Who is on each side · Part 10

Does the hardware hold its value?

12 investors who disagree about The Bubble Is in the Denominator. Each voice is attributed and paraphrased, dated where the date is known, with a link where one exists.

The DeepStack callSkeptical

The market is not cheap. It is capitalizing AI earnings before anyone knows who owns them.

Cheap is conditional · Conviction High (4/5)

Read the full call

6 voices

A model, not a price

  1. Michael Burry

    Acting on the expected AI-boom-to-bust thesis, the creator bought more MetLife (MET) 2029 puts far out of the money.

    Source

  2. Michael Burry

    The residual-value numbers on Nvidia's slide are not actual second-hand prices but the present value of current rental income run out over 8 years according to a continuous discounted cash flow model built on assumptions like 75% utilization and a…

    Source

  3. Ren

    Semiconductors trade at 17.1x forward earnings while the S&P 500 sits at 19.0x, so the sector powering the largest capex cycle in history is priced at a discount to the average American company, versus the same group trading at 35x two years ago.

    Source

  4. Ren

    Almost the entire move of this cycle has come from earnings growth rather than multiple expansion, unlike the run-up to March 2000 when the forward multiple on tech expanded roughly 250%, versus about 14% expansion since October 2022; a rally paid…

    Source

  5. Kakashii

    The creator calculates that Nvidia's plan to ship over 10 million chips of Blackwell and Rubin by the end of 2026 implies a total required power of roughly 17 to 23 GW, averaging around 20 GW, and questions whether enough data centers are under…

    Source

  6. Kakashii

    The creator's long-standing thesis, which they say has gained support and validation over time, is that Blackwell GPUs had to be shipped to warehouses because there were not enough compatible datacenters (with the needed cooling and equipment) to…

    Source

6 voices

Still renting at January prices

  1. David George

    Since 2023, tech has been the earnings growth story, contributing about 76% of the S&P 500's total earnings growth in 2026 as of late August.

    Source

  2. David George

    For now at least, the upward-inflecting demand curve for AI compute means older chips like A100s are still pretty useful, with rental rates and residual values not decreasing as expected and even A100 pricing at or above where it was at the start…

    Source

  3. JunkBondInvestor

    Contracted GPU lease rates and second-hand prices have held up even for chips launched 3 to 6 years ago, and compute is now about 60% of hyperscaler capex, so the scarce, short-lived asset is a growing share of what is being financed.

    Source

  4. JunkBondInvestor

    The entire AI buildout has been constrained by scarcity of power, GPUs, transformers, interconnection, skilled labor and ready-to-use capacity, and this scarcity has propped up the economics of everything financed against it by holding up lease…

    Source

  5. Tae Kim

    The creator believes that if OpenAI were public, its stock would have gapped up 30% or more on hitting $70 billion in ARR, a blowout versus expectations on the level of Nvidia in May 2023, and that the figure proves his August report that OpenAI's…

    Source

  6. Tae Kim

    Axios reported, and the Financial Times later confirmed the same day, that OpenAI's annual recurring revenue is nearing $70 billion, with enterprise sales more than doubled since July according to sources familiar with the financials.

    Source

Also in dispute · 15 voices

Who writes the checks?

8 voices

Not the customers, yet

  1. Torsten Slok

    Wall Street equity analysts work in sector silos, and adding up their forecasts produces internally inconsistent numbers.

    Source

  2. Torsten Slok

    Analysts covering tech expect the sector's operating cash flow to more than double to roughly $2.4 trillion by 2028, an increase of over $1.2 trillion, while analysts covering the other S&P 500 sectors, which are tech's customers, expect those…

    Source

  3. Michael Spencer

    The creator reads declining frontier model usage as a sign that enterprises, startups and power users are becoming more price conscious even as token prices fall, citing Ramp's index showing effective price per million tokens down 41% to $0.68 from…

    Source

  4. Michael Spencer

    A lot of the tasks agents are being asked to do won't translate into real economic value, despite requiring massive compute, which is actually how the entire AI thesis could break given the debt behind the AI infrastructure build-out.

    Source

  5. Mia Silverio

    Overall market earnings strength this year hinges on AI investment, since Goldman Sachs estimates half of earnings growth this year will come from AI investment, even though the S&P 493 (excluding the Magnificent 7) grew earnings 31.8% in Q2 2026…

    Source

  6. Mia Silverio

    There is an AI lab bubble, though not an AI bubble infecting the whole market, as shown by Nvidia's valuation having come down significantly; excessive excitement surrounds Anthropic and OpenAI because they are new and exciting even though they…

    Source

  7. Steve Eisman

    The visible, Nvidia-funded circular financing in AI (Nvidia investing in or lending to companies that then buy Nvidia chips) is out in the open and funded from Nvidia's own cash flow, and is not what disturbs the speaker; what disturbs him is the…

    Source

  8. Steve Eisman

    Meta financed a $27 billion Louisiana data center last year through an SPV called Beignet Investor LLC, owning only 20% while Blue Owl and Pimco hold the rest; Meta is building it, has promised to rent it for 20 years, and bears the cost of delays…

    Source

7 voices

Everyone, soon

  1. Aurelion Research

    Hyperscaler capex is durable rather than at risk of a sudden cut, because the largest spenders keep reaffirming the necessity of AI investment and their EBITDA and EBIT margins are still rising, giving them capacity to keep funding it.

    Source

  2. Aurelion Research

    The customer base is broadening beyond a concentrated set of hyperscalers to enterprises, AI-native companies and cloud providers, with AI-related infrastructure commitments expected to exceed $3.1 trillion, making the next phase of demand less…

    Source

  3. Tae Kim

    OpenAI near $70B a year.

    Source

  4. Kris

    AWS has had more demand than supply, with customers wanting more AWS capacity for AI than could be served despite large Nvidia orders and Amazon's own Trainium chips, and the creator considers this kind of capacity-constrained spending among the…

    Source

  5. Kris

    Broadcom and Nvidia can both be winners because ASICs that Broadcom helps hyperscalers and AI labs build are complementary to Nvidia's GPUs rather than competitors, contrary to the framing on social media.

    Source

  6. Daniel Romero

    OpenAI's combined annualized revenue with Anthropic reached $105 billion (OpenAI $40 billion, Anthropic $65 billion) in July 2026, which is nowhere near the figure needed to support the modeled 2027 buildout, and this understates the true gap since…

    Source

  7. Daniel Romero

    The semiconductor industry is supported by the AI data center buildout, which itself depends on Anthropic's and OpenAI's total addressable market since these two companies have driven almost every major AI advancement since 2022 and carry a…

    Source

Also in dispute · 10 voices

What does 5% do to the multiple?

2 voices

A growth signal near a peak

  1. Michael Howell

    The creator underestimated the pace of earnings growth in 2026: valuation multiples compressed as expected, but corporate earnings expanded far more rapidly than anticipated, fuelled by AI investment, which is why equities, notably technology…

    Source

  2. Michael Howell

    Higher bond yields reflect an economy growing faster than official policy settings can comfortably accommodate and a policy framework increasingly dependent upon maintaining abundant liquidity.

    Source

8 voices

A long end nobody caps

  1. TSCS

    The creator argues the front end cannot cap the long end: a hike is only about 4/10 of a percent of receipts and mostly lands as income with people who spend it, Treasury buybacks manage volatility rather than level, and the buyers who used to…

    Source

  2. TSCS

    The creator argues the marginal long-end buyer has changed: a Japanese lifer hedging dollars back to yen earned about 95bp less on a hedged Treasury than on a JGB on the one day with a published hedge cost, so it stops rolling rather than sells…

    Source

  3. Vitaliy Katsenelson

    Bond investors are rejecting the government-set price for Treasuries because of large, unaddressed budget deficits that may grow rather than shrink over the next 10 years, and because budget deficits imply inflation; the creator says they…

    Source

  4. Vitaliy Katsenelson

    The AI-focused stocks, which make up roughly 37 percent of the S&P 500 and drive a good chunk of daily market movements, are fueling optimism and a growing debt pile in the AI economy, while the rest of the economy struggles.

    Source

  5. Torsten Slok

    The tech silo is betting on a future in which demand for AI and tech services explodes, while the silos covering the companies that would pay for those services see a much more modest outlook, and both cannot be right at the same time.

    Source

  6. Torsten Slok

    Consensus expects hyperscalers' operating cash flow to more than triple by 2030, outpacing the sharp increase in AI-driven capex, based on FactSet and Apollo Chief Economist sources for Google, Meta, Amazon, Microsoft and Oracle.

    Source

  7. Michael Burry

    The entire extra present value shown for the oldest GPU (A100) in the model is attributable entirely to years 7 and 8 being added into the depreciation model, so the A100's current 25% ratio comes solely from those added years.

    Source

  8. Michael Burry

    Manufacturers such as Nvidia and hyperscalers such as Oracle and Google Cloud are offering residual value guarantees for chips not because they want to but because lenders require them, given that chip depreciation is a very real problem following…

    Source

Notes

Not a disagreement.

Where two voices only seem to differ, or where the thread runs back to an earlier Part.

Where they cross

Who answers whom.

The specific points where one voice meets another: the same evidence read two ways, or the same mechanism with a different sign.

What would settle it

The dated tests.

The same tests the story set, on the Docket; results land on the Results page.

  1. Late October

    Microsoft, Alphabet, Meta and Amazon report; depreciation, server lives and 29.5% growth will test the earnings denominator.

    Part 10: The Bubble Is in the Denominator Pending

  2. October

    Anthropic filing or listing reveals audited revenue, losses and concentration, testing the $400B revenue milestone.

    Part 10: The Bubble Is in the Denominator Pending

  3. At 5.5% 10-year

    A 5.5% Treasury yield tests Howell’s turning-point history against TSCS’s warning that the long end will keep rising.

    Part 10: The Bubble Is in the Denominator Pending

  4. Next used-GPU print

    Three-year-old chip rents and resale prices show whether scarcity or durable economics supports hardware values.

    Part 10: The Bubble Is in the Denominator Pending

  5. Productivity data

    A rise in total factor productivity would show the rest of the economy can write the checks; a flat reading leaves bondholders exposed.

    Part 10: The Bubble Is in the Denominator Pending

Analysis and opinion, not investment advice. Voices are paraphrased from public writing and attributed by name; DeepStack shows no score, ranking or accuracy for any person.