Broadcom Is 24% Off Its High and Just Unveiled a Custom AI Chip With OpenAI. Time to Buy the Dip?

Broadcom (NASDAQ: AVGO) has spent 2026 doing two things at once: posting some of the fastest growth of any large company in the market, and watching its stock sink anyway. Shares trade about 24% below their 52-week high of $495 as of this writing -- even after Broadcom and OpenAI unveiled Jalapeño, their co-developed artificial intelligence (AI) chip, in late June.

An accelerating business attached to a discounted stock is the raw material of every buy-the-dip debate. Is this one worth taking?

Image source: Getty Images.

The growth the sell-off is ignoring

"Q2 semiconductor revenue from AI of $10.8 billion grew 143% year-over-year, above our forecast, driven by increasing demand for custom AI accelerators and AI networking," said CEO Hock Tan in the company's fiscal second-quarter earnings release.

And the momentum is guided to steepen from there. Tan said Broadcom expects AI chip revenue to grow more than 200% year over year this quarter, to $16.0 billion -- within total revenue guidance of about $29.4 billion, up 84%. For a company this size, growth rates like these have almost no precedent outside the AI build-out itself.

Companywide, second-quarter revenue grew 48% to a record $22.2 billion, and adjusted earnings before interest, taxes, depreciation, and amortization came in at 69% of revenue -- profitability most software companies would envy, produced by a chipmaker.

Behind the numbers sits a short, remarkable customer list.

Tan told analysts that Broadcom now has six core custom-chip customers, including OpenAI, Anthropic, Meta Platforms, and Google parent Alphabet.

Jalapeño is the newest evidence that the model works. The inference processor -- built to run AI models for users, rather than train them -- reportedly went from initial design to completion in about nine months, with OpenAI's own models helping speed the engineering. The partners have said they plan to deploy racks of OpenAI-designed chips starting late this year, building toward systems that would ultimately draw 10 gigawatts of power. The two companies first revealed their plans last October, after 18 months of joint work behind closed doors.

The design choice matters for the industry, too. Jalapeno is an application-specific chip -- less flexible than a graphics processing unit, but cheaper to run for one dedicated job. Every workload that moves to silicon like this is one that no longer needs a general-purpose chip, which is a big part of why custom accelerators have arguably become the industry's fastest-growing niche.

Why are shares trading at a discount?

So why is a company growing this fast still 24% off its high?

Two worries seem to carry most of the weight. The first is concentration: a business increasingly tied to a handful of giant AI customers rises and falls with their spending decisions, and the whole AI trade stumbled into July. The second is the economics of custom silicon -- purpose-built chips are typically cheaper for customers than off-the-shelf processors, which is much of their appeal and limits their sellers' pricing power.

Has the stock's pullback made it a buy?

On trailing results -- still weighed down by acquisition-related amortization running through the income statement -- Broadcom trades at about 60 times earnings. On forecast earnings, however, the multiple falls to about 19, because profits are scaling with the AI ramp.

Then there's the company's impressive cash generation.

Broadcom produced $10.3 billion of free cash flow last quarter -- 46% of revenue -- and pays a $0.65 quarterly dividend besides.

Overall, I'd call this dip buyable, but in moderation. The demand signals are among the strongest anywhere in the chip industry. The customer list keeps deepening, and Jalapeño could be a game changer.

The risks -- customer concentration, custom-chip pricing, and an AI trade prone to sudden repricing -- are exactly why the shares sit this far below their high, and they argue for a modest position rather than a bold one. But at about 19 times forward earnings, helped by revenue growth guided to exceed 80%, the discount looks larger than the danger.

I might become more cautious about the stock if upcoming reports show a material step-down in AI revenue growth from the guided pace, or a pullback by one of those six anchor customers.

Until then, I'd rather own this dip than wait for a friendlier headline and a higher price.

Daniel Sparks and his clients have no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Alphabet, Broadcom, and Meta Platforms. The Motley Fool has a disclosure policy.