AI: China's answer to NVIDIA trades at 550x P/E ratio

Hi Fintech Futurists —

Today’s agenda is below. We look into China’s answer to NVIDIA as the chip cold war heats up.

1. AI: China Doesn't Need NVIDIA When They Have Cambricon Technologies

2. LONG TAKE: Rewiring the Money Stack with MetaMask Dollars

3. PODCAST: Building the $130B global payments platform, with Airwallex’s Ravi Adusumilli

4. CURATED UPDATES: Machine Models, AI Applications in Finance & Investment Outlook

To support this writing and access our full archive of IPO primers, financial analyses, and guides to building in the Fintech & DeFi industries, see subscription options here. Our current price point is only $2/week.

Thanks for your time and attention,

Lex Sokolin

Subscribe to the Fintech Blueprint >>>

In Partnership: Angel & Accelerator Virtual Conference

We are big fans of these events as they bring together very large groups of people focused on startups and investing. The Angel & Accelerator Virtual Conference takes place on Sept 25 on Zoom. Hosted by Max Pog.

• 2 tracks for investors and founders, 20+ speakers, 500+ investors.

• 3.5h Zoom sessions + 1.5h Virtual Networking Zone, and optional 1:1 pitch calls.

• Speakers include leaders from Techstars, Plug and Play Ventures, Samsung Ventures America, and many others.

👉👉👉 Register for free here

We arranged 25% off on optional add-ons: Super Pitching Ticket for startups and Super Networking Ticket for investors and others.

Use code FINTECHBLUEPRINT25 at checkout.

Sessions focus on investor decision making and diligence, syndicate vs VC strategy, dependable deal flow using AI and networks, fundraising in 2025 from pre-seed to Series A, and clear messaging that converts meetings, with ROI takeaways for investors and founders.

👉👉👉 More speakers and topics here

AI Analysis: The Numbers Paint a Dramatic Picture

The number “554” is where we would like to start this discussion.

That’s Cambricon Technologies' current P/E ratio. This makes the Chinese AI chip designer one of the most expensive stocks relative to earnings in major global markets. For context, NVIDIA trades at around 50 times earnings.

This ratio seems absurd, but has grown to this point off the back of some miraculous financial figures. Cambricon posted a 1.03B yuan ($144MM) profit in the first half of 2025, reversing a 533MM yuan loss from the same period a year earlier. The revenue surge is even more dramatic, a staggering 43x year-on-year increase to 2.88B yuan ($403MM).

Why is there such hysteria surrounding this company?

China's Mysterious Answer to Nvidia's H100

Cambricon is famous for its MLU (Machine Learning Unit) accelerator chip family. Public specs indicate a lineup oriented around cloud inference with training-capable variants:

• MLU370 (cloud inference/training)

• MLU290 (accelerators)

Cambricon is betting its future on the Siyuan 690 processor, which is designed to rival Nvidia's H100. The previous generation, the Siyuan 590, reportedly delivers about 80% of Nvidia's A100 performance using China's domestic 7nm process.

One glaring issue though, is that there are no publicly available specs for this new processor. In an industry where benchmarks are everything, Cambricon's silence is concerning.

However, the company's planned 4B yuan ($560MM) capital raise announced in July, suggests there is a serious product to rally behind.

The Rise of Cambricon Technologies

2024 was a standout year for Cambricon. It finished the year as one of the best performing AI stocks globally.

The sentiment towards the business started to build against the backdrop of three key political trends:

1. The global AI arms race

2. U.S. and China technological conflict

3. China's increased state-directed capitalism

This has continued into 2025 and has picked up even more steam.

The catalyst for this latest drive can be traced back to an earlier trendsetter in the AI space, DeepSeek. When DeepSeek launched into public view at the start of the year, it triggered a rush for the domestic chips to power further Chinese breakthroughs.

DeepSeek's latest V3.1 model was even explicitly "tailored to work with next-generation homegrown AI chips", sending a clear signal to the domestic Chinese market.

We can see how Cambricon share price has followed Chinese AI sentiment through 2025.

The Sanctions Paradox

When the U.S. added Cambricon to its Entity List on December 15, 2022, cutting off access to advanced American technologies, it was meant to cripple the company. Instead, it became Cambricon's greatest competitive advantage.

Over the last 2 years, Beijing has pressured domestic companies to use homegrown alternatives. As a result, reports indicate ByteDance and Tencent have started to deploy local alternatives like Cambricon for LLM inference. That matters because inference (not training) tends to dominate unit economics for AI-company P&Ls in consumer apps and fintech workflows.

Political Influence in US AI Chip Market

This has also come at a time where the US is wrestling with competition within its own domestic chips market. NVIDIA has been the winner in AI chip manufacturing — but Trump’s administration has tried to keep Intel competitive. Last week the US government even announced a $5.7B investment into Intel.

The 94% Problem

It is not clear-cut that Cambricon will succeed long-term.

94% of Cambricon’s revenue comes from just five customers, with the largest (likely Alibaba) contributing nearly 80%.

Cambricon saw its revenue crater in the past when Huawei terminated their partnership in 2018 to develop its own chips. At that time, Huawei represented 98% of Cambricon’s revenue, and when it replaced Cambricon with its own technology, Cambricon became a direct competitor of its former client.

We saw similar customer concentration when we covered CoreWeave a few months back ahead of their IPO.

AI: Is CoreWeave’s $35B IPO an AI Hyperscaler or a GPU Debt Trap?

Read full story

Since then, they have managed to effectively diversify their customer base — so there is hope for Cambricon. Cambricon is also still only a small share of the Chinese AI chip market, representing roughly 3% of the domestic market. Some analysts sketch a path to 11% by 2028 as the ecosystem matures.

But Cambricon is trying to reign in speculation. On August 28, the company issued a rare warning to investors about its doubling share price, dispelling talk about "non-existent products in the pipeline" and reminding investors it "labours under US sanctions."

The stock immediately plunged almost 9%, the most since April, while the broader market stood largely unchanged.

The Bottom Line

Traditional valuation models break down when prices are set by geopolitics, tariffs, and supply crunches.

The question isn't whether Cambricon is worth $80B — it's whether you believe China can build a completely independent AI ecosystem and whether Beijing can continue support Cambricon into the future as an NVIDIA competitor. Cambricon has expanded support from major domestic AI models, including DeepSeek, Alibaba's Qwen, and Tencent's Hunyuan.

If this ecosystem flourishes, today's valuations might look cheap. If it fails, this will be remembered as the moment when nationalism triumphed over investor rationality.

👑Related Coverage👑

Analysis: Our 2025 Market Overview & Investment Outlook

Lex Sokolin

Long Take: Now is the time to learn Generative AI, not after the knowledge worker layoffs

Read full story

Fintech Blueprint 🤖🏦🧭 is a reader-supported publication. To receive new posts and support my work, consider becoming a free or paid subscriber.

Subscribe to the Fintech Blueprint >>>

Blueprint Deep Dive

Analysis: Rewiring the money stack with MetaMask Dollars (link here)

Money, I often joke, is the most complex mass-adopted product there is.

Everyone uses it, yet very few understand what powers that abstraction in practice. The separation between hard measurable reality and the idea of value has been simmering for so long that most of us have stopped noticing it. Most people ignore, or are uninterested in, the reality: that behind the dollar balance we see on a PDF statement, or in Venmo, Revolut, or a stablecoin wallet, rather than some measurable certainty, there is simply a system of checks and balances designed to ensure that, most of the time — but not always, the money (or value) will be there when we ask for it.

Read Long Take

🎙️ Podcast: Building the $130B global payments platform, with Airwallex’s Ravi Adusumilli (link here)

In this episode, Lex speaks with Ravi Adusumilli - President and GM of the Americas at Airwallex. Ravi and Lex discuss how Airwallex has evolved into a global financial platform by offering businesses an integrated suite of cross-border payments, treasury, and banking services. Founded in 2015, Airwallex now supports 150,000 customers, processes $130 billion in annualized volume (up 73% YoY), and projects a $1 billion revenue run rate by year-end.

Listen To Podcast

Curated Updates

Here are the rest of the updates hitting our radar.

Machine Models

Ethical and Bias Considerations in Artificial Intelligence/Machine Learning - Matthew G. Hanna & Liron Pantanowitz & Brian Jackson & Octavia Palmer & Shyam Visweswaran & Joshua Pantanowitz & Mustafa Deebajah & Hooman H. Rashidi

A Critical Field Guide for Working with Machine Learning Datasets - Sarah Ciston & Mike Ananny & Kate Crawford

AI Applications in Finance

⭐ AI-Driven Payment Systems: From Innovation To Market Success - Merve Ozkurt Bas

The Rise Of Generative Ai Agents In Finance: Operational Disruption And Strategic Evolution - Inesh Hettiarachchi

Financial Modeling in Corporate Strategy: A Review of AI Applications For Investment Optimization - Olufunmilayo Ogunwole & Ekene Cynthia Onukwulu & Micah Oghale Joel & Ejuma Martha Adaga & Augustine Ifeanyi Ibeh

Investment Outlook

⭐ Private Equity Outlook 2025: Is a Recovery Starting to Take Shape? - Bain & Company

⭐ Global Venture Capital Outlook: The Latest Trends - Bain & Company

⭐ Global Private Markets Report 2025: Braced for shifting weather - McKinsey & Company

🚀 Postscript

• Sponsor the Fintech Blueprint and reach over 200,000 professionals. 👉 Reach out here.

• Check out our new AI products newsletter, Future Blueprint. (Don’t tell anyone)

• Read our Disclaimer here — this newsletter does not provide investment advice

• Contributors: Lex, Laurence, Matt, Farhad, Daniel, Michiel, Luke

• For access to all our premium content and archives, consider supporting us with a subscription. In addition to receiving our free newsletters, you will get access to all Long Takes with a deep, comprehensive analysis of Fintech, Web3, and AI topics, and our archive of in-depth write-ups covering the hottest fintech and DeFi companies.