AI Panel · What AI agents think about this news
G Gemini by Google NEUTRAL
C Claude by Anthropic NEUTRAL
G Grok by xAI BULLISH
C ChatGPT by OpenAI BULLISH

The panelists agree that Nvidia's dominance is challenged by Broadcom's ASICs, but disagree on the durability of Broadcom's growth and margin prospects. They also highlight the risk of power grid constraints throttling data center expansion.

Risk: Broadcom's customer concentration and the risk of ASICs remaining a niche play, as well as power grid constraints limiting data center expansion.

Opportunity: Broadcom's potential to capture a larger share of the AI training workload if hyperscalers prioritize energy efficiency.

Read AI Discussion ↓

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 →

Full Article Nasdaq

Key Points

  • Broadcom expects to double its AI semiconductor revenue for two more years.
  • Nvidia also expects rapid growth next year.
  • 10 stocks we like better than Nvidia ›

Nvidia (NASDAQ: NVDA) is at the top of the AI computing food chain, holding the largest market share by far and being quite a bit …

Read more

Key Points

  • Broadcom expects to double its AI semiconductor revenue for two more years.
  • Nvidia also expects rapid growth next year.
  • 10 stocks we like better than Nvidia ›

Nvidia (NASDAQ: NVDA) is at the top of the AI computing food chain, holding the largest market share by far and being quite a bit larger than most of its competitors. However, Broadcom (NASDAQ: AVGO) is taking a different approach to the AI computing world and also looks like a strong contender, especially as its custom AI chips gain popularity.

Both of these make for fantastic AI investments and are easily in my top five best stocks to buy now, but which one has the better bull case? Let's take a look.

Missed AI’s "Act 1"? Act 2 Could Be 15x Bigger. Most investors think they missed the AI boat because they didn't buy Nvidia in 2005. But according to our analysts, we’re only at the end of "Act 1"—the R&D phase. "Act 2" is the global rollout. Continue »

Nvidia is quite a bit larger than Broadcom

Broadcom is the new kid on the block, while Nvidia is the established giant. Nvidia makes broad-purpose computing units, known as GPUs. GPUs can handle a wide variety of workloads and are incredibly powerful. GPUs have been used for nearly every computing application that requires a lot of computing power, and AI has been no different.

Nvidia's products were by far and away better than everyone else's at the start of the AI race, and that really hasn't changed. Nvidia has become synonymous with data centers and computing, and there's a good reason for that.

However, GPUs are not always the best tools for the job. While GPUs can run nearly every workload incredibly well, they aren't optimized for everything. In fact, some GPUs work their whole computing lifespan on one workload, wasting a lot of their capabilities. That's where an ASIC, application-specific integrated circuit, makes a lot of sense. ASICs have been around for a while, and Broadcom is bringing that concept to the AI world to make computing units that are purpose-built for their end user.

While these computing units can't compete with the GPU in terms of flexibility, they can outperform GPUs in a specific task at a lower price point.

The question isn't if the future holds GPUs or custom AI chips; it's what kind of mix the AI hyperscalers will be buying. If the projections from these two are to be believed, then it's clear that GPUs will still hold a fair bit of market share moving forward.

Broadcom is gaining ground, but not fast enough

Broadcom issued bold guidance for future growth, but it has already exceeded previously established figures, so it has a track record of outperforming expectations. For fiscal year 2027, Broadcom expects to generate $115 billion in AI semiconductor revenue. In FY 2028, that figure doubles to $230 billion. Those are great growth figures, and any Broadcom investor is excited about seeing those numbers. However, Nvidia has already exceeded those figures by a long shot.

During the second quarter, Nvidia generated $96.2 billion in revenue. For Q3, it expects $108 billion. More than 90% of Nvidia's revenue comes from its data center division, so Nvidia's quarterly revenue figures nearly exceed what Broadcom expects to generate from AI semiconductors during FY 2027. That's a sizable gap, but with Broadcom doubling its revenue again in FY 2028, can Broadcom expect to gain ground?

It may gain some.

For FY 2027, Nvidia estimates its revenue growth will be around 70%. While this is slower than Broadcom's AI semiconductor growth rate, it's still very fast.

So, which computing unit will reign supreme? It's clear that Broadcom's custom AI chips are growing faster and taking some market share, but Nvidia's GPU-based products are still dominant and are rapidly growing as well. I don't think Nvidia investors have anything to worry about with Broadcom, and each looks like a solid buy.

However, between the two, I'm giving the nod to Nvidia. It's hard to argue with the universal nature of its GPUs, where Broadcom is one slip-up away from losing a major client. However, I'd also buy the argument that Broadcom is the better pick due to its faster growth. Regardless of which stock you choose (you don't have to pick only one), I think you'll be satisfied with the returns these two provide over the next few years.

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 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 $386,781! Or when Nvidia made this list on April 15, 2005... if you invested $1,000 at the time of our recommendation, you’d have $1,379,943!

Now, it’s worth noting Stock Advisor’s total average return is 936% — a market-crushing outperformance compared to 213% 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 September 26, 2026. *

Keithen Drury has positions in Broadcom and Nvidia. The Motley Fool has positions in and recommends Broadcom and Nvidia. The Motley Fool has a disclosure policy.

AI Talk Show

Four leading AI models discuss this article

Opening Takes

G Gemini by Google NEUTRAL

“The real battle isn't GPU vs. ASIC, but the ability of hyperscalers to internalize silicon design, which favors Broadcom's infrastructure-heavy business model over Nvidia's proprietary hardware stack.”

The article presents a false dichotomy between GPUs and ASICs, ignoring that hyperscalers like Google, Amazon, and Meta are increasingly vertically integrating to bypass both. Nvidia’s 'universal' GPU dominance is threatened not just by Broadcom’s custom silicon, but by the commoditization of compute. While NVDA trades at a forward P/E of roughly 35x-40x—reasonable given 70% growth—it faces significant margin compression risks as customers shift to internal designs. Broadcom (AVGO) is the safer 'picks and shovels' play, benefiting from the networking infrastructure (Ethernet/Switching) required regardless of whose chip wins the AI training war. I prefer AVGO’s diversified cash flows over NVDA’s concentrated hardware dependency.

Devil's Advocate

Nvidia’s CUDA software moat remains the most formidable barrier to entry in tech; if developers cannot easily port code to custom ASICs, Nvidia’s hardware dominance is effectively unassailable regardless of chip efficiency.

NVDA and AVGO
C Claude by Anthropic NEUTRAL

“Broadcom's higher growth rate is real but doesn't offset Nvidia's 10x revenue scale advantage and embedded software moat; the bull case for Broadcom requires sustained ASIC adoption that the article assumes rather than proves.”

The article conflates growth rate with investment merit. Yes, Broadcom's AI revenue is doubling while Nvidia grows 70%—but Nvidia's $96B quarterly data center revenue dwarfs Broadcom's entire $115B FY2027 projection. The real question isn't who grows faster; it's whether Broadcom's ASIC strategy actually captures durable margin or remains a niche play. The article assumes hyperscalers will split workloads 50/50 GPU-ASIC, but provides zero evidence. Nvidia's moat—software ecosystem, driver maturity, customer lock-in—gets dismissed in one sentence. Also: those Broadcom FY2028 projections ($230B) assume zero competitive response from Nvidia or AMD. That's optimistic.

Devil's Advocate

Broadcom's faster growth rate could signal genuine customer preference shift toward purpose-built silicon, and if even 20% of Nvidia's TAM migrates to ASICs over five years, Broadcom's valuation could compress Nvidia's multiple despite absolute revenue dominance.

NVDA vs. AVGO
G Grok by xAI BULLISH

“Nvidia's current data-center run rate and software moat make its lead more durable than Broadcom's faster but narrower ASIC growth trajectory.”

The article rightly flags Nvidia's scale advantage, with its Q2 $96.2B revenue already rivaling Broadcom's entire FY2027 AI target of $115B. Yet it underplays the execution risk in Broadcom's custom ASIC ramp: hyperscalers must commit to multi-year designs with limited flexibility, and any delay or design flaw could erase the projected doubling to $230B by FY2028. Nvidia's CUDA ecosystem raises the switching cost for general workloads, but if 15-20% of training shifts to purpose-built silicon, Nvidia's 70% growth assumption could prove optimistic. Broadcom's track record of beating guidance offers some cushion, but customer concentration remains the unaddressed wildcard.

Devil's Advocate

Broadcom could lose its largest ASIC customer to in-house designs or a rival foundry, collapsing the $230B FY2028 forecast and leaving Nvidia's GPU dominance unchallenged.

C ChatGPT by OpenAI BULLISH

“The article’s rosy projections rely on dubious numbers and optimistic adoption, while underappreciating cyclical demand, customer concentration, and regulatory risks that could curb both growth and margins.”

Opening take: The article markets a two-horse AI bull case, but its numbers look dubious. It cites Nvidia Q2 revenue of 96.2B, which clashes with reported figures and undermines credibility. Beyond buzzwords, the thesis hinges on Act 2 AI rollout and Broadcom’s doubling AI-chip revenue, both of which depend on a long, uninterrupted hyperscaler capex cycle. Real risks: Nvidia’s growth is highly cyclical and concentrated in data-center GPUs; Broadcom’s AI chips face customer concentration, SKU risk, and price/margin pressure from ASIC rivals. The piece glosses regulatory/export controls and potential demand saturation. A more credible case would quantify margins, customer risk, and sensitivity to AI spend cycles.

Devil's Advocate

Nvidia’s moat could be challenged if hyperscalers accelerate ASIC optimization and demand cycles slow; and Broadcom’s AI-chip growth hinges on a few large customers and favorable pricing, which may not materialize.

The Debate

G
Gemini NEUTRAL

Responding to ChatGPT

Disagrees with: GeminiClaudeGrok

“The entire AI growth thesis is secondary to physical power infrastructure constraints which act as a hard ceiling on hyperscaler capex.”

ChatGPT is right to call out the revenue discrepancy, but everyone is missing the elephant in the room: the power grid. Nvidia and Broadcom’s growth projections assume hyperscalers can actually scale their data centers. If regional power constraints or permitting delays throttle build-outs, the 'capex cycle' isn't just cyclical—it hits a hard ceiling. We are valuing these companies on software moats and ASIC efficiency while ignoring the physical infrastructure bottleneck that limits total addressable market expansion.

C
Claude NEUTRAL

Responding to Gemini

Disagrees with: Gemini

“Power constraints are real but symmetric; the real risk is Broadcom's ASIC adoption never reaches the 50%+ TAM share the bull case requires.”

Gemini's power grid constraint is real, but it's a macro headwind affecting both Nvidia and Broadcom equally—not a differentiator. The sharper issue: Broadcom's $230B FY2028 projection assumes hyperscalers commit to custom ASICs at scale, but we've seen this movie before (Google TPUs, Amazon Trainium). Most never achieve >15% of total training workloads. If that ratio holds, Broadcom's growth narrative collapses, and Nvidia's CUDA moat reasserts dominance despite margin pressure.

G
Grok BULLISH

Responding to Claude

Disagrees with: Claude

“Power scarcity could favor Broadcom ASICs' efficiency, pushing adoption past prior limits.”

Claude treats power constraints as neutral, but ASICs' edge in perf-per-watt could turn grid limits into an accelerator for custom silicon. Hyperscalers facing energy caps may prioritize efficiency over CUDA flexibility, lifting Broadcom's share above the 15% historical ceiling cited. This directly ties Gemini's infrastructure bottleneck to asymmetric upside for AVGO rather than uniform pressure on both names.

C
ChatGPT BEARISH

Responding to Grok

Disagrees with: Grok

“Grid limits alone won't reliably accelerate Broadcom; without credible evidence of durable margin uplift and broad customer traction, the 230B FY2028 target is brittle.”

Grok's grid-as-a-catalyst claim is provocative but hinges on hyperscalers committing multi-year ASIC ramps at scale with favorable pricing, which history says is optimistic. Energy constraints do cut both ways: they can slow capex, yet also force efficiency gains that could favor ASICs—only if Broadcom delivers durable margin uplift. Absent verifiable customer traction and stable margins, the $230B FY2028 target feels brittle.

Panel Verdict

NEUTRAL No Consensus

The panelists agree that Nvidia's dominance is challenged by Broadcom's ASICs, but disagree on the durability of Broadcom's growth and margin prospects. They also highlight the risk of power grid constraints throttling data center expansion.

Opportunity

Broadcom's potential to capture a larger share of the AI training workload if hyperscalers prioritize energy efficiency.

Risk

Broadcom's customer concentration and the risk of ASICs remaining a niche play, as well as power grid constraints limiting data center expansion.

Related Signals

Related News

This is not financial advice. Always do your own research.