Josh Brown: The Nvidia Dip Worth Buying Is in AMD, Not Nvidia
On Nvidia's $115.84B earnings day, shareholder Josh Brown pointed to AMD's Helios — and Anthropic's reported 3-day chip migration — as the dip worth buying.

Why it matters
- Nvidia's Q3 FY2027 revenue reached $115.84B, up 103.2% YoY, beating $108.73B consensus
- Josh Brown said on CNBC's Halftime Report (Nov 17, 2026): "The dip worth buying may not actually be in Nvidia"
- Brown cited an Anthropic filing describing a 3-day migration to AMD Helios chips through HPE
- AMD shares are up 137.22% YTD; Q3 Data Center revenue hit $8.55B, up 97% YoY
- Nvidia trades at 31x trailing earnings; AMD trades at 191x trailing earnings
On the day Nvidia reported a 103.2% revenue increase to $115.84 billion, one of its best-known shareholders told CNBC viewers to buy the dip somewhere else. Josh Brown, a Halftime Report regular introduced a week earlier as a "prominent shareholder" of the chipmaker, said on November 17, 2026: "The dip worth buying may not actually be in Nvidia."
The statement landed on the same day Nvidia posted third-quarter fiscal 2027 results that beat expectations across every metric. The gap between the numbers and the advice frames the current debate over AI chip stocks: whether Nvidia's blowout growth still justifies its position, or whether the balance of power between chip seller and chip buyer is finally shifting.
What did Brown actually say?
Brown framed his argument as the third of Nvidia's problems and, in his words, "maybe the most interesting for the viewers." His core claim: "the dip worth buying may be in some of the beneficiaries of a mess at Nvidia."
The logic runs through buyer leverage. For years, companies building AI infrastructure had to court Nvidia to secure GPU allocation, because no credible alternative existed. Brown argued that this dynamic may now be breaking, giving buyers genuine negotiating power for the first time since the launch of ChatGPT.
His evidence came from a regulatory filing. Brown said: "It turns out we were able to get up and running using AMD chips in the Helios architecture through HPE in 3 days, which is what Anthropic just told us in their latest filing." That description is Brown's characterization of the filing, which 24/7 Wall St. did not independently verify. The claim matters because a three-day migration to AMD hardware would undercut the assumption that Nvidia's CUDA ecosystem locks in its customers.
Brown then connected the point to valuation: "If we start to get a sense that this stranglehold that Nvidia has enjoyed over the AI trade since the launch of ChatGPT is maybe threatened regardless of whether Jensen comes back or not, then you'll understand why this stock is selling at a single-digit forward multiple."
The companies Brown means are Advanced Micro Devices (NASDAQ:AMD), whose Helios rack-scale systems reach some customers through Hewlett Packard Enterprise (NYSE:HPE).
How strong were Nvidia's numbers?
Nvidia's third-quarter fiscal 2027 results beat consensus on every headline line:
- Revenue: $115.84 billion, up 103.2% year over year, against a $108.73 billion consensus
- Non-GAAP EPS (earnings per share): $2.68, exceeding the $2.47 estimate
- Data Center sales: $107.60 billion, up 110% year over year
- Fourth-quarter guidance: $130.0 billion plus or minus 3%, with gross margin of 71-72%
The guidance assumes no Data Center compute revenue from China, meaning Nvidia projected this growth even while excluding one major market.
Interim CEO Colette Kress said in the release: "Vera Rubin is off to an explosive start, our roadmap is on track, and our ambition has never been greater." Vera Rubin is Nvidia's next-generation chip platform, and Kress's statement signals the product cadence has not slipped.
The filing also discloses risks that give Brown's leverage argument weight. It flags the leadership transition, an ongoing board investigation, and customer concentration: three direct customers accounted for 14%, 11%, and 11% of revenue. When three buyers represent over a third of sales, those buyers hold real bargaining power — precisely the dynamic Brown described.
The market reaction was muted relative to the beat. The stock rose 1.8% after hours to $158.26. Even after that gain, Nvidia was down 26.09% over the past month and down 14.94% year to date.
Why does AMD keep winning the rotation?
AMD shares are up 137.22% year to date, and the operating results have grown alongside the stock:
- Third-quarter revenue: $13.38 billion, up 44.7% year over year
- Data Center revenue: $8.55 billion, up 97% year over year
- Data Center operating margin: 34%, expanded from 25%
- Fourth-quarter guidance: approximately $17 billion plus or minus 3%, implying about 66% year-over-year growth
According to AMD's investor relations materials, Anthropic, Meta, Microsoft, OpenAI, and Oracle are all running workloads on the first Helios systems. That customer list directly supports Brown's argument: the largest AI labs — the buyers with the most leverage — are already diversifying across both suppliers.
CEO Lisa Su said: "We expect growth to accelerate further in the fourth quarter as we scale Helios and Venice production." Venice is AMD's next-generation AI accelerator, and its mention alongside Helios indicates the company expects supply to ramp, not plateau.
Does Brown's case hold up for long-term investors?
The diagnosis does. Nvidia's customer concentration, the unresolved question of who will lead the company, and the growing list of major AI customers running on AMD all point in the same direction: chip buyers are gaining leverage. AMD also starts from a much smaller Data Center base — $8.55 billion per quarter against Nvidia's $107.60 billion — so each new gigawatt-scale contract moves AMD's results far more than it moves Nvidia's.
The catch is price. Nvidia trades at 31 times trailing earnings with a 2.56% free cash flow yield (annual free cash flow divided by market capitalization). AMD trades at 191 times trailing earnings with a 0.81% yield. Investors are already paying a steep premium for AMD to keep taking share, which limits how much of the rotation story is left to buy cheaply.
For long-term holders, the analysis works best as a case for spreading AI chip exposure across both names. It works less well as a case for selling one to buy the other, because the valuations already price divergent outcomes for the two companies.
What should investors watch next?
Three checkpoints will test Brown's thesis in the coming quarters. First, whether AMD delivers its guided $17 billion quarter. Second, whether Nvidia's gross margin holds in the 71% to 72% range, which would indicate pricing power intact despite competition. Third, whether more AI labs report running production workloads on Helios, echoing Anthropic's three-day migration claim.
Each data point moves the leverage question in one direction or the other. If AMD clears its guidance and Nvidia's margins hold, the market's willingness to pay 191 times earnings for AMD and a single-digit forward multiple for Nvidia may both need revisiting.
Source: 247wallst.com
More from Marcus Bennett
Show full bio
Senior reporter covering companies, deals and strategy in artificial intelligence and machine learning at AI In Context.
745 articles
Related articles
- Analysts: Nvidia's Wobbling Chip Dominance Hands Samsung an Opening
- NVIDIA’s 27.28% Slide After CEO Exit Leaves $515 Bull Target
- NVIDIA’s 4% Drop Puts a $108B Earnings Test in Focus
- Nvidia Added $150 Billion to Its Buyback. One Strategist Says the Stock Is Still Cheap
- NVIDIA Falls 9% as Board Reportedly Prepares to Oust CEO Jensen Huang