Goldman Sachs' Urgent Warning: This AI Stock Crash Just Created A MASSIVE Valuation Disconnect — Transcript
Full transcript
- 0:05Wall Street is dumping semiconductor
- 0:07stocks. AI valuations are collapsing and
- 0:10investors are questioning whether the
- 0:12biggest technology spending boom in
- 0:14history has finally gone too far. But
- 0:16beneath the panic, something
- 0:18extraordinary is happening. Earnings
- 0:20forecasts are accelerating while stock
- 0:22prices are falling. Following November's
- 0:25midterm elections, fears surrounding
- 0:27data center regulations, electricity
- 0:30shortages, AI financing, and slowing
- 0:33infrastructure investment triggered a
- 0:35brutal semiconductor sell-off. Some of
- 0:37the industry's biggest names have fallen
- 0:39between 25% and 60% from recent highs.
- 0:43Yet, Goldman Sachs sees a valuation
- 0:46disconnect. Its semiconductor research
- 0:48highlights leading AI businesses trading
- 0:50at approximately 12 and 13 times
- 0:53conservative 2028 earnings assumptions.
- 0:56Meanwhile, institutional analysts are
- 0:59increasingly questioning whether markets
- 1:00have confused short-term uncertainty
- 1:03with deteriorating fundamentals.
- 1:05Remember Micron's extraordinary rally?
- 1:08The memory manufacturer surged from
- 1:10roughly $100 in early 2025 to above
- 1:14$1,200
- 1:15at its June 2026 peak, an increase
- 1:18exceeding 1,000%.
- 1:21But 2026 has also demonstrated how
- 1:24violently sentiment can reverse even
- 1:26when earnings continue climbing. Now,
- 1:29another disconnect is emerging. AI
- 1:31accelerators, custom silicon, high-speed
- 1:34storage, and advanced memory remain
- 1:36essential to the next generation of
- 1:38computing. Today, we're examining five
- 1:41companies positioned across that
- 1:43infrastructure chain. An emerging AI
- 1:46challenger, a custom chip powerhouse, a
- 1:49storage specialist, a memory giant, and
- 1:51the industry's dominant computing
- 1:53platform. Their valuations, earnings
- 1:56projections, institutional assessments,
- 1:58and upcoming catalysts reveal why this
- 2:01sell-off deserves a closer examination.
- 2:04First is Advanced Micro Devices, ticker
- 2:07AMD, where a sharp correction has
- 2:10collided with some of the strongest
- 2:11growth projections in semiconductors.
- 2:14AMD's third quarter revenue surged 45%
- 2:17to 13.4 4 billion, while data center
- 2:21revenue jumped 97% to $8.5 billion.
- 2:25Management expects fourth quarter
- 2:27revenue around $17 billion, with data
- 2:30center sales potentially rising another
- 2:3250% sequentially, but 2027 is where the
- 2:36numbers become extraordinary.
- 2:38CEO Lisa Sue projects revenue growth
- 2:41exceeding 80%, potentially lifting
- 2:44annual sales toward $94 billion. The
- 2:47catalysts are AMD's Instinct MI455
- 2:51accelerators, Helios Rackcale AI
- 2:54systems, and EPYC server processors.
- 2:57Demand for EPYC is accelerating as AI
- 3:00agents require CPUs to coordinate
- 3:03applications, retrieve data, and manage
- 3:05increasingly complex workloads. Google's
- 3:09next generation TPU infrastructure also
- 3:11incorporates EPYC processors, extending
- 3:15AMD's opportunity into custom AI
- 3:17systems. Despite these developments, AMD
- 3:20has experienced a correction approaching
- 3:2328%. Research analyst Giannis Zorpanos
- 3:26estimates fair value between $75 and
- 3:30$765
- 3:31per share, while Summit Research's
- 3:34discounted cash flow analysis suggests
- 3:36approximately $634.
- 3:39[music] Quarterly free cash flow
- 3:41declined to $554 million as inventory
- 3:45expanded ahead of anticipated shipments,
- 3:47making execution critical. The
- 3:50opportunity is clear. AMD is becoming a
- 3:53broader AI infrastructure supplier, not
- 3:56merely an alternative GPU manufacturer.
- 3:59Moving from AI computing challengers to
- 4:01custom silicon, our second company is
- 4:03Broadcom, ticker AVGO, whose growth
- 4:07prospects look dramatically different
- 4:09from its recent share price performance.
- 4:11Broadcom designs specialized AI
- 4:14accelerators, high-speed networking
- 4:16equipment, and enterprise infrastructure
- 4:18software. Its custom chips help
- 4:20hyperscalers optimize computing
- 4:22performance and reduce dependence on
- 4:25generalpurpose accelerators. Following
- 4:27November's sell-off, shares declined
- 4:29approximately 15%, including a single
- 4:32session loss exceeding 10%. Yet,
- 4:35Broadcom generated $ 31.9 billion in
- 4:39free cash flow over the first 9 months
- 4:41of its fiscal year. The company's next
- 4:43growth engine is custom AI
- 4:45infrastructure. Research discussing
- 4:48Anthropic's financing arrangements cites
- 4:50approximately $161 billion in
- 4:53infrastructure commitments and $63
- 4:56billion in secured financing supporting
- 4:58future deployments. These commitments
- 5:00are not equivalent to immediate Broadcom
- 5:03revenue, but they illustrate the scale
- 5:05of potential demand. Meanwhile, Google's
- 5:08expanding TPU ecosystem creates another
- 5:11substantial opportunity. Wall Street
- 5:13projections cited in recent research
- 5:15place fiscal 2027 revenue near $174
- 5:19billion, while earnings estimates imply
- 5:22approximately 70% growth. At around $300
- 5:26per share, Broadcom was trading near 15
- 5:29times projected fiscal 2027 earnings.
- 5:32Analyst Bowden Cuchery estimates fair
- 5:35value at $430,
- 5:38approximately 43% above that price. The
- 5:41intriguing disconnect. A cashg
- 5:43generating AI infrastructure leader is
- 5:46being repriced while its prospective
- 5:48custom chip deployments continue
- 5:50expanding. Our third opportunity shifts
- 5:52from processing information to storing
- 5:54it. SanDisk, ticker SNDK, a company
- 5:58whose stock has plunged roughly 60% from
- 6:01its highs despite extraordinary
- 6:03operating results. SanDisk produces NAND
- 6:06flash memory used in enterprise
- 6:08solid-state drives, cloud
- 6:10infrastructure, and consumer storage. As
- 6:13AI systems generate enormous data sets,
- 6:16storage capacity becomes increasingly
- 6:18important alongside computing power.
- 6:21SanDisk's fiscal first quarter revenue
- 6:23surged approximately 378%
- 6:26year-over-year to 11 billion, while
- 6:29gross margins reached an exceptional
- 6:3285%.
- 6:33Adjusted free cash flow approached $5.7
- 6:36billion, reflecting stronger memory
- 6:39pricing and constrained industry supply.
- 6:42Management's second quarter revenue
- 6:43outlook centers around 12.4 billion,
- 6:47while data center nan demand is
- 6:49projected to grow approximately 40% in
- 6:52calendar 2027. Now consider the
- 6:55valuation. Recent research places
- 6:57SanDisk's forward earnings multiple near
- 7:003.5 times, reflecting investor concerns
- 7:03that today's exceptional margins may
- 7:05eventually normalize. Analyst Gita Zizus
- 7:08estimates fair value near $1,163
- 7:12per share, approximately 40% above the
- 7:15recent price. SanDisk also reported
- 7:18around $5.1 billion in cash, no debt,
- 7:21and an upcoming 8for1 stock split
- 7:24scheduled for November 19th. Its
- 7:26emerging high bandwidth flash technology
- 7:29provides another potential catalyst for
- 7:31AI inference. SanDisk's investment story
- 7:34centers on whether today's storage
- 7:36shortage can sustain elevated earnings
- 7:38longer than the market expects. And that
- 7:40brings us to our fourth company, Micron
- 7:43Technology, ticker MU, where the AI
- 7:46memory boom could eventually evolve into
- 7:48something even bigger, a roboticsdriven
- 7:51demand cycle. Micron manufactures DRAM,
- 7:54high bandwidth memory, and NAND storage,
- 7:57making its products essential across AI
- 8:00accelerators, servers, smartphones, and
- 8:02autonomous systems. Despite declining
- 8:05approximately 35% from recent highs,
- 8:08Micron's fiscal 2026 net income reached
- 8:11roughly $85 billion. At a recent price
- 8:15near $690,
- 8:17shares traded around 9 times trailing
- 8:19earnings and approximately four times
- 8:21projected fiscal 2027 earnings. Demand
- 8:25visibility is another important
- 8:26distinction. Recent research indicates
- 8:29more than 75% of fiscal 2027 shipments
- 8:33are committed. While competitor SKHinix
- 8:36expects industry memory demand to remain
- 8:38exceptionally strong through 2030,
- 8:41robotics could extend that opportunity.
- 8:44Nvidia's Jetson Thor platform already
- 8:46supports 128 GB of LPDDR 5x memory,
- 8:51illustrating the growing memory
- 8:53requirements of physical AI. Micron's
- 8:56board has also authorized $ 35.16
- 8:59billion in share repurchases with
- 9:02December 9th identified as an important
- 9:04potential resumption date. Research
- 9:06analyst James Ford maintains a $1,500
- 9:10valuation objective for October 2027,
- 9:13while Bowden Cuchery estimates fair
- 9:16value near $1,791.
- 9:20The question is whether investors are
- 9:22anticipating another traditional memory
- 9:24downturn while overlooking a longer
- 9:26AIdriven expansion. Finally, we reach
- 9:29Nvidia, ticker NVDA, the company at the
- 9:33center of the AI revolution and perhaps
- 9:36the most closely watched valuation reset
- 9:38in the industry. Nvidia's competitive
- 9:41advantage extends far beyond GPUs. Its
- 9:44Blackwell accelerators, upcoming Vera
- 9:46Rubin architecture, high-speed
- 9:48networking and CUDA software ecosystem
- 9:51create an integrated platform for AI
- 9:53computing. Yet shares recently fell
- 9:56approximately 26% in 1 month. According
- 9:59to November research, Nvidia's forward
- 10:02earnings multiple compressed to around
- 10:0419 times compared with its 5-year
- 10:06average near 42 times. Its forward PEG
- 10:10ratio declined to approximately 0.38.
- 10:13Consensus estimates cited in that
- 10:15research project, fiscal 2028 earnings
- 10:18near $15.92
- 10:21per share, representing approximately
- 10:2371% growth. Institutional interest
- 10:27remains significant. Goldman Sachs has
- 10:29highlighted compelling semiconductor
- 10:31valuations, while Mizuo's Jordan Klene
- 10:34has described Nvidia's valuation as
- 10:36unusually depressed. NVIDIA also faces
- 10:39competition from custom accelerators,
- 10:42export restrictions and questions
- 10:44surrounding AI infrastructure financing,
- 10:47but hyperscaler investment, Blackwell
- 10:49deployments, networking demand, and the
- 10:52Vera Rubin product roadmap remain
- 10:54important growth catalysts. With another
- 10:56earnings report approaching on November
- 10:5817th, investors will be watching revenue
- 11:01guidance, margins, [music] and customer
- 11:03demand closely. The central question,
- 11:06has Nvidia's share price fallen faster
- 11:08than its underlying earnings
- 11:10opportunity?
About this transcript
This page contains the full transcript of Goldman Sachs' Urgent Warning: This AI Stock Crash Just Created A MASSIVE Valuation Disconnect by Millionaires Investment Secrets, generated from the public captions YouTube serves with the video. The transcript has 1,247 words across 252 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.
What you can do with it
Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.
Free YouTube transcript tool
YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.