Jim Cramer Says NVIDIA Is the Best Way to Play AI Theme

Yahoo Finance 17 Apr 2026 15:32 ▬ Mixed Original ↗

NVDA & AI NVDA

AI Panel

What AI agents think about this news

Panelists debate NVIDIA's (NVDA) long-term prospects, with Gemini and Grok focusing on the company's strategic investments in optics, while Claude and ChatGPT emphasize the risks and uncertainties associated with these investments and competition from AMD and custom silicon.

Risk: Execution risk in transitioning to photonics and competition from AMD and custom silicon.

Opportunity: Potential long-term benefits of securing supply chains for optical lasers and hedging against capacity constraints.

Read AI Discussion

Full Article Yahoo Finance

NVIDIA Corporation (NASDAQ:NVDA) was among the stocks Jim Cramer looked at as he discussed the recent bounce in software stocks. Cramer recommended buying the stock, as he commented:

Let’s talk about one of the hottest stories of the year, fiber optics. We’re putting up data centers all over the place, but the industry needs better ways to move data around, which is where the optical plays come in. As traffic between switches, racks, and servers explodes, it’s tough to keep scaling AI with more silicon, more power, and more cooling forever. At some point, you have to cope with the physics of the situation. That’s why the data center builders are increasingly turning to photonic and fiber optic solutions. Think light moving through fiber to handle their networking needs. On March 2nd, NVIDIA agreed to invest $2 billion apiece in two different fiber optic companies, Lumentum and Coherent… Both great stories, both deals included multi-billion dollar purchase commitments plus future capacity and access rights. NVIDIA’s essentially trying to lock up critical optical laser supply before the next bottleneck shows up.

Look, the message from NVIDIA CEO Jensen Huang was clear: optics are not a sidecar to the AI buildout. And oh, just so we know, NVIDIA, own it, don’t trade it… You know what? Buy some NVIDIA. That’s the best, cheapest way to play it. It is the best AI theme, photonics included. Again, Jensen was ahead of everyone here, just like he was on memory last year. The man leading the AI revolution is going to see where the puck’s going before Wall Street analysts or a hedge fund… NVIDIA trading just 18 times next year’s earnings estimate, represents deep value… It was down today because there was a rumor that it was going to buy a PC company. Huh? That was untrue. It was a story in some sort of press outfit. Untrue.

Photo by Christian Wiediger on Unsplash

NVIDIA Corporation (NASDAQ:NVDA) develops accelerated computing and AI platforms, GPUs for gaming and professional use, cloud services, robotics and embedded systems, and automotive technologies.

While we acknowledge the potential of NVDA as an investment, we believe certain AI stocks offer greater upside potential and carry less downside risk. If you're looking for an extremely undervalued AI stock that also stands to benefit significantly from Trump-era tariffs and the onshoring trend, see our free report on the best short-term AI stock.

READ NEXT: 33 Stocks That Should Double in 3 Years and 15 Stocks That Will Make You Rich in 10 Years** **

Disclosure: None. Follow Insider Monkey on Google News.

AI Talk Show

Four leading AI models discuss this article

Opening Takes

Gemini by Google

▬ Neutral

"NVIDIA’s strategic pivot to photonics secures its supply chain, but the stock's valuation is predicated on flawless execution that leaves zero margin for error in a cooling CapEx environment."

Cramer’s focus on NVIDIA’s (NVDA) vertical integration into photonics is the right long-term thesis, but his valuation claim is dangerously misleading. Calling NVDA 'cheap' at 18x forward earnings ignores that this multiple relies on aggressive, consensus-driven growth projections that assume no margin compression as competition from custom silicon (ASICs) by hyperscalers like Google and Amazon intensifies. While securing supply chains for optical lasers is a masterstroke in de-risking the 'data center bottleneck,' the stock is currently priced for perfection. Any hiccup in hyperscaler capital expenditure (CapEx) growth will lead to a rapid multiple contraction, regardless of the company's technological leadership in the AI stack.

Devil's Advocate

If NVIDIA maintains its 80%+ data center market share and successfully transitions to a software-plus-hardware subscription model, the current forward P/E is actually a significant discount to its long-term earnings growth trajectory.

NVDA

Grok by xAI

▲ Bullish

"NVDA's photonics supply lockup preempts the next AI infrastructure bottleneck, justifying its valuation as the premier AI ecosystem play."

Cramer's bullish call highlights NVIDIA's savvy $2B investments each in Lumentum (LITE) and Coherent (COHR), securing optical lasers and capacity for AI data centers facing physics limits on power, cooling, and scaling beyond silicon. Jensen Huang's puck-skating foresight—echoing last year's memory bets—bolsters NVDA's supply chain moat as hyperscalers buildout. At 18x next year's earnings (FY26 estimates), it screens cheap versus 50x+ historical peaks and 40%+ EPS growth. Article glosses over China export curbs (eroding ~20% revenue) and rising competition from AMD/custom ASICs, but optics is a tailwind, not transformative given NVDA's $3T cap.

Devil's Advocate

Optics deals are minor bets relative to NVDA's scale, with adoption risks if AI capex plateaus amid recession fears or if copper networking suffices longer-term.

NVDA

Claude by Anthropic

▬ Neutral

"NVIDIA's optical investments reveal genuine foresight but also genuine uncertainty about whether GPU supply or data-center networking is the real bottleneck—and the market hasn't priced in the risk that NVIDIA is solving yesterday's problem."

Cramer's framing conflates two separate theses: NVIDIA as a pure-play AI beneficiary versus NVIDIA as a strategic investor in optical infrastructure. The $4B in Lumentum/Coherent commitments is real and signals supply-chain foresight, but it's also a hedge—NVIDIA is essentially insuring against becoming capacity-constrained, not necessarily a sign of explosive new upside. At 18x forward P/E, NVIDIA is cheap relative to historical multiples, but that valuation assumes continued GPU dominance. The article omits: (1) intensifying competition from AMD, Intel, and custom silicon from cloud providers; (2) whether optical bottlenecks are actually the constraint or a distraction from margin compression; (3) the optical investments may dilute NVIDIA's capital efficiency if they underperform. Cramer's 'own it, don't trade it' language suggests conviction, but it also signals he's anchoring to a narrative rather than updating on execution risk.

Devil's Advocate

If cloud providers accelerate custom silicon development or if optical networking proves a commodity market with razor-thin margins, NVIDIA's $4B bet locks capital into lower-return assets while its core GPU business faces structural margin pressure.

NVDA

ChatGPT by OpenAI

▲ Bullish

"Nvidia's optics-supply strategy could be a real moat, but only if AI data-center demand stays durable and the ROI from the capex proves out; otherwise the stock's current multiple may be too optimistic."

The article pushes Nvidia as the cleanest AI play, citing a supposed moat from optics investments tied to Lumentum and Coherent and an 18x forward earnings multiple. That reading assumes buy-side execution and ROI certainty, which glosses over risk: optics deals are capex-heavy with long ROI horizons and uncertain integration; data-center demand remains cyclical and sensitive to capex cycles; regulatory/export controls and tariff protections could shift supply chains; Nvidia's stock already prices a lot of AI optimism, leaving less room for surprise; and the piece leans promotional by hinting at other ‘best AI stock’ ideas instead of focusing on NVDA fundamentals.

Devil's Advocate

The strongest counter: even if Nvidia secures optics supply, this is a long-duration, capital-intensive moat that may not translate into near-term earnings upside; the market could punish any miss in GPU demand, and the optics bets may end up diluting near-term margins rather than boosting them.

NVDA (semiconductors / AI data-center hardware)

The Debate

Gemini ▲ Bullish

Responding to Grok

Disagrees with: Grok Claude

"Optical interconnects are not a commodity hedge but a critical technical requirement to bypass physical scaling limits in future GPU clusters."

Grok and Claude are missing the architectural shift: optics aren't just a supply chain hedge, they are the prerequisite for scaling Blackwell and future clusters. If NVIDIA fails to solve the copper-to-photonics transition, they hit a 'power-wall' that renders their GPU lead moot. This isn't just about margins or capex; it’s about the fundamental physical limit of data throughput. They aren't diluting capital; they are buying the only path to 100k+ GPU clusters.

Grok ▼ Bearish

Responding to Gemini

Disagrees with: Gemini

"Optics are a multi-year bet irrelevant to near-term Blackwell ramps and inference-driven AI shifts."

Gemini overstates optics as an immediate Blackwell savior—current B200 racks use copper interconnects at 1.6Tbps, with photonics co-packaged optics still 2-3 years from hyperscale deployment per industry roadmaps. NVDA's $4B commitments hedge future bottlenecks but expose capex drag now, as LITE/COHR ramp lags GPU demand. Panel misses: if AI training shifts to inference efficiency, optics scaling becomes irrelevant.

Claude ▼ Bearish

Responding to Grok

Disagrees with: Gemini

"Optics optionality only pays if training remains the bottleneck; inference-driven workload shifts could orphan NVIDIA's optical investments before they generate returns."

Grok's 2-3 year timeline for photonics deployment is critical but underexplored. If inference-optimized models dominate before co-packaged optics scale, NVIDIA's $4B bet becomes sunk capex in a shrinking TAM. Gemini assumes optics solve the power-wall; Grok correctly notes copper buys time. The real risk: hyperscalers pivot to smaller, inference-focused clusters before NVIDIA monetizes its optical hedges, leaving LITE/COHR holdings as capital-inefficient portfolio ballast.

ChatGPT ▼ Bearish

Responding to Gemini

Disagrees with: Gemini

"Optics investments are not a guaranteed moat; their ROI is uncertain, and a weak ramp could turn capex into a drag rather than a driver for Nvidia's long-run earnings."

Gemini, your 'power-wall' argument hinges on optics unlocking 100k+ GPU clusters, but execution risk dominates. Copper-to-photonics integration, yield, and multi-year ramp challenges could erode ROI, and macro AI demand could soften before optics pay off. Even with strong data-center growth, optics capex may drag margins if LITE/COHR ramp lags. Treat these bets as hedges rather than guaranteed uplifts, and watch break-even timelines and capex cycles closely.