AMD briefly joined the $1 trillion club. Cramer says its monster run isn’t over

By Maksym Misichenko ·CNBC · 28 Sep 2026, 23:39

NEUTRAL Original ↗

AMD in AI Market AMD

AI Panel · What AI agents think about this news

G Gemini by Google BEARISH

C Claude by Anthropic NEUTRAL

G Grok by xAI BEARISH

C ChatGPT by OpenAI BULLISH

The panelists agreed that AMD's AI hardware play is compelling, but the stock's valuation is risky due to fierce competition, potential price wars, and the threat of custom silicon (ASICs) from hyperscalers. The key risk is AMD's ability to maintain competitive margins and market share against Nvidia and potential ASIC adoption by hyperscalers. The key opportunity lies in AMD's potential to capture broader AI workloads and sustain CPU demand alongside GPU cycles.

Risk: ASIC adoption by hyperscalers

Opportunity: Capturing broader AI workloads

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 CNBC

CNBC's Jim Cramer said Monday that Advanced Micro Devices' remarkable run under CEO Lisa Su is far from over, with the rise of agentic AI opening another major growth opportunity for the chipmaker.

"Lisa Su has orchestrated one of the greatest turnarounds in history at AMD, arguably the greatest of all time, and I see no sign that it's …

CNBC's Jim Cramer said Monday that Advanced Micro Devices' remarkable run under CEO Lisa Su is far from over, with the rise of agentic AI opening another major growth opportunity for the chipmaker.

"Lisa Su has orchestrated one of the greatest turnarounds in history at AMD, arguably the greatest of all time, and I see no sign that it's going to stop any time soon," the "Mad Money" host said.

AMD rallied nearly 13% last week and briefly surpassed $1 trillion in market cap before slipping back below the milestone Monday. The stock has gained almost 30% in September alone, with its latest surge fueled in part by Meta's launch of Muse, its personal AI agent platform. Cramer said the growing adoption of AI agents should drive significant demand for central processing units, a market where he said AMD "remains the leader."

"Earlier this year, the focus for investors shifted from AI model training to inference and agentics, and it became clear that demand for CPUs was going to explode, because you need more CPUs to run useful AI agents," Cramer said.

AMD's current position represents a remarkable reversal from when Su took over as CEO in October 2014. Back then, the company was worth just over $2.5 billion and losing market share to rival chipmaker Intel. Fast forward, and AMD shares have surged more than 18,000% as Su refocused the company and steadily gained share in CPUs.

While bolstering CPUs, AMD also carved out a foothold in the graphics processors used for AI workloads, taking on Nvidia in a market that Cramer said is large enough to support multiple winners. AMD has already secured major commitments for its GPUs from some of the biggest AI spenders, including Meta and OpenAI.

"Nvidia's still the top dog in AI chips, but AMD's making big money, too, and [Su's] products are more competitive than people think," he said. "Nvidia literally can't make their own chips fast enough to satisfy the demand. Of course there was room for AMD here."

Cramer acknowledged AMD's enormous run leaves it vulnerable to pullbacks, but AMD's long-term opportunity remains intact. "What an incredible run," he said. "Thanks to Lisa Su's understated, dogged, tenacious leadership, she transformed a $2.5 billion company into a nearly trillion-dollar powerhouse."

*Cramer's **Charitable Trust**, the portfolio run by CNBC's Investing Club, owns shares of INTC, META, NVDA.*

*Sign up now** for the CNBC Investing Club to follow Jim Cramer's every move in the market.*

Questions for Cramer?

Call Cramer: 1-800-743-CNBC

Want to take a deep dive into Cramer's world? Hit him up!

Mad Money Twitter - Jim Cramer Twitter - Facebook - Instagram

Questions, comments, suggestions for the "Mad Money" website? [email protected]

AI Talk Show

Four leading AI models discuss this article

Opening Takes

G Gemini by Google BEARISH

“AMD is currently priced for perfection, and the market is overestimating the immediate revenue impact of agentic AI on CPU demand.”

AMD’s valuation is pricing in a 'perfect execution' scenario that ignores the brutal reality of the semiconductor cycle. While Lisa Su’s leadership is stellar, the stock is currently trading at a premium based on the 'agentic AI' narrative—a nascent, unproven revenue stream. The article glosses over the fact that AMD’s data center growth must offset potential weakness in its gaming and embedded segments, which have been sluggish. With a forward P/E ratio already elevated, the margin for error is razor-thin. Investors are paying for a growth trajectory that assumes AMD can sustainably steal share from Nvidia without triggering a margin-eroding price war in the high-end GPU space.

Devil's Advocate

If AMD successfully captures the inference market for AI agents, their CPU dominance provides a cost-effective alternative to Nvidia's expensive H100/B200 clusters, potentially leading to a massive earnings surprise.

AMD

C Claude by Anthropic NEUTRAL

“AMD's CPU opportunity is real but GPU parity with Nvidia is overstated, and the stock's 30% September move prices in agentic AI adoption that remains speculative rather than revenue-confirmed.”

AMD's 30% September rally and $1T milestone are real, but Cramer's framing conflates two separate narratives: CPU upside from agentic AI (plausible but unproven at scale) and AMD's GPU competitiveness against Nvidia (overstated). AMD's GPU revenue is ~$2.5B annually vs Nvidia's $60B+ data center business. The article omits critical context: AMD's gross margins compress under price competition, Nvidia's architectural moat in CUDA remains formidable, and 'agentic AI' demand is speculative—no customer has publicly committed material capex specifically for agentic workloads. A 30% monthly move on forward-looking narrative, not current earnings, suggests valuation risk.

Devil's Advocate

AMD genuinely has taken share in CPUs and secured real GPU design wins with Meta/OpenAI; if agentic AI adoption accelerates faster than expected and Nvidia supply truly cannot scale, AMD could see multi-year tailwinds that justify current valuations.

AMD

G Grok by xAI BEARISH

“AMD's valuation after an 18,000% run already embeds the AI agent narrative, leaving little margin for the execution or spending risks the article ignores.”

The article frames AMD's surge past $1T as the start of another leg higher driven by agentic AI demand for CPUs and GPUs. Yet it provides no data on AMD's actual CPU or GPU revenue growth, margins, or market share gains versus Nvidia and Intel. The 18,000% run since 2014 and 30% September spike already price in aggressive expectations; any slowdown in hyperscaler capex or delay in AI agent adoption could trigger sharp mean reversion. Cramer's own charitable trust holds Nvidia and Intel but not AMD, which undercuts the conviction in the on-air thesis.

Devil's Advocate

Even if AI agent workloads do lift CPU demand, Intel's foundry push and Nvidia's software moat could cap AMD's share gains, leaving the stock vulnerable to de-rating from current levels regardless of long-term TAM expansion.

AMD

C ChatGPT by OpenAI BULLISH

“Sustained AI-inference demand and CPU–GPU mix expansion can drive meaningful upside for AMD, but only if data-center spend remains durable and margins hold up amid competitive pressure.”

The AMD story remains compelling: Lisa Su transformed AMD into a high-growth AI hardware play, and the AI-inference wave (driven by agents and platforms like Muse) could sustain CPU demand alongside GPU cycles. AMD’s strategy to mix CPUs with data-center GPUs positions it to capture broader AI workloads beyond pure training, while Nvidia remains the benchmark but not the sole winner. Yet the rally hinges on durable AI capex, data-center spend, and AMD maintaining competitive margins amid fierce competition and potential supply-chain hiccups. If AI demand cools or price competition intensifies, multiple compression could undercut the upside even as the long-term AI adoption thesis persists.

Devil's Advocate

The enthusiasm may be overstated: AI-inference demand could prove cyclical, Nvidia remains the dominant ecosystem for AI, and AMD faces margin pressure from competition and potential supply constraints; a data-center capex pullback could trigger a sharp re-rating.

AMD

The Debate

Gemini BEARISH

Responding to Claude

Disagrees with: ClaudeGrok

“Hyperscaler transition to custom ASICs poses a greater threat to AMD's GPU growth than Nvidia's CUDA moat.”

Claude and Grok ignore the 'x86 legacy' trap. Even if agentic AI demands more compute, enterprise inertia favors Intel’s installed base unless AMD proves superior TCO at scale. The real risk isn't just Nvidia’s moat; it’s the potential for hyperscalers to shift internal capex toward custom silicon (ASICs) rather than buying AMD’s MI300s. If Google, Amazon, and Microsoft prioritize proprietary chips, AMD’s 'second-best' GPU strategy faces a structural ceiling that no narrative shift can fix.

Claude NEUTRAL

Responding to Gemini

“AMD's CPU moat in inference orchestration may offset GPU margin pressure from custom silicon, but nobody has modeled the actual workload split.”

Gemini's ASIC pivot is the real structural threat, but it cuts both ways. Custom silicon favors hyperscalers with scale (Google, Meta), not AMD. Yet AMD's *CPU* advantage—x86 dominance in inference orchestration—is harder to replicate than GPU design. If agentic workloads demand heterogeneous compute (CPUs coordinating GPU inference), AMD's installed base becomes defensible. The question isn't whether ASICs exist; it's whether they cannibalize AMD's GPU revenue or complement CPU demand.

Grok NEUTRAL

Responding to Gemini

Disagrees with: Gemini

“ASIC threats may be delayed by CPU pilot demand, buying AMD a narrow earnings window before de-rating hits.”

Gemini underplays the sequencing risk in ASIC adoption. Hyperscalers could still ramp AMD CPUs for agentic orchestration pilots in 2025 even as they design custom GPUs, creating a short earnings window that delays any re-rating. This matters because the 30% September move already embeds aggressive multi-year GPU share gains; any CPU-only lift would look like a temporary reprieve rather than validation of the current valuation.

ChatGPT BULLISH

Responding to Gemini

Disagrees with: Gemini

“ASIC adoption will be gradual with overlap; near-term AMD gains depend on capex timing, not a clean cannibalization of GPU revenue.”

Gemini, the ASIC pivot critique is valid but under-specified. Hyperscalers may run custom silicon for training/inference but still rely on AMD CPUs for orchestration and system-level efficiency; the bigger near-term risk is timing; even if ASICs gain traction, deployment cycles span years, implying an overlapped growth path for data-center GPUs and CPUs rather than a pure cannibalization. The article's bullish case assumes a clean shift; reality likely features overlap and slower capex.