TheRaccoonAnalysis

This week highlights three separate themes: execution, diversification, and geopolitics. Tesla’s Austin robotaxi launch is a meaningful proof-of-concept that strengthens its AI and autonomy narrative, but a 20-vehicle deployment alone is unlikely to justify a major valuation rerating without evidence of safe, scalable expansion. Intel’s AI chip win and NVIDIA’s memory partnership suggest hyperscalers are broadening supply chains rather than relying on a single vendor, potentially creating opportunities across the AI ecosystem while preserving NVIDIA’s leadership. Meanwhile, restrictions on major Chinese firms underscore rising geopolitical risk, creating both valuation pressure and uncertainty for China-focused investors

The combination of a potential Strait of Hormuz disruption and stronger-than-expected U.S. payrolls creates a challenging backdrop: higher energy costs could fuel inflation while a resilient labour market may reinforce expectations for tighter monetary policy. For Singapore investors, the key question is whether the oil shock proves temporary or persistent.

The divergence between a record-high Dow and weakness in semiconductors suggests a rotation rather than a definitive breakdown in the AI theme. After an extended rally, profit-taking in high-beta AI names is unsurprising, while capital rotating into financials and value sectors reflects a broadening market leadership profile. Rising layoffs, softer hiring intentions and Bitcoin weakness point to pockets of economic caution ahead of payroll data, but do not yet imply a sharp deterioration in macro conditions. Ciena’s negative reaction despite strong results highlights elevated expectations in AI-linked hardware, whereas Blackstone’s strength reflects investor preference for earnings visibility and cash-flow resilience.

Broadcom’s post-earnings selloff highlights how elevated expectations have become across AI infrastructure. Strong revenue growth alone may no longer justify premium valuations when investors are pricing in near-perfect execution and sustained acceleration. If this pattern persists, AI-related hardware names could face greater sensitivity to guidance, margins, and capital expenditure trends rather than headline growth.

SpaceX’s planned debut at a valuation significantly above some fair-value estimates underscores IPO timing risk. While long-term fundamentals may remain attractive, initial trading could be driven more by sentiment and scarcity value than fundamentals, increasing the likelihood of heightened volatility and valuation recalibration.

Marvell Technology is emerging as one of the most important beneficiaries of the AI infrastructure buildout. While much investor attention remains focused on GPUs, Marvell operates in a critical segment of the AI ecosystem: networking, connectivity, and custom silicon that enable massive AI clusters to function efficiently.

As hyperscalers deploy increasingly larger AI systems, networking has become a strategic bottleneck. Thousands of GPUs must exchange data at extremely high speeds, making advanced networking silicon nearly as essential as the compute hardware itself. This positioning was highlighted when Jensen Huang specifically referenced Marvell at COMPUTEX, signaling industry recognition of its role in next-generation AI infrastructure.

The company’s investment appeal rests on several key drivers:

* Growing demand for AI networking and optical connectivity.

* Increasing adoption of custom AI ASICs by hyperscalers.

* Expanding AI-related revenue contribution.

* Long-term AI data centre capital expenditure growth.

Unlike many AI software companies whose valuations are driven largely by future expectations, Marvell’s growth is tied to physical infrastructure deployment, providing a more tangible connection to AI spending trends.

Key risks include customer concentration among a small number of hyperscalers, competition from larger rivals such as Broadcom, and the possibility that AI infrastructure spending moderates after the current investment cycle.

Overall, Marvell appears well-positioned as a critical “picks-and-shovels” provider for the AI era. The central question for investors is no longer whether the company will benefit from AI, but whether its future growth can exceed the increasingly ambitious expectations reflected in its valuation. If AI infrastructure investment remains strong over the next several years, Marvell has the potential to become one of the core AI infrastructure holdings alongside NVIDIA and Micron Technology.

$Marvell Tech(MRVL.US)

The AI infrastructure trade is broadening beyond GPUs. Marvell’s prominence in custom AI networking silicon underscores how critical interconnects have become for scaling hyperscale AI clusters, but a 32% one-day move likely pulls forward a significant portion of near-term optimism. Future upside may depend more on sustained design wins and execution than narrative expansion alone.

For SpaceX, the gap between targeted IPO valuation and secondary-market pricing suggests elevated expectations. Investors may weigh the company’s exceptional assets against the possibility of post-listing price discovery and volatility.

Anthropic’s potential IPO would likely be judged less on headline valuation and more on evidence of durable economics. Investors may look for a multi-billion-dollar revenue run rate, strong gross margins, high enterprise retention, and a credible path toward operating leverage before justifying an aggressive valuation.

For Alphabet, the negative reaction reflects near-term dilution concerns and uncertainty over AI returns. However, history suggests that large infrastructure investments can create significant long-term competitive advantages if demand materialises as expected.

The market narrative appears increasingly bifurcated. NVIDIA’s China exposure may be diminishing as a percentage of the investment case, but tighter export restrictions could still weigh on near-term revenue visibility and inventory planning. Berkshire’s acquisition of Taylor Morrison is likely less a rate call and more a long-duration wager on persistent US housing undersupply and demographic demand. Meanwhile, the S&P 500 reaching new highs despite weak sector breadth suggests leadership remains concentrated in a handful of mega-cap names. While not a definitive warning signal, narrow rallies historically warrant greater attention to valuation, earnings execution, and market participation trends.

Anthropic’s implied valuation versus OpenAI reflects investor willingness to price future AI infrastructure dominance before normalized earnings exist. The risk resembles prior liquidity-driven cycles, but unlike earlier tech bubbles, frontier AI firms are already generating meaningful enterprise revenue. Sustainability likely depends less on model hype and more on monetization durability, inference economics, and regulatory outcomes.

In China EVs, Li Auto appears to be prioritizing margin protection while XPENG is leaning into scale and share capture. H2 2026 may favor whichever strategy better converts software adoption and pricing power into sustainable cash flow amid intensifying competition.

If Hormuz supply risks ease materially, markets may rotate into lower energy-input beneficiaries such as airlines, chemicals, logistics and consumer sectors, while crude volatility and oil-linked equities could soften. However, positioning still depends on OPEC responses and broader geopolitical durability.

PDD’s revenue mix shift suggests Chinese e-commerce platforms are increasingly monetising fulfilment, logistics and transaction ecosystems over advertising alone, potentially pressuring peers like Alibaba Group and JD.com to deepen integrated commerce and merchant-service strategies.

Huawei’s “τ-law” may reflect genuine engineering progress, but timing near export-control reviews suggests strategic signaling as well. SMIC’s sharp rally looks vulnerable to profit-taking unless operational evidence confirms sustainable margin and yield improvements.

Singapore’s softer core CPI strengthens expectations that Monetary Authority of Singapore could adopt a more accommodative stance. That generally supports Singapore REITs through lower financing pressure, while banks may face modest net interest margin compression if rates trend lower.

Trump’s quantum package may accelerate funding, talent retention and procurement pipelines, but current valuations likely reflect policy momentum more than near-term commercial adoption. Quantum remains technologically promising yet commercially early-stage. Intuit’s sharp selloff highlights investor concern that AI could compress software margins and reduce headcount-dependent growth models, though markets may also be underestimating how incumbents can monetise AI productivity at scale. For C6L, Air India remains a high-risk strategic bet: if restructuring succeeds, long-term network value could emerge, but continued losses may increasingly weigh on capital allocation and shareholder confidence.

Singapore’s export surge suggests AI-linked hardware demand remains intact despite sharp corrections in storage names, implying markets may be repricing margins and inventory risk rather than end-demand collapse. A selective rotation into Singapore semi-test and precision engineering plays like AEM Holdings, UMS Holdings, and Frencken Group may offer cleaner exposure to AI infrastructure buildouts.

The Google-Blackstone JV, Nvidia’s agent-focused CPU push, and Meta’s workforce realignment reinforce a consolidation phase favouring hyperscale infrastructure, data centres, networking, and enterprise AI integration. In Singapore, Keppel DC REIT, Singtel, and ST Engineering could benefit through rising AI compute, connectivity, and sovereign AI demand.