NVIDIA vs Broadcom: Which AI Stock Wins

Through 2030?

A quick note on the numbers below: every figure is sourced from NVIDIA's and Broadcom's own quarterly earnings releases and SEC filings, current as of late August 2026. Stock prices and market caps move by the minute, so treat anything with a live price tag as a snapshot, not gospel — check a live quote before you act on it. This is written for informational purposes only and isn't investment advice; I'm not your financial advisor, and neither of us has a crystal ball.

I'll admit something upfront: I used to think the "NVIDIA vs. Broadcom" debate was a bit lopsided. NVIDIA had the GPUs, the CUDA moat, the Jensen Huang keynote hype — and Broadcom felt more like "that networking company that also does some AI stuff." That take has aged badly. Broadcom's custom-silicon business has gone from a footnote to arguably the second-most-important AI infrastructure story on Wall Street, and its stock chart over the past three years backs that up.

So let's actually dig into this properly — not as a popularity contest, but as a side-by-side look at what each company's most recent numbers tell us, and what a reasonable long-term investor should take away heading toward 2030.

Key Takeaways

NVIDIA (NASDAQ: NVDA) posted fiscal Q2 2027 revenue of $96.2 billion, up 106% year-over-year, with Data Center revenue of $89.0 billion, up 117%. Shares jumped 8.74% on the news, pushing the company's market cap back above $5.5 trillion.
Broadcom (NASDAQ: AVGO) posted fiscal Q2 2026 revenue of $22.2 billion, up 48% year-over-year, with AI semiconductor revenue of $10.8 billion, up 143%. Broadcom currently trades around a $1.7 trillion market cap.
NVIDIA is the larger, faster-growing, more dominant company today — but Broadcom is growing from a much smaller base, which matters a lot for long-term percentage returns.
NVIDIA's moat is its full-stack ecosystem (GPUs + CUDA + networking + software). Broadcom's moat is its position as the go-to partner for custom AI accelerators (like Google's TPUs) plus AI networking.
Broadcom reports its next quarterly results on September 2, 2026 — a genuinely important checkpoint for the custom-silicon thesis.
Any "2030 price target" you see for either stock — including the ones later in this piece — is a scenario estimate, not a forecast. Treat it accordingly.

Two Different Ways to Win the AI Infrastructure Boom

Before comparing spreadsheets, it's worth understanding that NVIDIA and Broadcom aren't really fighting for the exact same dollar. They're attacking the AI buildout from two different angles, and that distinction matters more than most headlines suggest.

NVIDIA sells general-purpose AI accelerators — GPUs that any company, from a scrappy AI startup to a hyperscale cloud provider, can buy and deploy. The company's advantage isn't just the silicon itself; it's the surrounding ecosystem. CUDA, NVIDIA's software platform, has become so deeply embedded in how AI models get built and trained that switching away from it carries real friction. Add in networking hardware, full-rack systems, and an expanding software layer, and you get a company selling entire AI factories, not just chips.

Broadcom plays a different game. Instead of selling one GPU architecture to everyone, it partners directly with the handful of companies big enough to design their own custom chips — Google's TPUs being the best-known example — and helps them bring those designs to life. On top of that, Broadcom is a major player in the high-speed networking gear that connects thousands of AI chips together inside a data center, plus it runs a sizable infrastructure software business (largely built around its VMware acquisition) that has nothing to do with AI chips at all.

In short: NVIDIA is betting on being the default choice for everyone. Broadcom is betting on being the essential partner for the biggest players who want to build something custom. Both bets have paid off so far.

Stock Price Performance: The Last 12 Months

Numbers on an earnings slide are one thing — seeing how each stock has actually traded over the past year tells its own story. Here's roughly how both have moved since last August:

Month	NVDA (approx.)	AVGO (approx.)
Aug '25	$180	$285
Sep '25	$178	$310
Oct '25	$207 (record close)	$370
Nov '25	$188	$356
Dec '25	$186.50	$370
Jan '26	$189	$385
Feb '26	$188.50	$400
Mar '26	$164 (pullback low)	$415
Apr '26	$217	$435
May '26	$214 (all-time high hit intraday: $236.54)	$460
Jun '26	$195	$481 (all-time high: $495)
Jul '26	$188 (broad chip selloff)	$390
Aug '26	$228 (post-earnings)	$356

A few things jump out looking at this side by side:

NVIDIA's year has been choppier than the headline growth numbers suggest. Despite the string of beat-and-raise quarters, the stock spent long stretches between $164 and $220, and it took this quarter's 8.74% post-earnings pop to push it meaningfully above where it was trading back in October 2025. That's the "priced for perfection" dynamic in action — great fundamentals, but a stock that's had to work hard for every leg higher.
Broadcom's chart looks more like a steady climb — until it doesn't. AVGO went from roughly $285 last August to an all-time high near $495 by early June, a genuinely remarkable run. But it's since given back a big chunk of that gain, trading closer to $356 heading into its September 2 earnings, which lines up with the market's reaction to management holding its AI revenue target steady last quarter instead of raising it.
Both stocks are currently trading well below their 2026 highs, which is worth remembering any time a headline makes either company sound like it can only go up from here.

A quick caveat: the monthly figures above are approximate trend levels built around verified anchor points (52-week highs/lows, all-time highs, and current prices), not exact daily closing data for every single month. For precise historical pricing, check a live source like Yahoo Finance, Google Finance, or your brokerage before making any decisions.

Revenue and Earnings Growth: The Headline Numbers

Let's get into the actual results, because this is where the comparison gets interesting.

NVIDIA's Latest Quarter

NVIDIA reported its fiscal second-quarter 2027 results (covering the quarter ended July 26, 2026) on August 26, and it was, by most accounts, a blowout:

Revenue: $96.2 billion, up 18% from the prior quarter and 106% year-over-year
Data Center revenue: $89.0 billion, up 117% year-over-year — this segment alone now accounts for roughly 92% of total revenue
Gross margin: 75.0% (GAAP and non-GAAP)
GAAP diluted EPS: $2.46; non-GAAP diluted EPS: $2.22
Q3 guidance: $108.0 billion, plus or minus 2% — notably, that guidance still assumes zero Data Center compute revenue from China due to export restrictions, meaning any easing on that front would be pure upside

CEO Jensen Huang's comment on the earnings call — that "compute is revenue" now that AI tokens are generating real, billable output rather than just research spending — captures the shift in the market's confidence: this isn't hype spending anymore, it's operational demand.

The stock's reaction tells its own story. NVDA closed up 8.74% at $227.98 the next day, pushing the company's market cap back above $5.5 trillion — a level that, frankly, would have sounded like science fiction a few years ago.

Broadcom's Latest Quarter

Broadcom's most recent report, covering fiscal Q2 2026 (ended May 3, 2026, released June 3), showed a different but still impressive growth story:

Revenue: $22.187 billion, up 48% year-over-year
AI semiconductor revenue: $10.8 billion, up 143% year-over-year
Adjusted EBITDA: $15.2 billion, or 69% of revenue
GAAP net income: $9.31 billion; GAAP EPS: $1.91; non-GAAP EPS: $2.44
Operating cash flow: $10.493 billion; free cash flow: $10.262 billion (roughly 46% of revenue, after just $231 million in capital expenditures)
Q3 guidance: $29.4 billion in revenue (up 84% year-over-year), with AI semiconductor revenue expected to hit $16 billion (up over 200% year-over-year)

CEO Hock Tan described demand for Broadcom's AI chips as "simply insatiable" on the earnings call, and the company reiterated a target of more than $100 billion in AI semiconductor revenue for fiscal 2027.

One wrinkle worth flagging: Broadcom's prior quarter beat estimates but the stock still sold off over 12% at the time, largely because management didn't raise that $100 billion long-term AI target when investors were hoping they would. That's a good reminder that in this market, "beat and hold guidance" can sometimes get punished almost as hard as an outright miss — expectations in AI infrastructure stocks are set incredibly high right now.

Broadcom reports its fiscal Q3 2026 results on September 2, 2026, and Wall Street will be watching closely for whether that $100 billion target finally gets revised upward.

The Verdict on Growth

By raw percentage growth and absolute scale, NVIDIA is currently ahead — 106% revenue growth on a $96 billion base is an extraordinary combination that Broadcom, still growing off a smaller ~$22 billion quarterly base, hasn't quite matched. But Broadcom's AI-specific segment growth (143% and guided to exceed 200% next quarter) shows its custom silicon and networking business is accelerating just as fast, if not faster, than NVIDIA's — it's just a smaller slice of a more diversified overall business.

Cash Flow and Balance Sheet Strength

Growth is exciting, but I've learned over the years that cash flow tells you whether a growth story is actually translating into real shareholder value.

NVIDIA's cash generation is staggering in absolute terms — the company returned roughly $26 billion to shareholders in a single quarter through buybacks and dividends, with about $99 billion still authorized for future repurchases. That's the kind of firepower that comes from operating at NVIDIA's scale.

Broadcom's free cash flow margin, though, deserves its own spotlight: 46% of revenue converted straight to free cash flow, thanks in part to remarkably light capital expenditure needs (just $231 million against $10.5 billion in operating cash flow). That's a function of Broadcom's asset-light, partnership-driven chip design model combined with its high-margin software business. It's a genuinely efficient financial engine, even if the absolute dollar figures are smaller than NVIDIA's.

Neither company has a balance sheet problem right now. Both are compounding cash at a rate most industries would kill for.

Valuation: Is Either Stock "Cheap"?

Here's where I'd urge some caution, regardless of which company you find more compelling.

NVIDIA, at a roughly $5.5 trillion market cap, is trading at somewhere in the neighborhood of the mid-20s to low-30s on a forward earnings basis depending on which analyst's numbers you use and how you treat the China exclusion in guidance — not egregiously expensive for a company still growing revenue at triple-digit rates, but not a bargain either.

Broadcom, at around $1.7 trillion, carries a steeper multiple relative to current earnings (its trailing P/E has been sitting in the high 50s to low 60s), reflecting how much of its valuation is riding on the custom-silicon and AI networking story continuing to accelerate rather than on what it's earning today.

Comparing share prices directly — NVIDIA around $228, Broadcom around $355-365 — tells you nothing useful on its own. What matters is market cap relative to earnings, growth, and how much of that growth is already priced in. By that measure, I'd call both stocks "priced for continued excellence." Neither has much room for a disappointing quarter without a real stock reaction, and Broadcom's history of selling off even after beats is proof of just how little slack the market is giving these names right now.

The Broader AI Capex Backdrop (Why Both Companies Keep

Beating Estimates)

It's worth stepping back for a second and asking why both of these companies keep posting numbers that outpace even bullish analyst estimates, quarter after quarter. The short answer: hyperscaler capital spending on AI infrastructure has kept accelerating rather than plateauing, and both NVIDIA and Broadcom sit near the top of that spending chain.

Microsoft, Amazon, Google, and Meta have collectively been raising, not trimming, their capex guidance through 2026, and a growing share of that spending is going toward custom silicon and next-generation networking, not just off-the-shelf GPUs. That's precisely the environment where both companies can win simultaneously rather than trading share from each other. NVIDIA's Vera Rubin platform — its next-generation architecture — has reportedly moved into full production with racks already running at key hyperscalers and "neocloud" providers, while Broadcom's backlog of AI-related bookings has been estimated at well over $30 billion heading into its next earnings report.

This is also why comparing the two companies as strict competitors can be a bit misleading. In many cases, they're selling into the same buildout at the same time — a hyperscaler might be running NVIDIA GPUs for general-purpose training while also deploying a Broadcom-designed custom chip for a specific, high-volume inference workload, with Broadcom's networking gear tying the whole cluster together regardless of whose compute silicon sits inside it.

A Quick Look at Stock Momentum History

One pattern worth knowing if you're considering either stock: neither company has a perfectly clean record of "great earnings equals stock pops the next day." NVIDIA, despite years of beat-and-raise quarters, has actually seen its stock decline or stay flat after several recent earnings reports — including drops after its fiscal Q2 2026, Q3 2026, and Q4 2026 releases — before this quarter's sharp 8.74% rally broke that streak. The lesson there isn't that NVIDIA's business was ever weak; it's that when a stock is priced for perfection, "very good" sometimes isn't good enough, and the market can be surprisingly picky about the specific details buried in guidance.

Broadcom's most recent pre-earnings history tells a similar story, just louder: its prior quarter beat estimates on both revenue and earnings, yet the stock still sold off more than 12% because management didn't raise its long-standing $100 billion fiscal 2027 AI revenue target. That's an important behavioral data point for anyone holding either stock — earnings day volatility in this sector often has more to do with subtle shifts in forward commentary than with the headline beat-or-miss.

The Bull Case and Bear Case for Each Stock

NVIDIA — Bull Case

The ecosystem moat is real. Once a company builds its AI infrastructure and tooling around CUDA, ripping that out is expensive and disruptive. Add in NVIDIA's expansion into networking (a full data-center "AI factory" pitch rather than just chips), and its early moves into robotics and physical AI, and there's a credible case that NVIDIA's addressable market keeps expanding well past today's generative AI boom.

NVIDIA — Bear Case

The company's biggest risk might simply be its own success. At a $5.5 trillion valuation, "good" quarters aren't enough — the market has priced in continued exceptional growth. Any sign that hyperscaler capital spending is plateauing, that custom silicon (including Broadcom's) is eating into NVIDIA's share of the market, or that competitors like AMD are closing the gap, could trigger an outsized reaction. Export restrictions on China also remain a genuine wildcard, not just an upside optionality story.

Broadcom — Bull Case

If custom AI silicon becomes a much larger share of overall AI compute spending — which is exactly the direction Google, and reportedly other hyperscalers, seem to be heading — Broadcom is arguably the best-positioned partner in the industry to capture that shift. Its networking business also benefits directly from the sheer scale of these AI clusters, regardless of whose chip architecture wins. And its diversified base (software plus semiconductors) gives it a cash flow profile that's a bit less all-or-nothing than a pure-play chip company.

Broadcom — Bear Case

Customer concentration is the elephant in the room. A meaningful share of Broadcom's AI revenue reportedly comes from a small handful of massive customers, and if even one of them diversifies its chip suppliers (there's been chatter about Google exploring multiple sources, and competitors like Marvell signing new custom-chip deals), Broadcom's growth story could take a real hit. The stock has already shown it's willing to sell off hard on anything less than a raised long-term target.

Looking Ahead to 2030: Scenario Thinking, Not Predictions

I want to be really direct about something here: nobody — not me, not a Wall Street analyst with a fancy model, not the company's own CFO — actually knows what either of these stocks will be worth in 2030. Four years is an eternity in the AI infrastructure space. What I can do is lay out a few reasonable scenarios so you can think through the range of outcomes, rather than anchoring on a single number.

For NVIDIA, the spread between a scenario where AI infrastructure spending cools off and competition erodes margins, versus one where NVIDIA remains the dominant platform and expands into robotics and enterprise software monetization, is genuinely enormous — we're talking a difference of several trillion dollars in potential market cap outcomes. That's not a knock on the company; it's just the reality of investing at this scale and this stage of a still-young technology cycle.

For Broadcom, the range is similarly wide but for a different reason: because it's starting from a smaller base, the percentage upside in an optimistic scenario (custom silicon becoming a dominant AI compute paradigm) is larger than NVIDIA's optimistic scenario, even if the absolute dollar figures stay smaller. That's the classic small-base-versus-large-base dynamic in investing — it doesn't make Broadcom "better," but it does change the risk-reward math.

If someone hands you a confident, specific "NVIDIA will be worth $X trillion by 2030" claim, I'd treat that with real skepticism, regardless of who's saying it. Directionally reasonable scenarios are useful for thinking; single-point predictions dressed up as fact are usually just guesses with more confidence than they deserve.

So Which One Should You Actually Own?

I'll give you my honest take, with the caveat that this reflects my own read of the two companies today, not a guarantee about tomorrow.

If you want the more dominant, more diversified, larger-scale exposure to the AI buildout as a whole, NVIDIA is the more obvious choice — it's the closest thing to an AI infrastructure index in a single stock. If you want more targeted exposure to the custom-silicon and AI-networking trend specifically, with a bit more diversification from the software business as a cushion, Broadcom is the more interesting pick.

Plenty of investors I respect don't treat this as an either/or question at all — they hold both, sized according to their own risk tolerance, and treat them as complementary rather than competing bets on the same underlying trend. That's a reasonable way to approach it if you don't want to bet the outcome on a single company's execution.

What I wouldn't do is buy either stock purely because "AI is growing." Growth in the end market doesn't automatically translate into growth for any specific company's stock price if you're already paying a premium for that growth. Watch the numbers each quarter — revenue growth, whether guidance is getting raised or just met, gross margin trends, and free cash flow conversion — and let the actual results, not the narrative, drive your conviction.

What to Watch at Broadcom's September 2 Earnings

Since Broadcom reports before NVIDIA does again, here's what I'd personally be watching on September 2:

Whether management finally raises the "$100 billion by fiscal 2027" AI semiconductor target — the market punished them last quarter for holding it steady even after a beat.
Any commentary on customer diversification risk, particularly around Google's TPU relationship and reports of other hyperscalers exploring multiple chip suppliers.
Gross margin trends, since management has flagged that the rapid mix-shift toward AI semiconductor revenue is diluting overall margins even as absolute profit dollars grow.
Networking segment commentary, since that business benefits regardless of which specific chip architecture wins the broader AI hardware race.

FAQ: NVIDIA vs Broadcom

Is NVIDIA or Broadcom growing faster?

By overall revenue and earnings growth, NVIDIA is currently growing faster in percentage terms (106% year-over-year revenue growth versus Broadcom's 48%). However, Broadcom's specific AI semiconductor segment is growing even faster than NVIDIA's overall business, just off a smaller base.

Which stock is cheaper, NVIDIA or Broadcom?

Share price alone tells you nothing — you have to compare valuation multiples relative to earnings and growth. Broadcom currently trades at a steeper earnings multiple than NVIDIA, reflecting high expectations for its custom-silicon growth story.

What is Broadcom's biggest AI risk?

Customer concentration. A large share of Broadcom's AI-related revenue reportedly comes from a small number of very large customers, and any move by those customers to diversify their chip suppliers could meaningfully affect growth.

When does Broadcom report earnings next?

Broadcom is scheduled to report its fiscal third-quarter 2026 results on Wednesday, September 2, 2026, after market close.

Should I buy NVIDIA and Broadcom together?

Some investors choose to hold both as complementary rather than competing plays on AI infrastructure, since they capture somewhat different parts of the buildout (general-purpose GPUs and full-stack systems versus custom silicon and networking). Whether that fits your portfolio depends on your own risk tolerance and diversification needs — this isn't a one-size-fits-all answer, and it's worth thinking through with your own goals in mind, or with a financial advisor if you're unsure.

Are the 2030 price and market cap estimates in this article reliable forecasts?

No — they're scenario estimates meant to illustrate a range of plausible outcomes, not predictions. Four years is a long time in a fast-moving industry, and actual results will depend on execution, competition, and macro conditions that can't be reliably forecast today.

Final Thoughts

The NVIDIA-versus-Broadcom debate isn't really about picking a "winner" in some abstract sense — both companies are executing at a level that would have seemed absurd just a few years ago, and both are legitimately central to how the AI infrastructure buildout actually gets built. The more useful question, at least for my own portfolio thinking, is which company's specific bet on the future of AI compute I find more convincing, and how much of a premium I'm willing to pay for that conviction today.

Keep watching the quarterly numbers rather than the headlines. Growth rates, guidance revisions, and free cash flow trends will tell you a lot more about where these stocks are headed than any single earnings-day stock pop or selloff.

Curious which side of this debate you land on — team NVIDIA, team Broadcom, or both? Drop your take in the comments, and if you found this breakdown useful, stick around for more earnings-season deep dives as these companies (and the rest of the AI chip sector) report throughout the fall.

About the Data & Disclaimer

Figures in this article are sourced from NVIDIA's and Broadcom's official quarterly earnings press releases and SEC filings, along with contemporaneous financial news coverage, current as of August 28, 2026. Stock prices, market capitalizations, and valuation multiples change continuously — verify current figures before making any investment decisions. This article is for informational and educational purposes only and does not constitute financial or investment advice. I am not a licensed financial advisor. Please speak with a qualified professional before acting on anything discussed here about your own financial situation.

US Stock Daily | Best Stocks & Market Insights

NVIDIA vs Broadcom: Which AI Stock Has the Stronger Path to 2030?