I Just Bought $74,000 in this Crushed AI Stock

Summary

Joseph Hogue argues that the sharp sell-off in AI-related stocks is creating buying opportunities, despite the risk that the AI boom could eventually become a bubble. He says demand for AI infrastructure remains strong, with companies continuing to commit large sums to data centers, chips, networking, memory, and power. He also argues that power availability and component shortages may slow deployments, regardless of political opposition.

His central caution is that AI-generated revenue will eventually need to justify the enormous capital spending. For now, he views the industry as being in the infrastructure-building phase.

Market context and macro themes

• The Nasdaq technology index was down about 6%, while the Global X Artificial Intelligence ETF (AIQ) had fallen nearly 10% in October. The Nasdaq was described as being in correction territory, more than 10% below its peak.
• The video gives varying one-month declines for some stocks, apparently reflecting different measurement periods. Nvidia is cited as down 14% early on and later 22%; Micron is cited as down 26% early and later 34%.
• Goldman Sachs estimates that four hyperscalers plus Oracle will invest more than $1 trillion in AI next year. Hogue says companies recently confirmed or raised spending plans, and that some construction is already contracted or financed.
• Cushman & Wakefield estimates $2.3 trillion of data-center projects are under development. It estimates that each 100 MW of new data-center capacity creates about 1,300 local jobs and $110 million in annual payroll.
• Hogue emphasizes the mismatch between infrastructure timelines: new power-generation and transmission capacity can take up to 10 years, while building and equipping a data center can take roughly 3–6 months.
• He expects political opposition to affect rhetoric and permitting more than underlying demand. He cites OpenAI’s stated willingness to work with local governments on energy, infrastructure, water, and noise issues.
• He warns that AI stocks could eventually face a “dot-com bubble” reckoning if revenues fail to support hundreds of billions or trillions of dollars in capital expenditures.

Stocks, sectors, and figures discussed

AI chips, networking, and infrastructure

Stocks discussed: Nvidia (NVDA), AMD (Advanced Micro Devices), Broadcom (AVGO), Micron Technology (MU), Taiwan Semiconductor (TSM), Marvell Technology (MRVL), and Arista Networks (ANET).

Reported one-month share-price declines include MRVL -50%, ANET -20%, AVGO -21%, NVDA -22%, TSM -22%, AMD -30%, and MU -34%. Marvell was also described as down 24% over five days; Arista was up about 2% over the week after its earnings report.

Company results and forecasts cited:

• AMD: Revenue rose 45% year over year, data-center revenue rose 97%, and operating profit rose 123%. Management forecast Q4 revenue of $17 billion, up 66% year over year. Demand was cited for EPYC CPUs and Instinct accelerators, with customers including Microsoft, Meta, Anthropic, and OpenAI.
• Arista: Revenue rose 58% to $3.6 billion, its fastest growth since its IPO, with operating margin near 50%.
• Broadcom: Reported $16.7 billion in AI revenue in the latest quarter and forecast $115 billion next year. The video says it works with Google, Anthropic, and Meta on TPU chips.
• Marvell: Raised its forecast for custom-chip revenue to $12 billion by fiscal 2029.
• Forward revenue-growth estimates: Nvidia 74%, AMD 55%, Broadcom 50%, Micron 104%, TSM 38%, Marvell 50%, and Arista 38%.

Hogue particularly highlights the gap between share-price declines and business results at AMD, Arista, Broadcom, and Marvell. He likes Marvell’s growth relative to its sell-off, while saying Arista’s valuation looks less attractive after its strong earnings report.

Memory, storage, and data-center bottlenecks

Stocks discussed: Micron (MU), Seagate Technology (STX), Western Digital (WDC), and SanDisk (SNDK).

• Reported one-month declines: Micron 34%, Western Digital 35%, Seagate 42%, and SanDisk nearly 50%.
• Micron reported 378% revenue growth in its latest quarter and an 87% gross margin. It also announced PCI Gen 6 SSDs for AI workloads, including KV Cache applications.
• Seagate and Western Digital were described as growing revenue around 55% annually, with gross margins near 57%. Seagate reportedly had its highest quarterly cash flow to date and is promoting HAMR (heat-assisted magnetic recording) to increase hard-drive capacity.
• The speaker cites a shortage of high-bandwidth memory and says Micron’s capacity is sold out years in advance. He remains skeptical that memory has ceased to be a cyclical industry, characterizing the current environment as a possible supercycle, rather than proof that cyclicality is gone.
• Forward revenue-growth estimates: Micron 104%, SanDisk 102%, and Seagate and Western Digital about 46% each.
• The subtitles contain an apparent sign or transcription inconsistency in the five-day figures for Seagate, SanDisk, and Western Digital. They describe the period as painful and as showing declines, but then render the figures as “up” 26–27%. Treat those weekly figures as uncertain.

AI power and energy

Stocks discussed: Bloom Energy (BE), Vertiv (VRT), Quanta Services (PWR), Monolithic Power Systems (MPWR), and Vistra (VST).

Reported one-month declines: Monolithic Power 19%, Vistra 22%, Quanta 24%, Vertiv 27%, and Bloom Energy 41%. The video also reports substantial weekly declines, including BE -15% and VRT -22%.

• Bloom Energy: Third-quarter revenue rose 179%. Its solid-oxide fuel cells can reportedly be installed on-site in months, rather than requiring a wait of years for conventional power infrastructure.
• Vertiv: Revenue grew 61% and orders grew 100% in the latest quarter. Management raised full-year guidance by more than $500 million and expected $2.8 billion in free cash flow for the year.
• Vistra: Completed its acquisition of Cogentrix Energy, adding 5.5 GW of natural-gas-fired generation capacity.
• Forward revenue-growth estimates: Bloom Energy 68–69%, Vertiv 34%, Quanta 27%, Vistra 10% (with potential upside from the acquisition), and Monolithic Power 43%.
• Hogue sees Bloom, Vertiv, and Monolithic Power as relatively direct AI-power and data-center exposures, while saying Quanta and Vistra could also benefit.
• He says the bottleneck creates potential for pricing power, but that stocks still need to be assessed on valuation and their ability to capture the growth.

Cybersecurity

Stocks discussed: Palo Alto Networks (PANW), CrowdStrike (CRWD), Zscaler (ZS), Okta (OKTA), and Fortinet (FTNT).

• One-month declines cited: Zscaler 6.8%, CrowdStrike 22%, Okta 22%, Palo Alto 25%, and Fortinet 37%.
• Hogue argues that cybersecurity demand should grow regardless of the pace of data-center construction, and that AI creates additional security needs.
• Company developments cited:
  • CrowdStrike launched Agent ID for monitoring AI agents.
  • Palo Alto reported 34% overall growth and 63% growth in next-generation security revenue.
  • Zscaler launched an Agent Security Center.
• Fortinet reported 22% quarterly revenue growth and $865 million in free cash flow, but lowered its annual service-revenue forecast by about $40 million.
• Forward revenue-growth estimates: Palo Alto 20%, CrowdStrike 23%, Zscaler 20%, Okta 10–11%, and Fortinet 15%.
• Growth-adjusted P/E/PEG figures cited: Palo Alto 4.0, CrowdStrike 5.9, Zscaler 1.8, Okta 2.4, and Fortinet about 2.1.
• Hogue says he owns all five, but considers CrowdStrike expensive and would wait for a lower price before buying more. He says he would buy more Zscaler and Fortinet after the declines.

Software and AI-related applications

Stocks discussed: ServiceNow (NOW), Snowflake (SNOW), Palantir (PLTR), Datadog (DDOG), AppLovin (APP), and Cloudflare (NET).

Reported one-month moves: AppLovin +40%, Palantir -16%, Snowflake -17%, Cloudflare -24%, ServiceNow -26%, and Datadog -35%. AppLovin was reported up 67% over five days after earnings; Datadog was down 22% over that period.

• ServiceNow: Reported 20% year-over-year revenue growth, $29 billion in contract commitments, and 70% growth in AI contracts to $1.7 billion. It raised its full-year forecast, although subscription-revenue and backlog growth had slowed relative to prior periods.
• Hogue says he bought 700 ServiceNow shares at $106 each, an investment of roughly $74,000 (about $74,200 at that price). He sees its orchestration platform as a way to manage interactions and costs among AI agents and models, and expects it to be important as companies adopt more AI agents.
• Forward revenue-growth estimates: ServiceNow 20%, Snowflake 30%, Palantir 65%, Datadog 27%, AppLovin 30%, and Cloudflare 31%.
• Growth-adjusted P/E figures cited: ServiceNow 1.06, AppLovin 0.78, Cloudflare 4.0, Datadog 2.4, and Palantir 1.7.
• Hogue says ServiceNow and Palantir look attractive to him at those figures. He is more cautious on AppLovin after its large recent rise and considers Cloudflare expensive at a growth-adjusted multiple of 4.0.
• The video notes that Cloudflare also has cybersecurity exposure and is working in AI-agent orchestration alongside ServiceNow.

Valuation method and investing framework

Hogue describes reviewing 29 AI-related stocks across infrastructure, memory and storage, power, cybersecurity, and software. His comparison process is to:

1. Review share-price performance over the past five days and one month.
2. Examine recent company results, industry developments, and management guidance.
3. Compare expected revenue growth over the next year.
4. Compare valuation using a non-GAAP, earnings-growth-adjusted P/E (PEG) ratio. He explains it as the forward P/E divided by expected earnings growth—not revenue growth.
5. Treat a PEG below 1 as potentially attractive, while still checking other fundamentals and business risks.

Selected PEG figures cited:

• AI infrastructure: Nvidia 0.40, AMD 0.80, Broadcom 0.46, Micron 0.07, TSM 0.49, Marvell 0.60, and Arista 1.4.
• Memory and storage: Micron 0.07, SanDisk 0.09, Seagate 0.15, and Western Digital 0.14.
• AI power: Bloom Energy 0.58 and Vertiv 0.38.
• Cybersecurity and software figures are listed in the relevant sections above.

The speaker cautions that low PEG ratios do not automatically make a stock a buy, particularly when earnings forecasts may be uncertain or industry cycles could turn.

Recommendations and cautions

• Hogue’s broad view is to buy selectively on market weakness, focusing on companies supplying infrastructure the AI industry needs.
• He is especially constructive on memory demand in the near term, while warning that memory chips remain a cyclical business and current pricing may not last indefinitely.
• He favors AI-power and data-center bottleneck exposures, particularly Bloom Energy and Monolithic Power, while also watching Vertiv.
• In cybersecurity, he favors Zscaler and Fortinet at current valuations, but considers CrowdStrike too expensive to add to aggressively.
• In software, he identifies ServiceNow and Palantir as attractive based on the growth-adjusted comparisons, while advising caution on AppLovin after its sharp rally and on Cloudflare’s valuation.
• His longer-term caution is that some AI companies may not generate enough revenue to justify the sector’s massive capital expenditure.

Promotion and disclosures

The video includes a promotion for Gigastar, a private-company investment offering on WeFunder. Hogue says he has invested in Gigastar and holds more than 9,000 shares. He says the platform’s offering had raised more than $2.8 million from over 1,300 investors, at a stated $58 million valuation and $3.52 per share. He also says the company had raised $9 million from venture capital.

He describes a separate investment in the channel’s YouTube revenue, under which investors receive 48% of future revenue, and says investors had received more than $98,000 in dividends across three rounds. He urges viewers to review the offering documents. These are promotional statements, not an independent assessment of the investment’s risks.

No explicit “not financial advice” disclaimer appears in the subtitles.

Presenter/source: Joseph Hogue, CFA, of Let’s Talk Money! He cites Goldman Sachs, Cushman & Wakefield, company earnings reports, and management guidance.