Oracle Crashes 11%: The Deadly ROIC Gap Could Kill the AI Hyperscalers in 2026
Oracle Crashes 11%: The Deadly ROIC Gap Could Kill the AI Hyperscalers in 2026
YouTube47 min 25 sec
Watch on YouTube
Note: AI-generated summary based on third-party content. Not financial advice. Read more.
Quick Insights

Consider investing in Tesla (TSLA) ahead of its unsupervised Robotaxi launch in Austin, expected in approximately three weeks, which could be a major catalyst. A core investment theme is rotating out of software and into the physical hardware and infrastructure powering AI, such as power, data centers, and optical fiber. For a specific hardware play, look at Corning (GLW), as its specialty glass and fiber optics are critical for the AI infrastructure build-out. Watch for continued outperformance in small-cap stocks, via the IWM ETF, as a sign of a broadening economic recovery and rotation away from large-cap tech. While cautious in the short-term, Bitcoin (BTC) is viewed as a highly attractive long-term asymmetric bet, with a sustained break above $92,000 signaling a potential major rally.

Detailed Analysis

Oracle (ORCL)

  • The stock had a bad week, plummeting 11%.
  • While earnings beat expectations, revenue came up short, causing concern among investors.
  • There are questions about whether the company's significant investments in AI are justified.
  • The company is facing balance sheet issues, with its Credit Default Swaps (CDS) rising, which indicates increasing perceived risk.

Takeaways

  • The sentiment surrounding Oracle is currently bearish due to its revenue miss and concerns over its spending and balance sheet.
  • For investors who are heavily invested in other AI stocks, the speaker suggests that Oracle CDS could be considered as a potential hedge. This would act as a form of insurance that pays out if Oracle's credit risk increases significantly, potentially offsetting losses elsewhere in an AI sector downturn.

AI Hyperscalers / Magnificent Seven

  • The speaker suggests that the Magnificent Seven and other AI hyperscalers may be at a relative top.
  • A major concern is the "ROIC (Return on Invested Capital) air gap": these companies are spending billions in capital expenditures (CapEx) on AI hardware now, but the revenue from these investments may not materialize for 3 to 4 quarters.
  • This heavy spending is expected to worsen their balance sheets, reduce their ability to do stock buybacks, and could lead to multiple compression (their stock price becomes lower relative to their earnings).
  • The transition to NVIDIA's Blackwell chip architecture is described as the "most complex in history," requiring a jump from 30 kilowatts to 130 kilowatts per rack, stalling some hardware progress in late 2024.
  • Competition is expected to increase dramatically with massive IPOs from companies like SpaceX, OpenAI, Anthropic, and XAI potentially coming in 2025-2027.

Takeaways

  • The sentiment is cautious to bearish in the short-to-medium term for the largest AI players.
  • Investors should be aware of the significant financial pressure these companies face due to massive, front-loaded spending. This could lead to stock underperformance even as AI technology progresses.
  • The influx of major AI IPOs will create more competition for investment dollars, potentially drawing capital away from the current leaders.

Investment Theme: AI Hardware over Software

  • A core thesis of the podcast is a major shift in value from software to physical hardware.
  • SaaS (Software-as-a-Service) companies are seen as being in a difficult position.
    • They are priced as growth companies but face an "uphill battle" against new, nimble AI-native startups that can operate with lower costs.
    • The speaker believes their valuations are at risk of being destroyed by hyper-competition.
  • Hardware and Infrastructure companies are the recommended area of focus.
    • There has been chronic underinvestment in the physical components needed for the AI revolution (power, data centers, cooling, optical fiber, etc.).
    • Demand for these physical goods is now "insatiable" and necessary to fuel the next stage of AI.

Takeaways

  • The speaker is very bullish on hardware and bearish on the long-term prospects of traditional SaaS companies.
  • Investors are advised to rotate their portfolios away from software and towards the companies that build the physical foundation for AI.
  • Specific areas of interest mentioned include:
    • Optical fiber and specialty glass (Corning was cited as an example).
    • Power infrastructure, including gas turbines, transformers, and solar names.

Tesla (TSLA)

  • Elon Musk announced that unsupervised, full self-driving Tesla Robotaxis will be operating in Austin, Texas, in approximately three weeks. These vehicles will have no one inside, not even a safety passenger.
  • This event is framed as the "opening act of embodied AI"—the moment AI moves from digital models to interacting with the physical world.
  • The build-out of Tesla's Colossus 2, the first gigawatt-scale data center, is a key enabler for processing the massive amounts of video data required for this.
  • The speaker notes that there is a significant negative bias against TSLA among institutional investors, meaning many are not positioned for this potential catalyst.

Takeaways

  • The sentiment is extremely bullish.
  • The upcoming Robotaxi launch is presented as a monumental catalyst that the market is underappreciating.
  • The speaker views TSLA as the single best-positioned company to capitalize on the "embodied AI" trend and suggests it is a compelling investment, particularly because many professional investors are not involved.

LIDAR Technology

  • While Tesla famously does not use LIDAR, the speaker notes that most other autonomous systems will, including those from Waymo, as well as military vehicles, mining equipment, and robotics.
  • The demand for LIDAR is expected to "absolutely increase" once the next generation of AI chips (NVIDIA's Blackwell) becomes widely available, as this will unlock the next wave of robotics and autonomous machines.

Takeaways

  • The sentiment on the LIDAR sector is bullish.
  • This is presented as a "picks and shovels" play on the "embodied AI" theme. As robotics and autonomy expand beyond just Tesla, the demand for this key sensor technology is expected to grow significantly.

Corning (GLW)

  • Corning is highlighted as a prime example of a hardware company that was stagnant during the software-dominated era but is now poised for growth.
  • Its products, including optical fiber and specialty glass, are critical components for the AI data center and infrastructure build-out.

Takeaways

  • The sentiment is bullish.
  • GLW is presented as a specific, actionable stock idea that fits the "hardware over software" theme. The thesis is that demand for its essential products will surge as the AI infrastructure build-out accelerates.

Small Caps (IWM)

  • Small-cap stocks, represented by the IWM ETF, have been outperforming large-cap tech (QQQ), with the IWM vs. QQQ ratio having its best week since August.
  • This trend is historically correlated with periods of rising PMIs (Purchasing Managers' Index), which signals an expanding economy.
  • The outperformance is part of a broader "reflation theme" and a potential rotation from growth stocks into value and more economically-sensitive companies.

Takeaways

  • The sentiment is bullish on small caps relative to large-cap tech.
  • The outperformance of IWM is a key market signal to watch. If the economy continues to strengthen and PMIs rise, this rotation could have significant room to run, favoring investments in smaller companies.

Bitcoin (BTC)

  • The speaker describes Bitcoin as the "most asymmetric bet" of his career, with a highly favorable risk/reward profile for the long term.
  • The long-term bullish case is based on its historical outperformance and its potential role as a store of value in a future heavily influenced by AI's impact on fiat currencies.
  • Short-term Technicals: The price is currently in a downtrend. A key signal for a reversal would be a break and sustained hold above the $92,000 - $93,000 level for three consecutive days.
  • Adoption is growing:
    • U.S. regulators are allowing banks to provide custody for customers.
    • The CFTC Chair sees crypto being used as collateral, unlocking trillions in value.
    • Michael Saylor of MicroStrategy is reportedly advising all major banks.

Takeaways

  • The outlook is extremely bullish long-term but cautious in the immediate short-term.
  • Despite the long-term potential, investors should be aware of the current downtrend. The $92k-$93k level is the critical resistance to watch.
  • The combination of low retail sentiment and increasing institutional adoption is seen as a powerful setup for a future price increase.

Ethereum (ETH)

  • The price of Ethereum recently had a false breakout and is now re-testing lower support levels.
  • The critical level to watch is $3,000. The price needs to hold above this level to maintain a bullish structure.
  • The speaker believes ETH's price action is highly dependent on Bitcoin. A significant move higher in ETH is unlikely without BTC first breaking out of its own downtrend.

Takeaways

  • The sentiment is neutral to cautiously bullish, contingent on the broader crypto market.
  • Investors should monitor the $3,000 support level. The primary driver for ETH in the near future is expected to be the price action of Bitcoin.
Ask about this postAnswers are grounded in this post's content.
Video Description
In this week's video, I break down the weekly movers in a busy week as small-caps hit new all-time highs alongside transports, midcaps, and microcaps. Markets are pricing in a higher probability of reflation with PMI-sensitive factors surging, copper breaking to five-year highs, and the CRB raw industrials index reaching new yearly highs. The Fed cuts rates and the Press conference was the most dovish in years while three FOMC members dissented and they announced bill purchases. Oracle's 11% weekly plunge highlighted balance-sheet risks facing hyperscalers as AI capex spending accelerates without immediate revenue visibility. The critical insight from Gavin Baker's interview with Patrick O'Shaughnessy: we're entering a 3-4 quarter "ROIC air gap" where billions in training capex generate zero revenue before inference monetization begins. The binding constraint has shifted from GPUs to electrical power, gas turbines, transformers, and grid interconnects now determine who wins. Blackwell's transition from 30kW to 130kW per rack requires liquid cooling and reinforced infrastructure, effectively stalling 2024 hardware progress while reasoning models bridge the gap. The shift from compute power to electric power infrastructure favors companies controlling bottlenecks over those waiting in line. With massive IPOs from SpaceX, OpenAI, and Anthropic likely by 2027 competing for growth capital, and software companies facing margin compression from AI-native startups, alpha is migrating toward physical infrastructure buildout. Michigan consumer confidence sits near 45-year lows—the opposite of bubble conditions—while PMI indicators, earnings revisions across 12 of 16 global sectors, and the beta-over-quality factor all signal cyclical acceleration ahead. Timestamps: • (00:00–02:22) Reflation signals: S&P down 60bps but small caps up third straight week, IWM/midcap/microcap new all-time highs, Aussie 10-year rates breaking higher, CRB raw industrials surging, copper at 5-year highs • (02:22–04:19) Factor rotations: Size rolling over (PMI bullish), growth-over-value compressing, earnings revisions accelerating in 12/16 global sectors, transports leading despite weak current data • (04:39–06:27) Disinflation evidence: Gas at pump down 16 cents since November, owner's equivalent rent back to 2019 levels, ECI declining, quits rate falling, 2-year inflation swaps at yearly lows • (06:52–09:35) Fed Chair Powell's key points: Labor market weaker than headline data, payrolls overstated 60k/month, AI layoffs now visible, AI primary GDP driver, productivity shift structural not temporary, service inflation cooling • (10:56–15:26) Bubble debate: Russell Napier/Howard Marks perspectives, key difference is demand remains ahead of supply, infrastructure lags digital demand, Michigan consumer confidence near 45-year lows (opposite of dot-com peak) • (15:44–17:20) Oracle warning: Down 32% from September highs, CDS spreads widening, balance sheet concerns as hyperscalers face revenue timing risk from heavy capex • (17:20–22:53) Gavin Baker insights: Must use top-tier paid models to judge progress, Blackwell transition most complex in history (30kW to 130kW racks), reasoning saved scaling progress, 3-4 quarter ROIC air gap before inference revenues arrive, SaaS companies face structural headwinds like post-fracking energy, Meta couldn't run 100k GPU cluster coherently • (26:36–27:55) ROIC air gap visualization: Training capex (Blackwell buildout) precedes revenue by quarters, creating balance sheet pressure for hyperscalers while opening opportunities in power/grid infrastructure • (33:04–35:46) Recursive self-improvement: Eric Schmidt/San Francisco consensus places this 2-4 years out, will accelerate AI beyond current pace, Pentagon forming AGI steering committee, military has vested interest in continued progress • (36:08–37:54) Tesla robotaxi: Elon Musk announced unsupervised full self-driving in Austin "in about 3 weeks" with no one in vehicle, marks opening act of embodied AI race, Colossus 2 first gigawatt data center enabling rapid progress • (38:11–40:18) Hardware cycle implications: 17-year software stagnation (like Corning) ending, VLM/VLA transition demands physical infrastructure, lidar for non-vision autonomous systems, gas turbines/transformers/solar all underinvested • (40:38–42:36) IPO wave ahead: SpaceX, OpenAI, Anthropic, xAI need capital for hardware buildout, creates growth-multiple compression risk for incumbent software, IWM-over-QQQ "Black Widow" trade works when PMIs rise • (42:57–47:06) Bitcoin/crypto: Most asymmetric career bet, still in technical downtrend until 92-93k breaks, Ethereum needs 3k hold, beta-over-profitability correlation, JP Morgan arranging Galaxy bond on Solana, banks seeking Saylor Bitcoin advice, tokenization enabling illiquid asset liquidity (private credit/CRE)
About Jordi Visser
Jordi Visser

Jordi Visser

By @jordivisserlabs

Empowering seasoned professionals to navigate the future of finance, technology, and AI. What We Offer: - Cutting-edge ...