Huawei
Vertically diversified hardware + software + services conglomerate. Reporting lines: ICT infrastructure (carrier + enterprise networking/5G/optical), consumer devices (smartphones/PCs/wearables/HarmonyOS), Huawei Cloud, Digital Power (solar/energy), and Intelligent Automotive Solutions (HiCar/ADS supplier to automakers). Held via a labor-union holding vehicle; profits distributed to 169,054 current and former employee virtual-shareholders (as of Dec 31, 2025), with founder Ren Zhengfei retaining ~0.59-1%.
Huawei has never raised external/venture capital and has no priced funding rounds - the only real, transacted valuation basis is the audited net asset value (total equity) at which the employee virtual shares are internally valued and redeemed. These points are audited book value, not a market valuation; a listable franchise value would likely be materially higher. RMB converted at ~7.2/USD (2023) and ~7.3/USD per Huawei's own 2024 USD figure.
Earnings, margins, COGS & capex
FY2025 revenue reached RMB 880.9B (~$122B), up 2.2% and the second-highest in company history (just RMB ~1.05B short of the 2020 peak), after a +22.4% rebound in FY2024. Net profit rose ~8.6% to ~RMB 68B (~$9.4B, ~7.7% net margin), recovering from the FY2024 dip (RMB 62.6B, -28%) that was driven largely by a collapse in 'other net income' (RMB 62.7B in 2023 to RMB 8.3B in 2024, absent the prior-year Honor sale gain) plus margin compression from internalizing the supply chain under sanctions. Growth is now moderate and reinvestment-heavy: R&D hit a record RMB 192.3B (21.8% of revenue), funding the Ascend AI and HarmonyOS self-sufficiency push. Automotive is the standout grower (+72.1%); Cloud external revenue fell 3.5% to RMB 32.16B as Chinese AI compute demand shifted to rivals. Audited total equity (book value) rose to RMB 544.62B (~$74.65B) at end-2024 from RMB 507.57B a year earlier.
Income statement — where each revenue dollar goes
% of revenueOf every $1 of revenue, ~56¢ is cost of goods and ~35¢ operating expense, leaving ~9¢ of operating profit (~8¢ net).
Revenue trend
Margins
down from 46.2% FY2023 as component sourcing costs and supply-chain internalization bit
down from ~14.8% FY2023
up - net profit +8.6% on flat-ish revenue after FY2024's -28% trough
up from 20.8% FY2024 (RMB 179.7B); ~113,000-114,000 R&D staff, over half of headcount
COGS structure
COGS is dominated by hardware bill-of-materials - semiconductors, RF/optical components, displays, batteries. Sanctions forced substitution of US/allied parts (Google GMS, TSMC leading-edge foundry, high-end EDA/lithography access) with domestic and self-designed alternatives (HiSilicon Kirin/Ascend fabbed at SMIC on ~7nm DUV), raising unit costs and depressing gross margin vs the pre-2020 baseline. Consumer-device COGS improved in 2024-25 as the domestic Kirin supply chain matured.
Capex
Not separately disclosed. Effective capital intensity is high and rising: Huawei is reportedly building its own advanced chip-production lines in Shenzhen for 7nm smartphone and Ascend AI processors, and plans to roughly double Ascend 910C output toward ~600,000 dies in 2026 (up to ~1.6M dies across the Ascend line) - a domestic-fab and packaging build-out that substitutes for lost TSMC access.
Latest earnings
Management characterized FY2025 as 'performance in line with forecast'; no external consensus exists (unlisted).
No formal guidance issued. Rotating-chair commentary signals continued heavy R&D reinvestment, restraint on over-expansion (Meng Wanzhou: 'restrain development boundaries'), and prioritizing technological sovereignty over near-term margin.
- FY2025 revenue
- RMB 880.9B (~$122B), +2.2%
- FY2025 net profit
- ~RMB 68B (~$9.4B), ~+8.6%
- R&D spend
- RMB 192.3B, 21.8% of revenue
- ICT infrastructure
- RMB 375B, +2.6%
- Consumer devices
- RMB 344.5B, +1.6%
- Digital energy
- RMB 77.3B, +12.7%
- Intelligent automotive
- RMB 45.0B, +72.1%
- Cloud (external)
- RMB 32.16B, -3.5%
Growth drivers
- Ascend AI accelerators (910B/910C) as the leading domestic Nvidia-alternative for Chinese datacenters - 910C output set to roughly double to ~600K dies in 2026; Alibaba, Tencent and DeepSeek among reported/prospective buyers (ByteDance also cited in press)
- Consumer-device recovery — Kirin-powered 5G smartphones drove a China market-share comeback; HarmonyOS NEXT (de-Androidized) with a large cumulative install base, going global in 2026
- Intelligent Automotive Solutions — fastest-growing line (+72.1% to RMB 45.0B FY2025); ADS autonomous-driving stack and HiCar supplied to AITO/Seres and other automakers, profitable since 2024
- Digital Power / energy (solar inverters, datacenter power, storage) - +12.7% to RMB 77.3B FY2025
- 5G/5.5G (5G-Advanced) carrier upgrades and enterprise networking outside sanction-restricted Western markets; strength across China, MEA, SE Asia, LatAm
- Kunpeng + Ascend developer ecosystems building a domestic compute stack
Bull & bear
The de-facto national champion of China's technological-sovereignty program: a diversified ~$122B-revenue conglomerate that survived a maximal US sanctions campaign, rebuilt its own chip-to-OS stack, and now owns the inside track on China's AI-compute and telecom spend with a 22%-of-revenue R&D engine and no shareholder short-termism.
- Survived the worst-case sanctions scenario and returned to near-record revenue (RMB 880.9B, 2nd-highest ever) with profit growing again (+8.6%) - proof of resilience
- Ascend is the only scaled domestic Nvidia alternative in a market Beijing is actively walling off; 910C output set to double in 2026 with named prospective buyers (Alibaba, Tencent, DeepSeek)
- Full-stack vertical integration (silicon + HarmonyOS + cloud + energy + auto) compounds - each cutoff spawned a new in-house franchise
- Automotive line +72.1% and profitable; Digital Power +12.7% - real diversification beyond the sanction-exposed core
- Employee ownership + RMB 192B annual R&D allows decade-long bets Western public rivals can't fund through the earnings cycle
- Structural policy tailwind: 'buy domestic' compute/telecom mandates make China's largest market a protected moat
A sanctioned, margin-compressed, un-investable-to-US-capital company stuck multiple process nodes behind the leading edge, growing only ~2%, dependent on a captive protected market and government support, with the single most important input - advanced lithography - permanently constrained.
- No path to EUV/leading-edge foundry: capped at ~7nm DUV means Ascend and Kirin can't match Nvidia/TSMC on performance-per-watt or cost at volume
- Growth has collapsed from +22% to +2%; the easy smartphone-rebound comp is behind it and the core ICT business is nearly flat (+2.6%)
- Cloud external revenue is shrinking (-3.5% to RMB 32.16B) - it's losing the domestic AI-services race to Alibaba/ByteDance even at home
- Margins are structurally impaired (net ~8%) and unlikely to recover to pre-sanction levels while the supply chain stays internalized
- Concentration + geopolitical tail risk: a Taiwan escalation or a fresh export-control round could sever remaining equipment/component access overnight
- Zero access for US public-market investors - no equity, no ADR, no priced round; the value simply cannot be owned, and disclosure is management-controlled and unaudited by public-market standards
What it is worth
No public valuation exists - Huawei is 100% privately held, unlisted, with no ticker, no ADR, no priced external round, and no S-1/F-1. Any 'valuation' is a third-party estimate or an internal book-value mark, not a traded or transacted market figure. Internal virtual shares are non-transferable book-value units redeemable only within the union scheme, valued off audited net assets (total equity RMB 544.62B / ~$74.65B at end-2024).
Process-node ceiling and tightening export controls cap Ascend economics; growth stalls near flat, cloud keeps losing share, and the value stays locked inside an opaque, un-investable, geopolitically fragile structure worth little more than its book equity to anyone who cannot transact it.
A ~$122B-revenue, ~8%-net-margin, ~2%-growth diversified hardware conglomerate with ~$75B audited book equity, a protected but capped home market, and structurally impaired margins - a stable national champion, not a high-multiple growth story.
If domestic 7nm yields scale and Beijing fully walls off Nvidia, Huawei captures the dominant share of a multi-hundred-billion-RMB domestic AI-compute + telecom market - implying a franchise value far above its ~$75B book equity, comparable to a global tier-1 semis+telecom conglomerate, were it ever listable.
For US public-market investors Huawei is fundamentally un-investable - there is no security to own. Exposure to the Huawei theme is only obtainable indirectly: long the Western beneficiaries of its exclusion (ERIC, NOK for telecom share; NVDA as the accelerator it competes against). Listed Chinese ecosystem names (SMIC 0981.HK, Xiaomi 1810.HK, Seres 601127.SS) are named as context only, not a buy/own recommendation. This dossier is informational, not financial advice.
SWOT
Strengths
- World's #1 telecom-equipment vendor with deep carrier relationships across China, MEA, SE Asia, LatAm
- Massive, sustained R&D (RMB 192.3B/yr, ~22% of revenue — ~113,000-114,000 R&D staff) and a huge patent portfolio
- Vertical integration — designs its own SoCs (HiSilicon Kirin/Ascend), OS (HarmonyOS), cloud, and energy hardware - resilience against single-supplier cutoffs
- Only credible domestic alternative to Nvidia AI accelerators inside China's protected, policy-favored market
- Employee-ownership model insulates it from quarterly-earnings pressure and hostile takeover; long-horizon capital allocation
Weaknesses
- Cut off from leading-edge foundry (TSMC), EUV lithography (ASML), and top-tier EDA - stuck at ~7nm DUV, a multi-node gap vs TSMC/Samsung
- Structurally lower margins (gross ~44%, net ~8%) than a comparable non-sanctioned peer would earn, due to supply-chain internalization costs
- Largely locked out of US, UK, Australia, Japan, and parts of the EU telecom markets on national-security grounds
- Cloud external revenue declined 3.5% in 2025 as Chinese AI compute demand flowed to rivals; HarmonyOS app ecosystem still thin outside China
- Opaque financials, no independent public-market scrutiny, and persistent Western allegations of state ties limit trust in restricted markets
Opportunities
- China's national AI-compute self-sufficiency drive — Beijing steering domestic buyers toward Ascend as Nvidia access is throttled both ways
- HarmonyOS NEXT global rollout (2026) as a genuine third mobile OS
- 5G-Advanced/6G leadership and datacenter/energy infrastructure for the AI build-out
- Automotive: becoming the 'Bosch/Qualcomm of China EVs' via ADS and HiCar supply deals
- Global South infrastructure (Belt-and-Road-aligned telecom, cloud, solar) where US restrictions carry less weight
Threats
- Further US/allied export-control tightening — Taiwan blacklisted Huawei and SMIC in 2025; secondary-sanction risk on suppliers (TSMC probe, TechInsights teardown fallout)
- Yield and capacity ceilings on domestic ~7nm production without EUV - caps Ascend/Kirin volume and cost competitiveness
- Nvidia/AMD closing back into China via policy shifts would undercut the captive-market advantage
- Geopolitical escalation (Taiwan) could sever remaining component and equipment access
- Rising domestic competition (Xiaomi, Cambricon, Biren, Alibaba T-Head) for the same sovereign-compute budget
Moats, dependencies & bottlenecks
Moats
High in unrestricted markets, eroding in the West #1 global telecom-equipment share; multi-generation carrier relationships and 5G/5.5G IP
HiSilicon + HarmonyOS + cloud + Ascend stack - sanctions forced the moat wider by internalizing dependencies
Wide but policy-dependent Sovereign-compute mandate favors Ascend; durability hinges on Beijing's stance and Nvidia's China access
One of the largest 5G SEP holders; cross-licensing leverage and royalty income
Large HarmonyOS device base; strong China consumer brand, weak app ecosystem abroad
Dependencies
Supplier / manufacturing Sole viable leading-edge fab for Kirin/Ascend at ~7nm DUV; itself Entity-Listed and yield/capacity-constrained without EUV
Regulatory / demand Protected-market demand, R&D support, and 'buy-domestic' mandates underpin the AI-chip and telecom thesis
Regulatory (adversarial) Entity List since 2019; each tightening (TSMC, ASML, EDA, Taiwan blacklist) removes an input - the dominant swing factor
HBM substitutes) AI accelerators need HBM-class memory; domestic substitutes lag, constraining Ascend performance
Platform / network effect Global viability of the de-Androidized OS depends on third-party app support outside China
Advantages
- Only scaled domestic Nvidia-alternative (Ascend) in the world's second-largest AI market
- Full-stack control from silicon to OS to cloud - no single external chokepoint can shut it down
- 22%-of-revenue R&D funded by patient employee-owned capital, immune to earnings-cycle cuts
- Global #1 telecom-equipment franchise with recurring carrier revenue
- Fast-scaling, now-profitable automotive-supplier business riding China's EV boom
- Deep government alignment as the national tech champion
Weaknesses
- Multi-node process disadvantage vs TSMC/Samsung with no near-term fix
- Structurally depressed margins from supply-chain internalization
- Growth decelerated to ~2%; core ICT roughly flat (+2.6%)
- Cloud losing share at home (-3.5%); HarmonyOS ecosystem thin abroad
- Un-investable to public-market capital; opaque, management-controlled disclosure
- Concentrated geopolitical/regulatory tail risk
Bottlenecks
- No access to EUV lithography (ASML) - hard ceiling on process node and therefore AI-chip performance and cost
- SMIC yield and wafer capacity at ~7nm - throttles Ascend/Kirin volume regardless of design quality
- HBM / advanced-memory supply for AI accelerators - domestic sources immature
- Leading-edge EDA tools and IP libraries restricted - slows design iteration
- Western-market access (US/UK/AU/JP/EU) largely closed on security grounds
- App-ecosystem depth for HarmonyOS outside China
Top signals & trends
Top signals
Signals confidence in domestic fab/packaging capacity and captive AI demand
The sanction-rebound tailwind has largely played out
Losing the domestic AI-services race to Alibaba/ByteDance
Cost internalization maturing; profitability stabilizing
Supply-chain noose tightening from multiple directions
Bullish (long-term) · Sustained sovereignty investment; near-term margin drag
Ambitious third-OS bid; execution and app ecosystem unproven abroad
Trends
Beijing throttling Nvidia both ways steers demand to Ascend - Huawei's single biggest tailwind
Closes Western markets and inputs, but forged the domestic-champion moat; net-negative on inputs, net-positive on protected demand
Core infrastructure and energy franchises benefit globally ex-restricted markets
ADS/HiCar supplier model scaling fast (+72.1%), profitable
HarmonyOS gains at home; global adoption faces Android/iOS lock-in
Ecosystem & competitor graph
Suppliers feed the company; customers pull from it. Line thickness shows the strength of each tie (supply-chain dependency, customer earnings contribution). Hover to isolate a tie.
Domestic foundry fabbing Kirin/Ascend at ~7nm DUV; the critical, capacity-constrained bottleneck
Former leading-edge foundry - cut off by US export controls since 2020; the input Huawei lost
EUV/DUV lithography supplier - advanced tools blocked to Huawei/SMIC; the hard technology ceiling
Private/unlisted Chinese DRAM/NAND makers substituting for restricted HBM/memory; still lagging
Largest carrier customer for 5G/network gear
Major carrier buyer of infrastructure
Carrier customer for RAN/core
Cloud/AI player and reported prospective Ascend buyer despite being a cloud competitor
Reported prospective Ascend AI-chip buyer for datacenter compute
Private; reported prospective large-scale Ascend AI-chip buyers
MEA, SE Asia, LatAm telecom operators - core of Huawei's ex-Western infrastructure demand
Direct rival in enterprise/carrier networking; largely uncontested by Huawei in US/allied markets
Primary Western 5G RAN competitor; gains share in markets that exclude Huawei
Other major Western telecom-equipment vendor; direct 5G/optical rival
The AI-accelerator standard Huawei's Ascend is explicitly built to displace inside China; multi-node performance lead
Premium-smartphone rival in China; Huawei's Kirin comeback directly pressured iPhone share
Competes in smartphones, memory, and foundry; also a leading-edge fab Huawei cannot use
Mobile-SoC and modem rival; Huawei's HiSilicon self-supply reduces reliance
China smartphone + IoT + EV competitor chasing the same domestic consumer wallet
Datacenter/cloud networking competitor in enterprise switching
Alternative AI/datacenter silicon Huawei's Ascend competes against in China