeholders blamed for Kakao server shutdown — Korea Herald“The ministry pinpointed lithium-ion batteries not being physically separated from the UPS as the main cause of the slow recovery.”https://www.koreaherald.com/article/3015077
- news Fire at data center causes unprecedented massive service disruption — Korea Times“A fire at a data center caused an unprecedented, massive disruption of internet services.”https://www.koreatimes.co.kr/southkorea/society/20221016/fire-at-data-center-causes-unprecedented-massive-service-disruption
- analysis Major data center fire highlights criticality of IT services — Uptime Institute Journal“According to police reports, the fire started in a battery room before spreading quickly to the rest of the building.”https://journal.uptimeinstitute.com/major-data-center-fire-highlights-criticality-of-it-services/
- news South Korea data center fire fallout: CEO resigns — Data Center Knowledge“Investing 460 billion won ($323 million) to complete new data center in Ansan by next year.”https://www.datacenterknowledge.com/management/south-korea-data-center-fire-fallout-ceo-resigns
- news Korea's ICT ministry gives Kakao, SK and Naver one month to implement emergency response — DatacenterDynamics“South Korea's ICT ministry has given Kakao and other players one month to come up with emergency response measures.”https://www.datacenterdynamics.com/en/news/koreas-ict-ministry-gives-kakao-sk-and-naver-one-month-to-implement-emergency-response-system/
- news Fire at data center causes KakaoTalk to go down for 11+ hours — allkpop“KakaoTalk went down for 11+ hours, the longest inoperable period of the app's history.”https://www.allkpop.com/article/2022/10/fire-at-data-center-causes-kakaotalk-to-go-down-for-11-hours-longest-inoperable-period-of-apps-history
- news SK C&C fire disrupts Kakao and Naver — The Register“Naver already operates one datacenter of its own and has another one planned for next year.”https://www.theregister.com/2022/10/17/sk_cc_naver_kaako_fire/
- aggregator SK C&C Pangyo data center fire — internet service outage (aggregator)“15:30 - Service failures began; power supply cut during firefighting.”https://en.namu.wiki/w/SK%20C&C%20%ED%8C%90%EA%B5%90%20%EB%8D%B0%EC%9D%B4%ED%84%B0%EC%84%BC%ED%84%B0%20%ED%99%94%EC%9E%AC%EB%A1%9C%20%EC%9D%B8%ED%95%9C%20%EC%9D%B8%ED%84%B0%EB%84%B7%20%EC%84%9C%EB%B9%84%EC%8A%A4%20%EC%9E%A5%EC%95%A0%20%EC%82%AC%EA%B1%B4
Sourced from public post-incident reports. Quotes are short attributed excerpts for provenance only; the analysis above is original and substantially shorter than its sources. Last verified 2026-07-31.