The Cost of Going Head-to-Head With Big Tech

Today, I’m digging into Kingsoft Cloud (KC).

Kingsoft Cloud is a China-focused cloud infrastructure provider that sells compute, storage, and AI-ready cloud services.

Here are a few things that stood out to me.

• Kingsoft Cloud’s gross margins sit in the mid-teens, while Amazon Web Services runs closer to ~50%. Ouch!

• That gap is due to huge differences in scale and utilization. AWS is more than 100x larger.

• AWS spreads fixed infrastructure costs across a massive global customer base and layers high-margin software on top.

• Even with improving margins, the economics are capped unless pricing power or utilization meaningfully improves.

The main take away for me is this: Given that they're China-native, they can have relationships with companies that it isn't so easy for U.S. or Europe-based companies to service. But that competitive advantage is definitely not showing up in their margins.

I try not to focus much on stock prices, but their stock is down 74% over the last 5 years. They are not in an enviable position.

All in all, this is one of the more ugly situations I've come across. I think there's hope for them given their China-native business, but it's ugly to see their current margins, given those advantages.

With that, I'll see you tomorrow!

Nick

TL;DR

• Kingsoft Cloud is a China-focused cloud infrastructure provider split between public cloud services and enterprise cloud projects, with AI workloads driving recent growth.

• The company is showing real margin improvement and narrowing operating losses, but this progress is fueled by extremely heavy CapEx tied to AI infrastructure.

• Public cloud now represents the majority of revenue, while lower-margin enterprise and CDN services are being deliberately deprioritized.

• The core lesson for operators is that strategy shifts that require large CapEx are really financing decisions, demand certainty matters more than vision.

The 30,000-Foot View

Kingsoft Cloud operates a classic infrastructure-as-a-service model in China, selling compute, storage, networking, and related cloud services. Its business is reported in two segments: public cloud services, which are usage-based and increasingly AI-driven, and enterprise cloud services, which are more project-based and tied to delivery, customization, and integration for regulated or traditional customers.

The revenue mix has been steadily shifting toward public cloud, which carries better long-term scalability but requires massive upfront investment in data centers, GPUs, and networking equipment. This mix shift is intentional and central to management’s strategy.

Revenue mix (FY2024):

• Public cloud services: 64.3%

• Enterprise cloud services: 35.7%

Key Stats

• Market cap: $3.3B

• TTM Revenue: $1.29B

• TTM Gross Margin: 16%

• Employees: ~12,300

• Industry: Cloud computing and IT infrastructure services

Company History

• 2012: Company incorporated, cloud operations established.

• 2013–2015: Series A, B, and C financing rounds fuel early expansion.

• 2018: Series D financing raises ~$201M, supporting infrastructure buildout.

• 2020: IPO on Nasdaq.

• 2021: Acquisition of Camelot strengthens enterprise digital-solutions capability.

• 2022: Dual primary listing by introduction on the Hong Kong Exchange.

• 2023: CEO transition as company pivots toward margin discipline and AI focus.

• 2024–2025: Aggressive AI infrastructure investment, CDN retrenchment, margin recovery.

Show Me the Money

Standout financial features:

• Terrible gross margins.

• Operating losses narrowed significantly but still ugly.

• CapEx intensity surged due to AI investment.

• Net debt flipped from net-cash to leveraged as borrowing increased.

• Related-party balances add effective leverage beyond reported debt.

Financial Data

The N.O.O.B. Nine — Competitive Powers

The Nerd Out on Business Nine is made up of Hamliton Helmer's famous "7 Powers" of competitive advantage (Scale Economies, Network Economies, Counter-Positioning, Switching Costs, Branding, Cornered Resource, and Process Power) combined with two of my own (Data Flywheel and Distribution Advantage).

Average Score: 2.4/5 - Limited structural moat, with competitiveness driven more by execution than defensibility.

Memorable Marketing

Kingsoft Cloud’s marketing is pragmatic and B2B-focused, relying on ecosystem credibility, partner proof, and performance signaling rather than mass-market branding.

Xiaomi and Kingsoft Ecosystem Co-Sell (2025)

• Core idea: Borrow trust from the broader Kingsoft and Xiaomi ecosystem.

• Channels: partner sales, enterprise relationships

• Why it worked: Trust transfer reduced sales friction.

• Result: Management cited ecosystem-driven revenue expansion.

AI Gross Billing Disclosure (2025)

• Core idea: Report AI-specific commercial metrics to prove traction.

• Channels: earnings materials, sales decks

• Why it worked: Buyers want evidence of real AI workloads.

• Result: AI gross billing reached ~$104M in a quarter.

KS3 Extreme Speed with Solidigm (2024)

• Core idea: Turn infrastructure performance into a case study.

• Channels: partner marketing, case studies

• Why it worked: Benchmarks beat abstract claims.

• Result: Bandwidth improvements exceeding 100x cited in partner materials.

Low-Margin CDN Exit (2025)

• Core idea: Frame retrenchment as strategic focus.

• Channels: earnings calls, customer targeting

• Why it worked: Focus narrative attracts higher-quality demand.

• Result: Public cloud growth offset CDN decline.

Tactical takeaways:

1. Use partner trust to shorten enterprise sales cycles.

2. Publish one metric that proves momentum.

3. Replace adjectives with benchmarks.

4. Frame revenue exits as strategic upgrades.

AI Uses & Opportunities

Current uses:

• AI-driven public cloud growth and reporting.

• AI infrastructure buildout via GPU-heavy CapEx.

• AI-optimized object storage for training and inference workloads.

Future opportunities:

• AIOps to reduce outages and staffing costs.

• AI-powered FinOps tools sold as premium features.

• GPU capacity marketplaces to improve utilization.

• Vertical AI stacks for regulated industries.

• AI-driven credit risk and collections optimization.

Bumps in the Road

• Persistent losses despite improvement.

• Large impairment charges highlight prior capital misallocation.

• Rising leverage and related-party exposure.

• Extreme CapEx sensitivity to demand fluctuations.

• Intense competition from hyperscale Chinese cloud providers.

Your Swipe File

• Markets often dictate pricing.

• If your competitors have much more scale, you need to find unique differentiation.

• If you have competitive advantages related to regulation or your geography, they need to show up via better pricing.