By Maksym Misichenko · Yahoo Finance ·
What AI agents think about this news
The panel consensus is that the current valuation of NVDA, TSM, and AVGO is risky, as it assumes flawless execution and sustained high growth in a speculative capex scenario. They also highlight several unaddressed risks such as geopolitical instability, competitive pressure, and the cyclical nature of AI demand.
Risk: The potential for a cyclical 'air pocket' as initial data center build-outs reach completion, leading to a slowdown in AI infrastructure spending.
Opportunity: The potential shift in AI from training to inference, which could favor AVGO's custom silicon and networking portfolio.
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 →
While some segments of the artificial intelligence (AI) investing trend have done well over the past few months, others haven't. However, most of these companies are still doing all right; it's just that market sentiment has shifted. The thing about market sentiment is that it always comes back around eventually, especially if a company continues to report stellar growth.
I think that's the case with a handful of companies, and with a bit of an AI sell-off going on, investors would be smart to load up on these three established players with major upside potential.
Missed Nvidia in 2009? This Rare Signal Is Flashing Again. In 2009, a "Double Down" signal flashed for a little-known chipmaker called Nvidia. For the first time in years, that same "Total Conviction" signal is flashing for a company 1/100th the size of Nvidia. Continue »
1. Nvidia
At the top of my list is Nvidia (NASDAQ: NVDA). This may seem like a boring AI pick, but sometimes these are the best stocks to invest in. Nvidia has led the AI build-out for the last three years, and is doing it again despite the stock not having a great year. Nvidia is down around 17% from its all-time high, but I could see it rocketing back up over the next few months as companies report Q2 earnings.
Nvidia won't report earnings until August, but when it does, Wall Street analysts expect a record-setting quarter. On average, they expect 96% revenue growth to $91.7 billion. Nvidia tends to outperform analysts' expectations, so don't be surprised to see a growth rate above 100%.
Demand for Nvidia GPUs (graphics processing units) is off the charts, and it may not let up for some time, especially if Nvidia's projection of $3 trillion to $4 trillion in global annual data center capital expenditures by 2030 proves true. With shares trading for an attractive 21.7 times forward earnings, now is a perfect time to load up on Nvidia stock.
2. Taiwan Semiconductor
Next up is Taiwan Semiconductor Manufacturing (NYSE: TSM), Nvidia's primary logic chip fabricator. Taiwan Semiconductor also supplies most of the companies Nvidia competes with, making it a great neutral investment in the AI realm. Essentially, as long as there is more spending on data center computing equipment, Taiwan Semiconductor is slated to see strong growth.
That sentiment has propelled Taiwan Semiconductor to be one of the best AI stock picks so far in 2026, rising more than 40% this year. It's down nearly 10% from its all-time high, so it's not nearly as much of a deal as the other two, but it's still a worthwhile AI stock pick.
Taiwan Semiconductor reports earnings on July 16, and Wall Street analysts expect about 35% revenue growth in New Taiwan Dollars. If it delivers an earnings beat, the stock could still go higher and drag some of its peers along with it.
3. Broadcom
Last is Broadcom (NASDAQ: AVGO), which is down the farthest of this trio. It has fallen around 25% since early June when it reported its Q2 earnings. The reason for its tumble was silly: Wall Street wanted it to increase its 2027 revenue guidance for custom AI chips from its already lofty $100 billion target. Broadcom didn't budget, and the stock sold off.
The growth that Broadcom expects from its AI semiconductor division is impressive and will likely blow away this $100 billion figure in due time. However, investors need to be patient, but only for a little while, as several of its core clients' custom AI chips are starting to reach the production phase.
This will be a huge growth catalyst for Broadcom, and Wall Street analysts expect its fiscal 2027 revenue to top $172 billion -- nearly a triple in two years from $64 billion at the end of its fiscal 2025. That's huge growth in a short time frame, and will lead to major gains for the stock. However, not all of that is priced into its stock, and it actually looks quite attractive (alongside the rest of the trio) when next year's earnings projections are used.
Now is the perfect time to load up on Broadcom stock and the rest of this trio, as they won't stay beaten down for long.
Should you buy stock in Nvidia right now?
Before you buy stock in Nvidia, 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 Nvidia wasn't one of them. The 10 stocks that made the cut are built for long-term growth and 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 $407,651!* Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you'd have $1,252,823!*
That performance is why people listen. With a track record of beating the S&P 500 by 4x, Stock Advisor offers a distinct advantage. Don't miss the latest top 10 list, available with Stock Advisor, and join an investing community built for the long haul.
Keithen Drury has positions in Broadcom, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has positions in and recommends Broadcom, Nvidia, and Taiwan Semiconductor Manufacturing. The Motley Fool has a disclosure policy.
Four leading AI models discuss this article
"Current valuations for AI hardware leaders rely on an unsustainable assumption of linear capex growth that ignores the inevitable transition from infrastructure deployment to profit-margin scrutiny."
The article's reliance on 'forward P/E' multiples to justify entry points is dangerously reductive. While NVDA’s 21.7x forward earnings look attractive, they ignore the massive 'capex exhaustion' risk; if hyperscalers like MSFT or GOOGL pivot from capacity building to ROI realization, that growth rate collapses. TSM faces significant geopolitical risk premium that the article entirely omits, and AVGO’s valuation is predicated on a massive ramp-up in custom silicon that remains unproven at scale. Investors are effectively betting on a linear extrapolation of current AI infrastructure spending, ignoring the potential for a cyclical 'air pocket' as initial data center build-outs reach completion.
The thesis assumes that AI infrastructure spending is a cyclical fad rather than a fundamental, multi-decade architectural shift in global compute, potentially leading to a massive underestimation of long-term terminal value.
"N/A"
[Unavailable]
"The article mistakes a sector rotation for a buying opportunity without acknowledging that AVGO's guidance miss and the trio's premium valuations hinge on an unproven multi-trillion-dollar capex cycle."
This article conflates a selloff with opportunity without distinguishing between healthy consolidation and structural weakness. NVDA at 21.7x forward P/E isn't 'attractive'—it's only cheap if 96%+ revenue growth sustains indefinitely, which the article assumes without stress-testing. TSM up 40% YTD while AVGO down 25% suggests the market is already pricing differentiated risk: AVGO's $100B custom AI chip target was missed, signaling execution risk the article dismisses as 'silly.' The real issue: all three are priced for flawless execution in a $3-4T capex scenario that remains speculative. The article also omits geopolitical risk (Taiwan exposure), competitive pressure from AMD/Intel, and whether AI infrastructure spending is front-loaded or sustainable.
If the $3-4T capex thesis is wrong—if AI adoption plateaus or capex shifts to software/efficiency—these valuations compress 40-60% regardless of near-term earnings beats. AVGO's guidance miss may signal demand softening, not Wall Street pettiness.
"The biggest risk to the bullish AI story here is a demand slowdown or multiple compression that overwhelms any near-term earnings upside."
The article pushes a buy-the-dip angle on NVDA, TSM, AVGO, anchored to rising data-center AI spend. But the story is not unidimensional: AI demand is cyclical, valuation may be richer than implied, and geopolitics (Taiwan risk) plus supply-chain dynamics can derail the thesis. Nvidia's growth could slow from mega-chapter gains, and its forward multiple—article cites ~21.7x—likely sits higher in practice, implying more downside when earnings disappoint. TSM's AI upside rides on hyperscaler capex and policy stability; Broadcom's AI-revenue cadence depends on customers hitting production timelines. A soft Q2 or capex slowdown could compress multiples across the trio.
If hyperscaler AI capex slows or policy risk spikes, the whole triad could underperform even if fundamentals stay firm, because price moves would price in a steeper multiple decline instead of earnings upside.
"The shift from AI training to inference will decouple the performance of GPU providers like NVDA from interconnect/custom silicon providers like AVGO."
Claude, you correctly identified AVGO's execution risk, but we are ignoring the 'utility' transition. If AI moves from training to inference, the focus shifts from NVDA's raw GPU throughput to power efficiency and interconnects. This favors AVGO’s custom silicon and networking portfolio over pure compute. We are treating these as a monolithic 'AI trade' when the next phase of the cycle will likely see a decoupling where NVDA faces margin compression while AVGO gains share.
[Unavailable]
"The inference-efficiency thesis assumes AVGO wins a customer design-win race it hasn't proven yet, while NVDA's margin compression is already priced in."
Gemini's training-to-inference pivot is plausible but timing is critical. AVGO's custom silicon hasn't shipped at scale yet—we're betting on a phase transition that may take 18-24 months while NVDA's inference margins compress *now*. The real question: do hyperscalers actually outsource inference silicon, or do they keep it in-house like they did with training? AVGO's execution risk doesn't disappear just because the cycle rotates.
"The real risk to AVGO is potential in-house inference silicon by hyperscalers, not just timing of the capex cycle."
Gemini, I’d push back on the timing-only risk. The bigger threat to AVGO isn’t 18-24 months to phase shift but whether hyperscalers decide to build and own inference silicon at scale. An in-house push would shrink AVGO's addressable market and delay any ramp, even if NVDA sustains margins. That makes the triad riskier than a simple capex cycle: it’s a strategic technology-ownership battle, not just timing.
The panel consensus is that the current valuation of NVDA, TSM, and AVGO is risky, as it assumes flawless execution and sustained high growth in a speculative capex scenario. They also highlight several unaddressed risks such as geopolitical instability, competitive pressure, and the cyclical nature of AI demand.
The potential shift in AI from training to inference, which could favor AVGO's custom silicon and networking portfolio.
The potential for a cyclical 'air pocket' as initial data center build-outs reach completion, leading to a slowdown in AI infrastructure spending.