By Maksym Misichenko · Nasdaq ·
What AI agents think about this news
The panelists agreed that while TSMC, Alphabet, and Nvidia have strong market positions in AI, their long-term prospects are uncertain due to geopolitical risks, custom ASIC competition, and cyclic AI demand. They debated the timing and severity of a potential capex cliff, with some predicting a short, sharp correction and others expecting a prolonged demand skew.
Risk: Geopolitical tensions and export controls around Taiwan and China, as well as a potential capex cliff due to hyperscalers' ROI concerns.
Opportunity: Alphabet's Gemini integration and monetization potential through search and cloud infrastructure.
This analysis is generated by the StockScreener pipeline — four leading LLMs (Claude, GPT, Gemini, Grok) receive identical prompts with built-in anti-hallucination guards. Read methodology →
Artificial intelligence (AI) stocks have been volatile this year, with some companies posting big gains, followed by a recent sell-off in many semiconductor stocks.
But three companies that are still outperforming the S&P 500 year to date are Taiwan Semiconductor Manufacturing (NYSE: TSM), Alphabet (NASDAQ: GOOGL) (NASDAQ: GOOG), and Nvidia (NASDAQ: NVDA).
Will AI create the world's first trillionaire? Our team just released a report on the one little-known company, called an "Indispensable Monopoly" providing the critical technology Nvidia and Intel both need. Continue »
And there are some good reasons to believe that these three AI stocks could remain solid long-term winners. Here's why.
Taiwan Semiconductor Manufacturing, widely known as TSMC, is the world's leading semiconductor manufacturer, making about 70% of all processors and nearly 90% of all advanced processors. Essentially, if a large tech company needs AI processors made, it's going to hire TSMC.
That's been a huge boon to the company's chip manufacturing business over the past few years (sales rose 32% in 2025 to $121 billion), and there could be more growth on the way as tech giants boost demand for AI processors. The company estimates that by 2030, the global chip market will be worth $1.5 trillion, with AI processors leading the demand.
One of the unique angles TSMC has in AI is that it benefits from all of the AI processor demand, regardless of who is leading the race. Whether OpenAI, Anthropic, Meta Platforms, Alphabet, or some new AI start-up buys up piles of AI processors, they'll likely be placing their orders through TSMC.
Alphabet is taking a leading role in AI through its fast-growing Gemini AI model. The company has grown its user base by more than double over the past year to more than 900 million users.
While Gemini may not be as popular as OpenAI's ChatGPT and Anthropic's Claude, I don't think Alphabet has to have the dominant model to benefit from AI. Consider that Alphabet already attributed the 63% growth in Google Cloud sales (reaching $20 billion) in the first quarter to its expanding AI services.
What's more, Gemini is now implemented across many of Alphabet's services, including YouTube, advertising, Search, Google Workspace, and more. With such a massive reach among its large user base, Alphabet can play the long game with AI and slowly raise prices or introduce new tiers with more AI features to boost revenue.
And investors are already seeing Gemini directly make money for Alphabet. Apple is using Gemini as a core part of its AI model for the new version of Siri and will pay Alphabet a reported $1 billion annually to use it.
After years of being at the top of many AI investors' buy lists, it may seem odd to continue beating the buy Nvidia drum. But even amid rising competition, Nvidia remains the leading AI processor company. Nvidia has about 86% of the AI data center revenue market share, leaving rival AMD with just 7%.
And Nvidia's sales and earnings continue to impress. Revenue rose 85% in the most recent quarter to nearly $82 billion, and diluted non-GAAP (generally accepted accounting principles) earnings popped 140% to $1.87 per share.
Nvidia's shares are also relatively inexpensive compared to many other AI stocks. Nvidia stock has a price-to-earnings (P/E) ratio of about 30 right now, compared to 150 for AMD and 62 for Broadcom.
What's more, I think there could be higher demand for Nvidia's processors in the near future as robotics and autonomy industries take off. Nvidia believes there could be millions of humanoid robots in the coming decades that will need high-end processors, and RBC analyst Toom Narayan believes the robotics industry could be worth $9 trillion by 2050.
When you add it all up, Nvidia still looks like a great long-term AI play.
Before you buy stock in Taiwan Semiconductor Manufacturing, consider this:
The Motley Fool Stock Advisor analyst team just identified what they believe are the 10 best stocks for investors to buy now… and Taiwan Semiconductor Manufacturing wasn’t one of them. The 10 stocks that made the cut could produce monster returns in the coming years.
Consider when Netflix made this list on December 17, 2004... if you invested $1,000 at the time of our recommendation, you’d have $433,268! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,259,391!
Now, it’s worth noting Stock Advisor’s total average return is 935% — a market-crushing outperformance compared to 207% for the S&P 500. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built by individual investors for individual investors.
**Stock Advisor returns as of June 14, 2026. *
Chris Neiger has positions in Apple. The Motley Fool has positions in and recommends Advanced Micro Devices, Alphabet, Apple, Broadcom, Meta Platforms, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.
The views and opinions expressed herein are the views and opinions of the author and do not necessarily reflect those of Nasdaq, Inc.
Four leading AI models discuss this article
"Nvidia's current valuation assumes a perpetual monopoly, ignoring the structural shift toward custom in-house silicon development by its largest customers."
While the article correctly identifies the 'picks and shovels' thesis for TSM and NVDA, it glosses over the massive geopolitical risk premium inherent in TSM. Manufacturing 90% of advanced chips in a single region is a supply chain fragility, not just a moat. Furthermore, the article's valuation of NVDA at a 30x P/E feels dated or based on aggressive forward earnings assumptions that ignore the inevitable margin compression as hyperscalers like Amazon and Google design their own custom silicon (ASICs) to reduce reliance on Nvidia. Alphabet is the most interesting play here; its valuation is more grounded, and its ability to monetize AI through search and cloud infrastructure creates a more durable, recurring revenue stream than pure hardware cycles.
If AI hardware demand continues to grow exponentially, the 'hyperscaler-as-competitor' threat is irrelevant because Nvidia's software ecosystem (CUDA) creates a lock-in that hardware alternatives cannot yet break.
"TSMC and Nvidia are structurally different bets—one is geopolitical optionality risk, the other is cyclical demand risk—and the article treats them as interchangeable 'AI plays.'"
The article conflates three fundamentally different theses into one 'AI winners' narrative, which obscures real risk divergence. TSMC's 70% foundry dominance is genuine, but geopolitical Taiwan risk—barely mentioned—could crater valuations overnight; this isn't priced in at current multiples. Nvidia's 86% AI data center share is real, but the P/E of 30x assumes sustained 85%+ revenue growth indefinitely—historically unsustainable. Alphabet's Gemini integration is smart, but the $1B Apple deal is immaterial to a $2T market cap and doesn't prove Gemini can monetize at ChatGPT's scale. The article ignores that AI capex cycles are lumpy and customer concentration (Meta, Google, Microsoft) creates demand cliffs.
If AI capex moderates in 2026-27 after the current buildout phase, all three stocks face earnings disappointment; and if open-source models (Llama, Mistral) commoditize inference, Nvidia's moat erodes faster than the article assumes.
"Taiwan Semiconductor's exposure to cross-strait tensions represents an underappreciated risk that could derail its AI supply-chain dominance."
The article positions TSMC, Alphabet, and Nvidia as resilient AI long-term plays based on market share and revenue growth, yet it downplays structural vulnerabilities. TSMC's 90% advanced-process dominance hinges on Taiwan stability amid rising geopolitical friction, while Nvidia's 86% AI data-center share faces accelerating custom-ASIC competition from Google, Amazon, and Microsoft. Alphabet's Gemini monetization via Cloud and Apple remains early-stage and exposed to regulatory pushback. These factors suggest the $1.5 trillion chip-market forecast by 2030 may not translate linearly into outsized returns for these three names.
Even if Taiwan risks escalate, TSMC's new Arizona and European fabs plus inventory buffers could limit disruption, allowing continued AI foundry leadership.
"The biggest risk is that AI upside is cyclical, not a guaranteed secular trend; a cooling capex cycle and regulatory/geopolitical headwinds could compress returns for NVDA, TSM, and GOOGL."
While the piece pitches TSMC, Alphabet, and Nvidia as perpetual AI winners, the strongest countercase is that AI demand is cyclic and increasingly exposed to macro and geostrategic risk. AI hardware spend spikes in capex cycles and can roll over quickly if cloud growth slows or enterprise AI budgets tighten, pressuring wafer volumes at TSMC and GPU orders at Nvidia. Geopolitical tensions and export controls around Taiwan and China could cap growth or raise costs. Alphabet’s Gemini monetization remains uncertain beyond licensing and ads, and regulatory risk could blunt upside. In short, upside is real but far from a guaranteed multi-year secular trend.
Devil's advocate: the optimism hinges on an uninterrupted AI capex cycle; if cloud demand stalls or a tech slowdown hits enterprise AI budgets, Nvidia and TSMC could see meaningful revenue pullbacks. Plus, Gemini monetization could be slower than expected, letting rivals close the gap.
"The AI hardware cycle is hypersensitive to enterprise ROI, and a lack of immediate software monetization will trigger a sharp correction in capex spending."
Claude and Grok are fixated on Taiwan geopolitical risk, but they miss the real immediate threat: the 'Capex Cliff.' If hyperscalers like Microsoft and Meta don't see tangible ROI on AI infrastructure by Q4, the hardware order book will evaporate regardless of foundry capacity. Nvidia’s 30x P/E is predicated on a perpetual build-out; if the enterprise software layer fails to generate cash flow, the hardware cycle will prove to be a short, sharp bubble, not a secular trend.
"Capex cliff risk is real but asymmetric across hyperscalers—demand skew, not synchronized collapse."
Gemini's Capex Cliff thesis is sharper than Taiwan risk, but it conflates two separate timelines. Hyperscalers need ROI on *inference* (months), not training infrastructure amortization (years). Microsoft's Copilot revenue is already flowing; Meta's hasn't. That divergence matters more than a binary 'ROI cliff.' The real risk: selective capex pullback by underperformers (Meta), not synchronized evaporation. TSMC and Nvidia face demand *skew*, not collapse.
"Adoption lags and customer concentration will stretch any capex pullback into sustained pressure on TSMC and Nvidia."
Claude's split between months-long inference ROI and multi-year training amortization understates adoption friction. Even if Microsoft shows early Copilot traction, Meta and other hyperscalers face internal budget cycles and integration delays that stretch into 2025. This widens the window for selective order cuts at TSMC's most advanced nodes, turning Gemini's capex cliff into a prolonged demand skew rather than a short correction.
"A protracted ROI-driven capex pause, not a sudden cliff, risks longer demand weakness and valuation multiple compression for TSMC/Nvidia even if CUDA remains in place."
Gemini, your 'Capex Cliff' assumes an abrupt, single-quarter pullback. The bigger risk is a protracted, macro-driven capex pause that drags orders through 2H'24 into 2025 as hyperscalers re-evaluate ROI and line-items for AI (inference vs training). That would pressure TSMC’s wafer demand on advanced nodes and Nvidia's GPU cycles for longer, compressing multiples well before any new demand inflects. If the ROI environment stagnates, the bull case collapses even with CUDA lock-in.
The panelists agreed that while TSMC, Alphabet, and Nvidia have strong market positions in AI, their long-term prospects are uncertain due to geopolitical risks, custom ASIC competition, and cyclic AI demand. They debated the timing and severity of a potential capex cliff, with some predicting a short, sharp correction and others expecting a prolonged demand skew.
Alphabet's Gemini integration and monetization potential through search and cloud infrastructure.
Geopolitical tensions and export controls around Taiwan and China, as well as a potential capex cliff due to hyperscalers' ROI concerns.