Hi friends,

💡 Investing Doesn’t Have to Be Complicated, R2R is built on simple, time-tested principles:

From Sep 28,2022, through Jan 9,2026

No speculation. Just disciplined portfolio housekeeping.

This week, I conducted a comprehensive analysis on TSM, Google, and Meta, leading to a decision to add exposure to Meta and TSM. My portfolio weightings have shifted slightly to reflect this conviction.

Latest R2R Research Published:

On Wednesday, January 14, CEO Elon Musk announced that Tesla will stop selling the Full Self-Driving (FSD) capability as an upfront purchase (previously ~$8,000) effective February 14, 2026. Moving forward, FSD will be available exclusively as a monthly subscription, priced at $99 per month.

My Take:

This is a pivotal business model pivot. By killing the large upfront cash payment, Tesla is betting entirely on the Recurring Revenue (SaaS) model.

It lowers the barrier to entry. More drivers might try FSD at $99/month than would commit $8,000 upfront, potentially increasing the total size of the data-gathering fleet. It also smoothes out revenue, making quarterly earnings less "lumpy."

In the short term, this will likely hit cash flow and margins. Recognizing $8k instantly is very different from collecting $99 over 7 years. The market needs to see if the "take rate" (percentage of owners subscribing) jumps significantly to offset the loss of upfront capital.

U.S. safety regulators (NHTSA) granted Tesla a five-week extension, until February 23, 2026, to respond to an ongoing probe regarding FSD safety and alleged traffic law violations. The investigation, which covers nearly 2.9 million vehicles, was triggered by reports of FSD-enabled cars running red lights and entering lanes into oncoming traffic.

My Take:

Regulatory overhang remains the single biggest "wildcard" risk for the stock right now.

The extension is neutral-to-positive; it buys the legal team time, but it doesn't resolve the underlying issue.

The core fear here isn't just a fine, it’s a potential forced recall or a "pause" on FSD features. If regulators force Tesla to "nerf" the software to be ultra-conservative, it could degrade the user experience and hurt the new subscription-only strategy mentioned above. Watch the February 23 deadline closely.

Elon Musk posted that AI5’s design is “almost done,” that they’re aiming for a ~9-month design cycle, and that AI6 is in early design stages (with additional forward-looking chip mentions).

My Take
Tesla’s in-house silicon matters if autonomy is the endgame: better inference performance per watt and tighter vertical integration can be a durable advantage. The key for me is whether these chip cycles translate into observable capability jumps (safety + reliability + cost to deploy) rather than just roadmap velocity. Hardware progress is only valuable if it measurably compounds software capability and fleet economics.

On Wednesday, January 14, the Bureau of Labor Statistics released the CPI data for December 2025. Headline inflation came in slightly hotter than expected at 0.3% month-over-month, keeping the annual rate stuck above the Fed's 2% target. Following the report, futures markets reduced the probability of a March 2026 rate cut from 70% to roughly 40%.

My:

This is the "Good News is Bad News" dynamic returning.

On Thursday, January 15, Tencent repurchased 1 million shares on the open market for a total consideration of approximately HK$635.6 million. This implies an average purchase price of roughly HK$635.60 per share. This action continues the company's long-running buyback program, executed despite the stock trading significantly higher than its mid-2025 levels (approx. HK$517 in May 2025).

My Take:

This is a massive signal of confidence. Usually, companies slow down buybacks when their stock price rallies ~20% in six months, but Tencent is accelerating.

Management effectively stated this week that they believe the stock is still undervalued at HK$635. They are using their cash pile to defend this price level.

Tencent is transitioning from a "growth at all costs" company to a "shareholder yield" machine. For long-term holders, this reduces the float and supports earnings-per-share (EPS) even if top-line growth is modest.

Reports emerged this week highlighting the diverging strategies between Tencent and its arch-rival ByteDance regarding AI hardware. While ByteDance launched updates for its "Doubao" AI assistant and is aggressively pursuing AI glasses (expected Q1 2026), Tencent has opted for a partnership model, teaming up with AR maker INMO, rather than building its own device.

My Take:

This makes me nervous. We are seeing the next platform war unfold: "Super Apps" (WeChat) vs. "Super Devices" (AI Glasses).

If AI glasses eventually replace smartphones as the primary daily interface, ByteDance is positioning itself to own the hardware, whereas Tencent is just an app on someone else's hardware.

This reminds me of the early smartphone era. Tencent won mobile with WeChat, but if they miss the hardware transition, their "moat" could be breached by ByteDance owning the user's field of view.

On Wednesday, January 14, details surfaced regarding new U.S. export controls that place a "volume cap" on sales of advanced AI chips (specifically Nvidia's H200s) to China. The regulations limit the total compute capacity China can import. Additionally, the Department of Defense continues to label Tencent a "Chinese Military Company," complicating the compliance certifications required to import even the permitted chips.

My Take:

This is the single biggest cap on Tencent's Cloud and AI potential.

You can write the best code in the world (Hunyuan model), but if you don't have the silicon (H200s) to run it, you lose.

If Tencent is starved of chips, their cloud service becomes a "second-tier" offering compared to AWS or Azure. This forces them to rely on domestic chips, which are currently less efficient, hurting their margins and slowing down their AI training cycles.

Meta began laying off more than 1,000 employees from Reality Labs, representing roughly 10% of the division's 15,000-person workforce. The restructuring redirects resources from virtual reality and metaverse products toward AI wearables and phone features. Multiple VR studios closed, including Armature Studio, Twisted Pixel, and Sanzaru. Supernatural, a VR fitness app acquired for $400 million in 2023, was moved into maintenance mode with a skeleton crew and no new content development . Reality Labs has reportedly lost $73 billion over five years, and spending was cut by 30% as part of broader cost management.

My Take:

This restructuring is the most concrete evidence to date that Meta's metaverse bet, once positioned as the company's defining multi-year initiative, has been fundamentally deprioritized. The 30% budget cut to Reality Labs signals that management views the metaverse as a failed investment with limited near-term commercial potential. The $73 billion cumulative loss is staggering and represents capital that cannot be recovered. For investors evaluating Meta's capital allocation discipline, this pullback demonstrates management's willingness to acknowledge past misjudgments and reallocate resources toward areas perceived as more immediately productive. However, it also raises questions about how many other multi-billion dollar bets within Meta carry similar risk profiles with uncertain returns. The decision to put Supernatural into maintenance mode after acquiring it for $400 million three years ago exemplifies the human and financial cost of misaligned strategic bets.

Meta announced landmark agreements to extend and expand three nuclear power plants and support development of advanced nuclear technology in partnership with Vistra, TerraPower, and Oklo. These agreements make Meta one of the most significant corporate purchasers of nuclear energy in American history. The agreements support construction of multiple Oklo Aurora Powerhouse reactors and will create thousands of construction and long-term operations jobs in Ohio and Pennsylvania. The company states that it will pay the full costs for energy used by its data centers.

My Take:

The nuclear energy commitment is a critical and revealing signal about Meta's confidence in its AI trajectory and the scale of infrastructure investment required. Nuclear power provides the firm, reliable baseload electricity that hyperscale data center operations demand. By securing long-term partnerships and making multi-year commitments, Meta is essentially betting that its AI infrastructure needs will justify sustained high-capacity energy consumption for years to come. This is a vote of confidence in both the demand for Meta's future AI applications and in the stability of its business model to support such capital outlays. However, it also reveals operational reality: Meta's estimated 2026 capex is $106-119 billion, with $121 billion in non-cancelable commitments. The nuclear energy partnerships underscore that this spending is not speculative, it is locked in and committed. The company is essentially mortgaging future cash flows to current infrastructure investments. If the AI monetization thesis disappoints, Meta will still bear the full cost of this energy infrastructure.

Meta announced at CES 2026 that the Ray-Ban Meta Display glasses (priced at $799) are experiencing unprecedented demand with waitlists extending well into 2026. Due to limited inventory, Meta paused planned international expansion to the UK, France, Italy, and Canada originally scheduled for early 2026 to focus on fulfilling US orders . The company released software updates including teleprompter functionality, pedestrian navigation in 32 cities, and neural handwriting capabilities through the Meta Neural Band wristband, which allows EMG-based surface gesture control .

My Take:

The strong demand for Ray-Ban Display glasses validates Meta's wearables strategy and provides near-term revenue momentum in a hardware category that was previously speculative. The fact that the company is pausing international expansion due to inventory constraints suggests either production capacity limitations or genuine consumer appetite exceeding supply expectations. Both scenarios are operationally positive, they indicate the product is resonating rather than struggling with demand. The neural band integration and teleprompter features demonstrate product-market iteration and expanding use cases. However, this remains a small-scale business relative to Meta's core advertising platform. EssilorLuxottica's CFO stated the company will reach 10 million unit capacity earlier than originally planned by the end of 2026 , but even at that volume, wearables revenue would be modest relative to Meta's $164 billion total revenue base. The wearables segment is a strategic hedge against future platform risk and demonstrates technological competence, but it is not yet a material revenue contributor.

Meta acquired Manus, an AI agent startup, in a deal reportedly valued at over $2 billion. Manus reportedly has approximately $100 million in annual recurring revenue . Meta is shifting away from open-source LLaMA models toward closed-source Avocado and Mango models for commercialization, with the Avocado model scheduled for release in Q1 2026 .

My Take:

The Manus acquisition represents Meta's bet on agentic AI, a category where autonomous AI agents perform complex tasks on behalf of users. At a $2 billion price for $100 million in annual recurring revenue, the deal implies a 20x revenue multiple, which is elevated but not unusual for high-growth AI infrastructure companies. However, the strategic pivot from open-source LLaMA to proprietary closed-source models (Avocado and Mango) is more significant. Open-source models helped establish Meta's presence in the AI ecosystem and generated goodwill in the developer community. The shift to proprietary models suggests management believes the company can monetize proprietary AI capabilities more effectively than the industry-benefitting model that comes with open-source distribution. This also indicates competitive intensity with OpenAI and Google, if Meta cannot build proprietary defensibility around AI models, it risks becoming a commoditized infrastructure provider. The Avocado rollout in Q1 2026 will be closely watched by investors as a test of whether Meta's AI capabilities can compete with leading closed-source models from OpenAI and Anthropic.

On Monday, January 12, Meta announced the appointment of Dina Powell McCormick as President and Vice Chairman. Powell McCormick, a former Goldman Sachs executive and Deputy National Security Advisor, will oversee the company's overall strategy, global affairs, and execution, reporting directly to Mark Zuckerberg.

My Take:

This is a calculated "Political Hedge." Zuckerberg is bringing in a heavy hitter with deep ties to the current administration and global finance. By creating a "President" role for her, he is effectively outsourcing the geopolitical and regulatory shield wall. With antitrust scrutiny high and AI regulation looming, having a seasoned diplomat/banker at the top allows Zuckerberg to focus purely on product and AGI (Artificial General Intelligence). This reduces "Key Man Risk" regarding political slip-ups.

On Thursday, January 15, TSMC released Q4 2025 results that exceeded Wall Street expectations across the board.

My Take:

The most impressive number here isn't the revenue, it’s the Gross Margin (62.3%).

Usually, when a manufacturer ramps up new technology (like 3nm), margins initially dip due to yield issues. TSMC is defying gravity here. They are essentially charging a "monopoly tax" on AI chips because customers like Nvidia and Apple have nowhere else to go.

The guide for Q1 implies the "AI supercycle" hasn't even peaked yet. If they can maintain 60%+ margins while growing revenue at 25%, the valuation (even at all-time highs) looks reasonable compared to software peers.

During the earnings call, TSMC announced it is raising its 2026 Capital Expenditure (CapEx) budget to between $52 billion and $56 billion. This is a significant jump from the ~$41 billion spent in 2025 and is well above analyst projections of ~$48 billion. Roughly 70-80% of this budget is allocated to advanced process technologies (2nm and A16).

My Take:

This is the loudest "Buy Signal" for the entire AI sector we've heard in months.

TSMC is notoriously conservative. They do not spend $56 billion unless they have firm orders locked in from customers. This effectively confirms that the "Big Tech" spending spree (Microsoft, Meta, Google) will continue through 2027.

This CapEx number single-handedly lifted the entire semi-equipment sector (Applied Materials, ASML) on Friday. It tells me the "infrastructure build-out" phase of AI is far from over.

Management confirmed that 2nm (N2) chip production officially reached high-volume manufacturing (HVM) in late 2025 and is ramping up faster than 3nm did. Additionally, they introduced the roadmap for A16 (1.6nm) chips, scheduled for volume production in the second half of 2026. The company noted that 2nm demand is so high that it will generate more revenue in its first year than 3nm or 5nm did.

My Take:

The competitive gap is widening, not closing.

Reports confirm Samsung and Intel are still struggling with yields on comparable nodes. TSMC is effectively the only shop in town for the next generation of AI accelerators (Nvidia’s Rubin, Apple’s M6/A20).

Because they are the sole supplier for 2nm, TSMC can dictate pricing. This protects them from the "commoditization" risks that usually plague hardware manufacturers.

Reports surfaced late in the week (Friday, Jan 16) that the U.S. and Taiwan are nearing a trade agreement that could lower tariffs on Taiwanese exports to 15% (down from 20%). In exchange, TSMC is reportedly discussing expanding its U.S. footprint further, potentially adding up to five more fabs in Arizona over the next decade.

My Take:

This turns a liability into an asset.

The biggest risk to TSMC has always been "China invasion risk." By deeply integrating its manufacturing roadmap with U.S. domestic policy (building fabs in Arizona), TSMC makes itself indispensable to the U.S. military-industrial complex.

Building in the U.S. is expensive and hurts margins (see Point #1). However, if this trade deal subsidizes those costs or lowers tariffs, it neutralizes the margin drag. It’s a geopolitical win-win.

Stay patient. Stay focused. Turn volatility into cash flow. Let compounding do the heavy lifting.

—-William | Relax to Rich Club

⚠️Disclaimer

I am not a licensed financial advisor. The information presented here is for educational purposes only and does not constitute investment advice. Past performance is not indicative of future results. Please review the full legal disclaimer here: