• Avoid treating the recent SPY and QQQ declines as a confirmed buying opportunity; wait for clearer signals from earnings, Treasury yields, and oil prices before adding broad-market exposure.
• Keep NVDA and other semiconductor exposure sized cautiously: AI demand remains a potential growth driver, but custom-chip competition and the sector’s cyclicality could mean further volatility.
• Treat CRWV as high risk rather than a straightforward AI-growth investment; its backlog and demand must translate into revenue despite debt and financing concerns.
• Watch AMZN for evidence that AWS growth and wider Trainium adoption can justify its heavy AI spending; the cited $350 price target is an analyst forecast, not a guarantee.
• Rising oil—near $100 for U.S. crude—and a 10-year Treasury yield around 5.1% are key near-term risks to stocks and housing; monitor them before increasing risk exposure.

S&P 500 and Nasdaq (SPY, QQQ)

• The Nasdaq 100 fell 5.5% for the week, its largest weekly decline since April 2025. The S&P 500 had two declines of more than 1% after roughly 120 days without one.
• The host described the broader market as close to its highs, with the S&P 500 less than 2% below its all-time high, even as many individual stocks had suffered much larger declines.
• The transcript cited the S&P 500’s forward earnings multiple at about 19.7×, slightly below a historical average of roughly 20.3×. It also noted that global equities’ 24-month forward multiple was about 14×, compared with a 10-year average of 16.2×.
• Goldman Sachs data cited on the show showed unusually high put-option volume, suggesting investors were buying protection against further declines. The host viewed this as a possible sign of capitulation, while acknowledging it could also reflect justified concern.
• Risks discussed included higher Treasury yields, elevated oil prices, inflation, uncertainty about AI investment returns, and the possibility that earnings growth may not meet expectations.
• Yardeni’s outlook included an S&P 500 target of 7,900 by year-end, 8,400 by mid-2027, and 10,000 by the end of the decade. He said the earnings outlook was strong but valuations remained uncertain. These were his forecasts, not recommendations.

Takeaways

• The discussion presented a genuine divide: index valuations and investor sentiment may support a rebound, but rising yields, oil prices, and uncertainty over earnings could extend the decline.
• For a general investor, the transcript supports watching earnings and interest rates before treating a short-term bounce as evidence that the downturn is over.

AI Semiconductors and Compute

• The host said compute demand and earnings growth remain central to the bull case for AI-related stocks, but noted that last week’s sell-off raised questions about how durable that demand will be.
• GPU rental prices were said to be rising, which the host viewed as evidence that compute demand still exceeds supply.
• The transcript also emphasized that semiconductors can be cyclical. A change in expectations about the AI investment cycle could cause sharp declines even in companies with strong businesses.
• A BTIG comment cited on the show warned that semiconductor stocks could be forming a longer-term topping pattern, and suggested a rebound might be an opportunity to reduce exposure. This was a quoted market view, not a firm forecast.

Takeaways

• The transcript’s central investment question is whether AI demand translates into sustained, profitable growth—not simply whether companies are spending heavily on data centers.
• Investors considering semiconductor exposure should account for the sector’s cyclicality and the possibility of large price swings if expectations change.

NVIDIA (NVDA)

• NVIDIA was described as having fallen from nearly $242 to below $210 during the sell-off, before trading near $218 in premarket trading.
• Amazon’s decision to make its Trainium chips available to more PyTorch developers was discussed as competitive pressure, as were Microsoft’s custom chips. The host noted that major customers are also developing competing products.
• The transcript also highlighted NVIDIA’s expanding CPU opportunity. It cited a reported deal to supply Vera CPUs to AI infrastructure company Modal, and NVIDIA’s claim that its CPU business could reach $20 billion in revenue this fiscal year.
• NVIDIA CEO Jensen Huang’s view that the semiconductor industry could grow substantially over the next several years was presented as a key part of the long-term bull case.

Takeaways

• NVIDIA’s opportunity depends on continued demand for AI compute, but competition from customers developing their own chips is a risk explicitly raised in the discussion.
• The host’s broader message was that investors need to believe AI infrastructure demand will persist; a short-term price rebound alone does not resolve that question.

AMD (AMD)

• AMD was said to have fallen from about $658 to roughly $547 during the previous week, before rebounding in premarket trading.
• The host pointed to CPU demand as a potential growth driver, while noting that NVIDIA’s Vera CPU push could add competition.
• The discussion suggested that demand may be large enough to support multiple CPU suppliers, but did not establish that AMD would necessarily gain share.

Takeaways

• AMD’s investment case is tied to both continued AI-related compute spending and its ability to compete as NVIDIA expands beyond GPUs.
• Watch company guidance and customer demand rather than assuming that industry growth will benefit every chipmaker equally.

CoreWeave (CRWV)

• CoreWeave was repeatedly described as a focal point for market anxiety about neocloud companies and AI-related debt.
• Its shares were reported to have fallen below $60 at one point, rebounded to about $70, and then returned to roughly $61. The host said the company’s credit-default swaps and bonds were signaling concern about its ability to finance its expansion.
• The host cited a reported backlog of about $130 billion, potentially approaching $150 billion when the company reports results. The key questions raised were whether that backlog is dependable, how concentrated revenue is in Microsoft and OpenAI, and how expensive future borrowing will be.
• The host argued that a reported Microsoft contract cancellation involving another neocloud was caused by power-supply limitations at a specific site, not a lack of customer demand. He cautioned that this distinction matters when assessing the AI infrastructure market.
• Goldman Sachs reportedly assigned CoreWeave a $100 price target. The host contrasted that bullish sell-side view with the more cautious signals from the bond market.

Takeaways

• CoreWeave represents a high-risk way to invest in AI infrastructure: its upside depends on converting demand and backlog into revenue, while its financing costs and debt load are major risks discussed in the transcript.
• The transcript did not resolve whether the market’s credit concerns are temporary or reflect deeper problems. Investors should distinguish customer demand from the company’s ability to finance and deliver capacity.

Nebius (NBIS)

• Nebius was described as having experienced a steep decline during the week, at one point approaching its prior lows, before trading around $170–$180 during the market session.
• The company was grouped with other neoclouds exposed to investor concerns about data-center funding, credit conditions, and the AI build-out.
• The host said the market was trying to determine whether the sell-off reflected temporary fear and trading pressure or a more fundamental reassessment of AI infrastructure spending.

Takeaways

• Nebius offers exposure to AI infrastructure, but the transcript highlighted financing and valuation uncertainty as material risks.
• A rebound in the share price would not, by itself, confirm that those risks have passed.

Micron (MU) and SanDisk (SNDK)

• The host argued that AI data centers and custom chips will require substantial memory, and cited industry reports suggesting memory supply could remain constrained until at least 2028.
• Micron was described as having suffered a sharp sell-off despite those demand arguments. The discussion also noted concerns about future supply from Chinese memory producer CXMT and the possibility of pressure on pricing.
• SanDisk was highlighted as a memory stock that had fallen sharply but showed more resilience during the session. Both companies were discussed in the context of the wider memory cycle.
• The host acknowledged that memory is cyclical and that strong current demand does not guarantee that pricing or earnings will remain strong.

Takeaways

• The memory bull case depends on AI-related demand continuing to outpace supply; the principal risk discussed was that new capacity or competition could weaken pricing.
• Investors should compare company guidance and supply expectations with the broad AI-demand narrative rather than assuming all memory stocks will move together.

Broadcom (AVGO), Marvell (MRVL), Intel (INTC), and Custom Chips

• The transcript discussed hyperscalers building or expanding their own chips, including Amazon’s Trainium and Microsoft’s Maya chips. This could reduce reliance on outside suppliers over time.
• Marvell was mentioned as having ties to hyperscalers, including reported collaboration with Google. The host noted that the market may be rewarding hyperscalers for bringing more of the AI supply chain in-house while penalizing some external suppliers.
• Intel was mentioned as another potential competitor in CPUs, but the host did not suggest that NVIDIA’s CPU expansion would necessarily eliminate demand for other suppliers.

Takeaways

• Custom-chip development is a long-term competitive risk for outside chip suppliers, but the transcript also emphasized that overall compute demand may be large enough to support multiple vendors.
• The key distinction is whether a company can retain a role in customers’ infrastructure as hyperscalers increasingly design their own hardware.

Applied Optoelectronics (AAOI), Lumentum (LITE), and Coherent (COHR)

• These optical and photonics-related companies were described as having suffered very large declines during the sell-off, with some names down roughly 30%–50% from recent highs.
• The host characterized them as part of the AI “picks and shovels” trade, but warned that they can behave cyclically when investors question the pace of data-center investment.
• The stocks showed some attempted rebounds during the session, but the host cautioned that early trading gains could reverse.

Takeaways

• These companies offer exposure to data-center connectivity and AI infrastructure, but the transcript highlighted substantial volatility and sensitivity to changes in the investment-cycle narrative.
• A sharp decline can create an opportunity only if the underlying demand and earnings outlook remain intact; the discussion did not establish that they had bottomed.

Amazon (AMZN)

• Amazon shares rose during the session after the company announced that its Trainium AI chips would be available to more PyTorch developers.
• The transcript also cited CEO Andy Jassy’s view, relayed by a podcast guest, that compute shortages could persist beyond 2028.
• Morgan Stanley reportedly raised its price target to $350 and forecast AWS growth of 47% in 2027 and 42% in 2028. These were analyst estimates, not company guidance.
• The host said the market may be underestimating AWS’s growth and discussed the possibility that Trainium could become a large business in its own right.
• Risks discussed included high capital spending, negative free cash flow, and uncertainty about how much AWS growth depends on OpenAI, Anthropic, and other AI customers.

Takeaways

• Amazon’s case combines AWS growth with its effort to sell in-house AI chips. The transcript suggests the market may reward that vertical integration, but the return on large capital commitments remains important.
• Watch AWS growth, chip adoption, and the concentration of AI-related demand when assessing the thesis.

Microsoft (MSFT)

• Microsoft was discussed as a major AI customer and a developer of its own Maya chips, as well as a key customer of CoreWeave.
• The transcript said investors were concerned about the exposure of AI infrastructure providers to Microsoft and OpenAI, and questioned how durable those commitments would be.
• Microsoft’s shares were described as having rebounded from recent lows, but the host also discussed reducing some Microsoft exposure in favor of Amazon as one investor’s portfolio decision—not as a general recommendation.

Takeaways

• Microsoft remains closely tied to the AI build-out, but its own custom-chip efforts and its role as a customer of third-party infrastructure providers create a complex competitive picture.
• The transcript emphasized the importance of assessing capital spending and the eventual revenue generated by AI services.

Meta Platforms (META)

• Meta was due to report earnings during the week. The host said investors would look for details on future capital spending, whether its AI products—particularly Muse—can be monetized, and whether AI has broader use cases.
• Meta was described as an important test of whether hyperscaler spending on AI can be justified by future growth.
• The host expressed concern that Meta’s recent earnings calls had not communicated its strategy effectively. Meta’s shares were also discussed as having fallen below $700 during the session.
• Morgan Stanley’s strategist, as quoted in the transcript, said earnings and management commentary could help determine whether investors continue to view AI adoption as a multi-year trend.

Takeaways

• Meta’s results and commentary could affect broader sentiment toward AI capital spending, not just the company’s own shares.
• Investors should focus on evidence of monetization and returns on AI investment, rather than capital-spending plans alone.

Tesla (TSLA)

• Tesla was expected to report earnings during the week, with attention on growth and its plans for autonomous driving and AI-related products.
• The host said Tesla had rebounded in premarket trading after a difficult period and noted that its shares had previously risen following a Roadster announcement.
• A reported claim from Satrini Research said Tesla could be acquired by SpaceX for $420 per share, potentially near year-end. The host said the claim was unconfirmed and that the transcript did not establish whether such a deal would happen.
• Tesla’s proxy materials reportedly did not mention a merger, which the host viewed as a reason not to assume a near-term transaction.

Takeaways

• Tesla’s investment case in the discussion rested more on earnings, robotaxi progress, and AI ambitions than on the unconfirmed SpaceX acquisition rumor.
• Treat the $420 figure as a reported rumor, not a confirmed offer or price target.

SpaceX (private)

• SpaceX was expected to report during the week and was discussed as an AI and data-center-related company as well as a potential participant in a possible Tesla transaction.
• The host said investors would look for evidence that its annualized recurring revenue from data-center activities is continuing to grow.
• The possible acquisition of Tesla for $420 per share was described as an unconfirmed research claim. The transcript did not provide confirmation from either company.

Takeaways

• SpaceX was presented as a private investment theme tied to AI infrastructure and possible corporate transactions, but the discussion provided no confirmed terms for an investment or merger.
• The company’s reported growth and any formal disclosures would be more informative than speculation about a Tesla deal.

Netflix (NFLX)

• Netflix was scheduled to report during the week. The host described it as a test of whether a legacy media company can demonstrate growth and whether consumers remain willing to pay for entertainment.
• The transcript noted that Netflix had already made layoffs and that its CEO had indicated growth was weaker than expected.
• The host expected the earnings reaction could be negative but acknowledged that the stock’s prior decline could make the share price sensitive to any positive surprise. No specific price target was given.

Takeaways

• The key issues are subscriber or revenue growth and consumer demand, not merely whether the company meets near-term expectations.
• The transcript highlighted the risk that weak growth could weigh on the shares, while also noting the possibility of a relief rally if expectations are already low.

Bloom Energy (BE) and Western Digital (WDC)

• Both companies were scheduled to report during the week and were framed as AI-related plays—Bloom Energy through energy and Western Digital through memory and storage.
• Western Digital and Seagate were described as having experienced steep recent declines, with Western Digital reportedly down about 20% over a month before a session rebound.
• The host said their results could help clarify whether demand for data-center power, memory, and storage remains strong.

Takeaways

• These earnings reports may offer evidence about the broader AI infrastructure cycle beyond the largest chip companies.
• Investors should distinguish company-specific results from short-term price rebounds.

Anthropic and OpenAI (private)

• The transcript discussed Anthropic’s reported pledge of 5% of its shares to a public-benefit fund, as well as co-founders’ and employees’ stated plans to donate equity. Elon Musk was quoted as supporting giving shares directly to the public.
• The host said the market was questioning whether Anthropic and OpenAI can generate enough revenue to justify the scale of AI infrastructure spending. A figure cited in the discussion suggested the two companies might need to reach $1.5 trillion in combined revenue by the end of the following year for some estimates of the build-out to make sense; the host said that prospect was alarming.
• Anthropic and OpenAI were also discussed as important customers for cloud providers and data-center companies, creating concern about customer concentration.
• The host argued that the AI investment thesis may not depend solely on either company, but said investors still need to understand who will ultimately pay for the infrastructure.

Takeaways

• These private companies are important indicators of AI demand and monetization, but the transcript underscored that ambitious funding and spending plans are not proof of sustainable returns.
• For public-market investors, the relevant signals are customer commitments, revenue growth, and whether AI services produce enough value to support continued infrastructure spending.

Cerebras and Tenstorrent (private companies, as discussed)

• Cerebras was described as having fallen to about $123 after a reported IPO allocation price of $180. The host cited a valuation of roughly 75× sales, even after the decline.
• The discussion said Cerebras’ technology is viewed positively by some investors, but stressed that the valuation remained high and that the stock could face selling from investors who do not want to hold a sharply declining IPO.
• Tenstorrent, founded by Jim Keller, was reported to have raised $1.5 billion at a $6.2 billion valuation. The host said this could add competition but could also indicate investor interest in the market for alternative AI chips.

Takeaways

• The transcript distinguished confidence in a company’s technology from confidence in its valuation. Both matter, particularly for high-growth, high-multiple AI companies.
• Competition is increasing, and the discussion did not conclude that Cerebras had reached a bottom.

Palantir (PLTR)

• Palantir was mentioned as a potential provider of open-weight AI models and as a possible beneficiary of the U.S. government’s move to restrict the use of DeepSeek-related products in civilian government systems.
• The host said the policy could favor U.S. models, including open-weight offerings, but did not present it as a major change to the broader AI investment outlook.

Takeaways

• The transcript identified government adoption and restrictions on foreign AI products as potential opportunities for domestic providers, but did not quantify the financial impact for Palantir.
• Treat the policy as a possible catalyst, not evidence by itself of increased revenue.

APLD (Applied Digital)

• The host noted that Applied Digital had fallen roughly 31% over the preceding month.
• The company’s CEO and president were reported to have bought about $6.7 million of shares. The host viewed the insider purchases as notable, though no detailed company-specific investment thesis was provided.

Takeaways

• Insider buying may be worth monitoring, but it does not establish that the stock has bottomed or that the company’s financing and demand outlook are secure.
• The transcript’s broader concerns about neocloud financing remain relevant to data-center companies.

Defense and Drone Stocks: Anduril (private), AeroVironment (AVAV), Kratos (KTOS)

• The host said he had begun building a position in Anduril and viewed drone-related companies—including AeroVironment and Kratos—as potential buying opportunities after recent declines.
• The stated rationale was their outlook for earnings and EBITDA growth, the size of the market opportunity, and a favorable view of Anduril’s partnership with Palantir.
• Anduril is private; the transcript did not provide a ticker or price target.

Takeaways

• The host’s thesis was that the recent share declines may create an opportunity if expected growth and the addressable market remain intact.
• The transcript did not provide company-specific valuation analysis or evidence that the sell-off had ended, so the bullish view remains conditional.

Nubank (NU), SoFi (SOFI), and Robinhood (HOOD)

• A viewer asked about Nubank, citing its reported 140 million users, low customer-service costs, and potential to grow revenue per customer. The host said he had not done enough work to invest in it, but agreed with the general argument that its valuation could be low relative to its reported growth.
• The host also said higher interest rates could be difficult for financial technology companies, including SoFi, Robinhood, Affirm, and Coinbase.
• Robinhood and SoFi were mentioned during the market session, but the transcript did not provide detailed company-specific analysis.

Takeaways

• Nubank’s user growth and ability to increase revenue per customer were presented as potential strengths, but the host also flagged Brazil-related and interest-rate uncertainty.
• For fintechs, assess customer growth and monetization alongside the impact of rates and broader financial conditions.

Bitcoin (BTC)

• Bitcoin was described as trading around $82,000, after briefly reaching roughly $79,000 during the preceding week.
• The host said some investors believe Bitcoin could help address long-term concerns about government debt, but noted that Bitcoin was behaving like a risk-on asset and therefore depended on liquidity conditions.
• The transcript linked the outlook for risk assets to inflation, Federal Reserve policy, oil prices, and whether the conflict with Iran de-escalates.

Takeaways

• The discussion did not present Bitcoin as a hedge that would necessarily rise during market stress. Its price was described as sensitive to liquidity and risk appetite.
• Investors should consider the transcript’s stated link between Bitcoin, interest-rate expectations, and broader market sentiment.

Oil, Treasury Yields, and Housing

• Oil was reported near $100 for U.S. crude, with Brent near $110 at one point. The host said higher energy costs could keep inflation elevated and make it harder for the Federal Reserve to ease policy.
• The 10-year Treasury yield was around 5.1% and rising during the session. The discussion linked higher yields to pressure on borrowing costs and stock valuations.
• A home-builder sentiment index was reported at 29, its lowest level since 2012 and below the level of 50 associated with negative sentiment. Builders cited higher mortgage rates, fuel costs, and construction costs.
• The host said higher mortgage rates and weak housing affordability were risks for homebuilders and housing-related companies.

Takeaways

• Oil prices and Treasury yields were presented as major cross-market drivers: they can affect inflation, borrowing costs, and valuations across both growth stocks and housing-related investments.
• The transcript suggested that a lasting equity-market recovery may be difficult if yields and energy prices continue rising.