By Victor ([@vcmktasa](https://x.com/@vcmktasa)) · Mr. Z ([@168MrZ](https://x.com/@168MrZ))
In early June 2026, the COMPUTEX Taipei event is in full swing. Jensen Huang has declared Marvell as the next trillion-dollar company, causing the optical interconnect sector to surge. Simultaneously, the macro environment is far from peaceful; the Strait of Hormuz has been blocked for 100 days, oil prices remain stuck around $90, and the market is holding its breath for the May CPI release on June 13. Meanwhile, SpaceX is about to go public with a valuation of approximately $1.75 trillion, and Anthropic has secretly filed its documents. The liquidity landscape for the second half of the year is being reshuffled.
At this intersection of hardware fever and macro variables, 168X invited Frank (@qinbafrank), one of the few investors who can simultaneously track macroeconomics, US tech stocks, the AI supply chain, Crypto, and global capital flows. Frank’s "top-down, cross-market" framework, honed through mobile internet, entrepreneurship, and VC experience, provided a comprehensive judgment in this two-hour dialogue: He does not believe AI is a bubble; he believes the entire commercialization process has just passed the monetization inflection point and is only halfway through. What truly warrants caution are "reset-style" tail risks to macro logic. He starts with the three great debates of Nvidia, dismantling penetration dividends, the capital expenditure war, optical interconnects and CPO, Nokia and edge computing, the three capital logics of scarcity/upgrade/long-term, and finally the liquidity shock of the SpaceX IPO and the future of Bitcoin.
👉
Global capital is pouring into the semiconductor industry, with Taiwan stocks at the forefront. The 168X team is based in Taipei, and we have opened a
Taiwan-US Stock Research Exchange Group (Group 1 filled up instantly; this is a special additional group)
to continue sharing assets worth watching and cutting-edge information. Welcome to join!
This article is a summary of the 168X (@168X_Fortune) program, a top-tier dialogue platform deeply connecting Eastern wisdom with Western innovation, focusing on AI, blockchain, robotics, space tech, and bioengineering to explore how technology, capital, and human wisdom will reshape the future of human civilization.
Hosts: Mr. Z (@168MrZ) · Victor (@vcmktasa) & Guest: Frank (@qinbafrank
Listen to the interview: Original X Space Audio | YouTube Version (Better Quality)
Table of Contents
- 1. Research Framework: Cross-market investment path from mobile internet to VC
- 2. Three Great AI Debates: Using Nvidia's stock price to verify trends, computing power, and profitability
- 3. Commercialization Inflection Point and CapEx War: Why penetration dividends determine this bull market
- 4. The Era of Limited Easing: Liquidity, interest rates, and stock selection logic in a "structured bull market"
- 5. Differentiation in Software Stocks: Which are replaced by AI and which are strengthened
- 6. Killing Valuations, Earnings, and Logic: Three scenarios for individual stock corrections
- 7. Optical Interconnects and CPO: How connectivity becomes the next AI main line
- 8. Nokia, Edge Computing, and Physical AI: Extending from the cloud to the physical world
- 9. Scarcity, Upgrades, Long-term: Dismantling the three logics of AI capital rotation
- 10. SpaceX IPO and the Liquidity Black Hole: The impact of three giants going public
- 11. Large, Medium, and Small Corrections: Only a "reset of macro logic" is a true crash
- 12. Bitcoin and the New Crypto Normal: Differentiation, capital paradigms, and AI integration
- 13. Advice for Listeners: Balance offense and defense, dive deep into fundamentals
1. Research Framework: Cross-market investment path from mobile internet to VC
Mr. Z: Frank is one of the few top investors on Twitter who understands macroeconomics, US tech, the AI supply chain, Crypto, and global capital flows. Frank, please introduce yourself and tell us how this market research framework was formed.
Frank: Mr. Z mentioned "top investor," but I really don't reach that level; I'm just an ordinary investor. There's an old Chinese saying, "A full bucket makes no sound, but a half-full bucket sloshes around." That describes me—I know a bit about everything, but I might not be an expert in any single area.
I usually invest in US stocks and Crypto, mainly in the secondary market, while sharing my thoughts on macro, industry, and individual stock logic on X. My career is divided into two parts: over ten years ago, I worked in China's mobile internet industry for seven or eight years as a product manager, started my own business, raised several rounds of funding (though it didn't succeed), and then transitioned to VC. I entered the crypto space in late 2017/early 2018, continuing my primary market VC mindset to invest in projects—first called ICOs, then private placements. Around 2021-2022, as Chinese policies tightened, I stopped primary market business and moved to secondary market investments. These years, I've been an individual investor.
This framework is actually a continuation of my previous work. My first job after graduation was in the internet sector, just as the world was transitioning from PC to mobile. Working at big Chinese firms, starting my own business, and doing VC gave me a foundation: you have to dismantle an industry's development status, trends, and dividends from an industrial perspective. This naturally helped me understand the pros and cons of various business types and commercial models in the tech industry.
I opened my first US stock account around 2013. To be honest, early on, I just rode the industry dividend without truly understanding the market. What really forced me to understand the market were the successive crashes I experienced after entering crypto: the 2018 trade war and Fed rate hikes, the 2020 COVID black swan, and the subsequent emergency rate cuts and unlimited QE.
That wave made me think: Why is this happening? So in 2020, I read almost every book I could find on Fed history, monetary finance, central banks, and asset allocation. I wanted to understand the core logic. Over those two years, my macro cognitive framework slowly formed. Whether looking at RWA or AI, I see them through the same lens: treat them as tech industries and see where the business is turning and which paths are viable.
2. Three Great AI Debates: Using Nvidia's stock price to verify trends, computing power, and profitability
Mr. Z: The AI craze in US, Taiwan, and Korean stocks feels like the 2020 liquidity surge. But the Fed hasn't cut rates. How do you see the current market? Where is the AI money now?
Frank: My view is that today is different from 2020. 2020-2021 was the end of a golden era driven by 0% interest rates and unlimited QE. Today is an era of "limited easing." I predicted in mid-2024 that the Fed would maintain limited easing for 2-3 years: rates between 3% and 3.5%, likely no QE, but small-scale balance sheet expansion as needed. In this environment, assets without fundamentals suffer.
AI is different because it's a massive, recognized trend. I wrote about the "Three Great AI Debates" last year. You can see it in Nvidia's stock price, which is divided into three stages.
First stage (late 2022 to late 2023): Is AI a trend? ChatGPT 3.5 excited the market, but the semiconductor industry hadn't recovered yet, and earnings hadn't been released. It was pure speculation. Nvidia stagnated for about six months in late 2023.
Second stage (early 2024 to Q1 2025): Do we need this much computing power? The turning point was the Davos Forum in January 2024 and the release of Sora. More importantly, Nvidia's earnings began to release on a massive scale. Even when DeepSeek shook the US tech world with low training costs in early 2025, the logic held because model parameters and data richness required even more power.
Third stage (Q4 2025 to Q2 2026): Can we make money from this CapEx? In February, I wrote about the "CapEx War." I judged that the Q1 earnings season in April would be critical. Big cloud providers (Microsoft, Amazon, Google) delivered explosive reports, moving the market from "doubt" to "verification." This is why the market has risen so fast since April: we finally saw the data. Anthropic's ARR grew from $30 billion in March to $45 billion in May. This was the monetization inflection point.
3. Commercialization Inflection Point and CapEx War: Why penetration dividends determine this bull market
Frank: How far has AI developed? First, look at penetration. In tech, 10% is the inflection point. Once you pass 10%, adoption accelerates rapidly. US enterprise AI procurement penetration was about 9.7% last year and reached 18% by March/April this year. We are in a high-growth stage.
Second, look at commercialization. Demand is growing "parabolically," as Jensen Huang says. Big tech companies are accelerating CapEx. Google recently announced $80 billion in financing. They are doubling down.
However, parabolic growth is unsustainable; there will be adjustments. My view is that we are in the middle of a long-term cycle, but macro risks remain.
Mr. Z: US liquidity is shrinking. Why hasn't the AI sector cooled down? How long can this last?
Frank: First, define this: the current US stock market is not a comprehensive bull market; it is a "structured" market. Unlike 2020, not every sector is rising. In crypto, only BTC, SOL, and BNB hit new highs. In US stocks, the leaders are AI and semiconductors. To rise, you need either expectations or earnings.
4. The Era of Limited Easing: Liquidity, interest rates, and stock selection logic in a "structured bull market"
Frank: In a limited easing environment, once the market overextends, there will be a correction. The key to judging a correction's level is whether AI commercialization is slowing down. As long as cloud revenue exceeds expectations, the business logic remains intact. Even if the market drops 20-30% due to overvaluation, it will recover once new evidence of growth appears.
People compare this to the 2000 dot-com bubble. There are similarities in the parabolic rise, but the difference lies in penetration and business model maturity. In 1999, internet penetration was low and business models (ads, e-commerce) weren't found until 2002-2006. Today, AI faces 4-5 billion smartphone users. A revolutionary technology can complete the mobile internet's decade-long journey in 3-5 years. This means corrections will be smaller because the infrastructure and monetization are already there.
5. Differentiation in Software Stocks: Which are replaced by AI and which are strengthened
Victor: Software stocks have been volatile. How do you see their logic?
Frank: Software will differentiate. General-purpose software that agents or LLMs can do without third-party tools is in danger.
However, three types are strengthened:
- Vertical software with deep know-how: LLMs are trained on public data, but private enterprise data is a moat. Companies like MongoDB or AI-native data containers are growing.
- Software-hardware integrated companies: Like Cloudflare. You can't replace physical CDN nodes with AI. As agents increase, the need for low-latency physical infrastructure grows.
- Deep industry software: Manufacturing or pharma data is proprietary. LLMs can't access it.
6. Killing Valuations, Earnings, and Logic: Three scenarios for individual stock corrections
Frank: Corrections happen for three reasons:
- Killing Valuations: The company is good, but the price got too high. It drops until it's attractive again.
- Killing Earnings: Not necessarily losses, but "growth missing expectations." If the market expects 50% growth and you give 48%, you get punished.
- Killing Logic: This is fatal. Like Meta in 2022 when the Metaverse pivot seemed to fail, or if a company is kicked out of Nvidia's supply chain.
7. Optical Interconnects and CPO: How connectivity becomes the next AI main line
Victor: Marvell surged after Jensen Huang called it a trillion-dollar company. How do you see the optical sector?
Frank: Connectivity is becoming more important because agent-era computing is distributed. Marvell's CEO Matt Murphy talked about the "distance-less data center" where storage and compute are pooled and connected by light.
But the optical sector will differentiate. Some, like Innolight, are seeing real earnings. Marvell's switch business is also growing. However, CPO (Co-Packaged Optics) is still new. Cloud providers might not deploy CPO at scale until 2026-2027 because of stability concerns. For now, 1.6T and 3.2T pluggable modules are the focus. The core indicator is when cloud providers start massive CPO switch procurement for rack-to-rack interconnects.
8. Nokia, Edge Computing, and Physical AI: Extending from the cloud to the physical world
Victor: You mentioned Nokia is in the optical and edge AI space. How do you see it?
Frank: Nokia's optical connectivity business is already delivering. They are a system-level supplier. Edge computing is currently a "valuation option." Nvidia is the conductor here; they recently separated edge computing in their financial reports.
Physical AI is the advanced form of edge computing: AI entering the real world via bots that perceive and act. Nokia's AI-RAN (turning base stations into mini data centers) is interesting but early.
I also tracked BlackBerry. It's a "distressed reversal" play. They partnered with Nvidia to integrate QNX into autonomous driving platforms. It's a bet on the underlying security infrastructure for Physical AI.
9. Scarcity, Upgrades, Long-term: Dismantling the three logics of AI capital rotation
Frank: Capital rotation follows three logics:
- Scarcity Logic: First GPU, then HBM, now storage and CPU. In the agentic era, CPU tasks (scheduling/orchestration) are more important. I'm bullish on ARM architecture here.
- Upgrade Logic: Optical interconnects (pluggable to CPO), data center power (800V HVDC), and advanced packaging (3D stacking).
- Long-term Logic: Edge computing and Physical AI. This depends on the mass production of robots like Tesla's Optimus.
10. SpaceX IPO and the Liquidity Black Hole: The impact of three giants going public
Mr. Z: Will the SpaceX IPO cause a liquidity black hole?
Frank: It will have an impact. It's a $1.75 trillion valuation seeking $75 billion. Institutions might sell other positions to participate. SpaceX has unique lock-up rules and will be added to indices quickly, forcing passive funds to buy.
With Anthropic and OpenAI also likely to IPO, we might see a liquidity-driven correction in the second half of the year. If this coincides with sticky inflation (CPI), the correction could be medium-sized. But as long as AI revenue growth doesn't stall, it won't be a 2000-style crash.
11. Large, Medium, and Small Corrections: Only a "reset of macro logic" is a true crash
Frank:
- Small corrections (single digits): Valuation adjustments or minor macro noise.
- Medium corrections (~15%): Major macro shocks, like the 2023 bond yield surge or the 2024 Yen carry trade unwind.
- Large corrections (25%+): These "reset the macro logic." Examples include the 2008 financial crisis, the 2020 pandemic, or the 2022 inflation/rate hike shock.
For a large correction now, we'd need a total failure of AI commercialization or a collapse of the US-led global order.
12. Bitcoin and the New Crypto Normal: Differentiation, capital paradigms, and AI integration
Frank: Crypto has entered a "new normal."
- Differentiation: As the market matures with ETFs, capital seeks certainty (BTC) and growth. Altcoins without fundamentals are ignored.
- Capital Paradigm: Crypto is a "capital paradigm," not a "productivity tool" like AI. Its 10x efficiency is in asset issuance and 24/7 global liquidity.
- AI Integration: The future of crypto lies in RWA (Real World Assets) and AI. Bots and agents will need smart contracts and crypto for payments (Agent Finance).
On Bitcoin's outlook: We are in a period of uncertainty, but BTC has proven value. Sub-$70k is a good accumulation zone, even if we don't know the exact bottom.
13. Advice for Listeners: Balance offense and defense, dive deep into fundamentals
Frank: First, balance offense and defense. You can't predict everything, so leave room to maneuver. Second, dive into business fundamentals. Understand a company's competitiveness in the supply chain. Buy when it's at a value price, and trim when it's overvalued.
Mr. Z & Victor: Thank you, Frank, for sharing so generously. If you enjoyed this, follow 168X on X and YouTube. See you next time.
👉
Taiwan-US Stock Research Exchange Group (Group 1 full; special Group 2 open)
About 168X
168X is a platform connecting Eastern wisdom and Western innovation, focusing on AI, blockchain, and space tech. Hosted by Mr. Z and Victor.
