Theses
The questions the AI economy turns on.
Each thesis is one question about the AI economy, argued on both sides by named voices, with dated tests that settle it in public. They are grouped by where the money moves along the AI value chain.
14 theses
Semiconductors & Compute Silicon
Who designs, makes and equips the logic chips that do AI compute.
Does the concentration of leading-edge chipmaking give TSMC pricing power over its customers?
TSMC's revenue for September 2026 was NT$511.9 billion, 54.6% above September 2025. The tests read its monthly revenue as the Taiwan Stock Exchange publishes it.
Next test: Nov 10, 2026 · 14 tests Call · 72% yes
Will CoreWeave report year-over-year revenue growth below 30% in any quarter reported on or before 2027-12-31?
As the workload mix shifts from training to latency-sensitive, cost-conscious inference served on hyperscaler custom silicon near users, CoreWeave's merchant GPU-rental revenue growth is tested against a 30% year-over-year floor through 2027.
Next test: Nov 3, 2026 · 6 tests
Will a panel-level or substrate-less advanced packaging line (CoPoS, CoWoP, or OSAT 310x310mm panel) reach announced mass production for an AI accelerator product by the end of 2027?
The CoWoS shortfall forces panelization and substrate-less relief paths into real mass production by end-2027, moving part of the AI package bottleneck out of TSMC's wafer-based lines.
Next test: Nov 12, 2026 · 6 tests Call · 60% yes
Will HBM supply, rather than advanced packaging capacity, be named by Nvidia or TSMC as the primary accelerator supply constraint in any 2027 earnings call?
The binding constraint migrates from CoWoS interposer capacity to HBM stacking and DRAM supply, so the named bottleneck in 2027 is memory rather than packaging.
Next test: Nov 3, 2026 · 6 tests Call · 72% yes
Will Nvidia's reported GAAP gross margin fall below 70% in any fiscal quarter reported between 2026-11-15 and 2027-12-31?
Foundry and advanced-packaging pricing power may pass cost increases to chip designers; the question is whether that pressure is visible in the largest designer's reported GAAP gross margin falling below 70%.
Next test: Nov 10, 2026 · 6 tests
Will a mainland Chinese foundry other than SMIC publicly confirm volume production at a 7nm-class (N+2) or better logic node before 2027-12-31?
Process learning diffuses from SMIC to Hua Hong/HLMC and other domestic fabs, widening the choke point so that entity-list controls aimed at single firms stop determining outcomes.
Next test: Oct 28, 2026 · 7 tests
Will US-headquartered frontier model providers (OpenAI, Anthropic, Google, Meta) still be unavailable for direct commercial service to mainland China customers as of 2027-09-30?
Regulatory and sanctions barriers keep Western frontier models out of China while Chinese models (Doubao, Qwen, DeepSeek, Kimi) hold the domestic market, locking in two separate AI product markets.
Next test: Nov 10, 2026 · 6 tests
Will large software firms' gross margins drop a point or more by Q3 2027 while Nvidia's stays at or above 70%?
The idea: AI inference costs are climbing faster than software revenue, squeezing software profitability while chipmakers keep their fat margins, moving profit upstream to silicon.
Next test: Nov 18, 2026 · 6 tests Call · 84% yes
Will the PHLX Semiconductor Index (SOX) total return exceed the iShares Expanded Tech-Software ETF (IGV) total return over the period 2026-01-01 to 2027-09-30?
Bottleneck rent at the silicon layer keeps semiconductor equity returns ahead of software equity returns through the AI capital cycle.
Next test: Nov 12, 2026 · 6 tests Call · 72% yes
Will Amazon disclose by 2027-12-31 that its own AI chips lead a named AWS AI workload, or that AWS operating margin exceeds 38%?
AWS's Trainium/Inferentia price-performance advantage drives enough inference migration off merchant GPUs to show up as a disclosed workload share or an expanded AWS operating margin.
Next test: Nov 18, 2026 · 7 tests
Will Nvidia disclose in an SEC filing or earnings release, on or before 2027-06-30, that it recognized revenue from H200-class (or successor China-approved) data center GPU sales to customers in China?
Chinese compute demand outruns Huawei/SMIC supply, so Beijing permits and Nvidia books material H200-class GPU revenue in China within the test window.
Next test: Nov 10, 2026 · 6 tests
5 theses
Memory & Storage
The memory and storage that feed the accelerators and hold the data.
Will memory makers keep earning scarcity profits from AI demand through 2027?
Micron's revenue for the quarter ended September 3, 2026 was $54.2 billion, against $11.3 billion a year earlier. The tests read Micron's next quarter and Korea's chip exports, most of which are memory.
Next test: Oct 29, 2026 · 17 tests Call · 68% yes
Will Korea's monthly memory semiconductor exports set a new all-time monthly record in any month between 2026-11-15 and 2027-12-31?
Inference workloads are read by some as memory-bound rather than compute-bound, pulling HBM, DRAM and CXL demand toward record levels, while this side tests whether Korean export data actually registers a new all-time monthly high.
Next test: Oct 27, 2026 · 8 tests Call · 76% yes
Will DRAM contract or spot prices fall at least 30% from their 2026 peak at any point before December 31, 2027, per a dated public price series?
Memory scarcity pricing reverses into a glut, with DRAM prices falling 30%+ from peak as greenfield capacity and AI efficiency gains close the gap.
Next test: Oct 29, 2026 · 7 tests Call · 80% yes
Will DRAM contract prices (TrendForce DXI or published DDR5 contract price index) decline by 20% or more from their 2026 peak at any point before 2027-12-31?
Capacity additions, Chinese entrants, and architecture workarounds break the memory shortage, producing a 20%+ drawdown in DRAM contract pricing from the 2026 peak.
Next test: Oct 29, 2026 · 6 tests
Will Nvidia's data center revenue growth rate for any quarter reported between 2026-11-15 and 2027-12-31 be explicitly attributed in its earnings call to memory (HBM) supply constraints?
Memory, not GPU die or foundry capacity, is the binding constraint, and Nvidia management will publicly cite HBM/memory supply as a limiter on shipments or product configuration.
Next test: Oct 29, 2026 · 6 tests
6 theses
Networking, Interconnect & Optics
What moves data between chips, servers, racks and data centers.
Will the U.S. have in force, as of 2027-06-30, a tariff or import restriction specifically targeting Chinese-origin optical transceivers?
A proposed U.S. restriction on Chinese-origin optical transceivers is weighed against China's leverage over upstream optical materials, with one side expecting the restriction to be paused or shelved and the other expecting it to take force by mid-2027.
Next test: Nov 2, 2026 · 7 tests Call · 69% yes
Will Taiwan-listed and US-listed optical transceiver suppliers report 1.6T module volume shipments as their largest datacom revenue contributor in any quarter reported by 2027-12-31?
Bandwidth per socket doubles with each accelerator generation, forcing an 800G-to-1.6T transition that raises networking dollars per GPU on schedule.
Next test: Nov 3, 2026 · 7 tests Call · 72% yes
Will Coherent Corp report datacom transceiver-driven Networking (Communications) segment revenue growth of at least 40% year-over-year in a fiscal quarter reported between 2026-11-15 and 2027-09-30?
The 800G-to-1.6T transition sustains a multi-year optical upgrade cycle, showing up as sustained 40%+ year-over-year growth in the leading Western transceiver supplier's communications revenue rather than a one-year spike.
Next test: Nov 3, 2026 · 6 tests Call · 67% yes
Will a major optical transceiver supplier attribute a margin or revenue decline to transceiver pricing pressure between 2026-11-15 and 2027-12-31?
Transceiver capacity additions outrun deployable AI cluster capacity, producing visible price erosion and a public margin or guidance cut by at least one major supplier.
Next test: Nov 3, 2026 · 6 tests
Will Ciena report fiscal-year 2027 revenue growth of at least 20% year-over-year, with cloud/hyperscaler customers cited as the primary driver?
Power-constrained, multi-site AI training forces hyperscalers to buy long-haul coherent scale-across capacity at a pace that lifts Ciena's growth far above its historic mid-single-digit trend.
Next test: Nov 3, 2026 · 7 tests
Will Arista Networks' GAAP gross margin fall below 60% in any quarter reported before 2027-12-31?
CPO deletes the board-level signal-integrity engineering that justified branded switch premiums, pushing hyperscalers to whitebox/ODM sourcing and compressing incumbent network vendor margins.
Next test: Nov 3, 2026 · 7 tests
17 theses
Hyperscalers, Cloud & Compute Services
Who rents compute and runs models for others, and spends the capex.
Will capital spending at Microsoft, Alphabet, Amazon and Meta outrun the cash their businesses generate?
Their purchases of property and equipment came to 96.1% of operating cash in the quarter ended June 30, 2026, against 68.7% a year earlier. The tests read the next two quarters in the companies' own filings, watching whether that ratio stays above or falls below 85%.
Next test: Nov 9, 2026 · 8 tests Call · 62% yes
Will NVIDIA state by Nov 30, 2027 that inference is over 50% of its data center compute demand?
As AI shifts from building models to running them in everyday products and agents, inference calls could outgrow training runs. The question is whether NVIDIA says so on the record.
Next test: Nov 9, 2026 · 6 tests Call · 72% yes
Will spot/on-demand hourly rental prices for NVIDIA H100 80GB SXM capacity on major public GPU marketplaces be at or above their November 2026 level as of the stated test dates in 2027?
Deployed accelerator capacity stays scarce enough that even prior-generation H100/A100 rental prices hold or rise rather than collapsing as Blackwell-class supply floods in, confirming there is no idle-GPU glut.
Next test: Nov 9, 2026 · 6 tests Call · 72% yes
Will at least two of Microsoft, Alphabet, Amazon and Meta state on a 2027 earnings call that they remain capacity- or power-constrained in AI compute?
Through 2027 the largest cloud operators continue to describe themselves as short of AI capacity rather than long, confirming installed accelerators stay fully utilized, though whether any will use that precise phrasing on an earnings call is itself uncertain.
Next test: Oct 28, 2026 · 8 tests Call · 82% yes
Will three or more distinct operators publicly offer surplus internal AI compute for external resale (or disclose underutilized AI capacity) between 2026-11-15 and 2027-12-31?
The no-glut claim breaks down: firms that over-ordered ahead of site readiness or lost their internal AI programs begin marketing surplus accelerators externally, revealing idle capacity.
Next test: Nov 9, 2026 · 6 tests Call · 72% yes
Will aggregate 2027 capex guidance from Alphabet, Amazon, Microsoft and Meta imply year-over-year growth of less than 15%?
Cash constraints force capex growth to decelerate toward the pace of operating cash flow, with 2027 guidance implying under 15% aggregate growth.
Next test: Nov 9, 2026 · 6 tests
Will cloud revenue growth at Microsoft, Google, and AWS outpace their combined depreciation growth near Q3 2027?
Record spending on AI data centers is pushing up depreciation costs. The question is whether cloud revenue is growing even faster.
Next test: Nov 9, 2026 · 6 tests Call · 72% yes
Will Alphabet, Microsoft and Amazon each post positive free cash flow for calendar 2027?
Cloud and AI spending is surging, but so is capex. This tracks whether all three hyperscalers still generate more cash than they spend over the trailing year through 2027.
Next test: Nov 9, 2026 · 6 tests
Will Microsoft, Amazon, or Alphabet reveal that one customer makes up 10% or more of their revenue or future contracted business?
A handful of richly funded AI labs may already loom so large in cloud sales that one of the big three cloud providers has to name them in official filings before the end of 2027.
Next test: Nov 9, 2026 · 6 tests
Will D&A as a share of revenue rise year-over-year for at least 3 of 4 big hyperscalers in Q3 2027 vs Q3 2026?
Years of heavy AI infrastructure spending are moving from construction into active use, and depreciation is catching up with that spending curve.
Next test: Nov 9, 2026 · 6 tests
Will any of Microsoft, Alphabet, Amazon, or Meta disclose a further extension of the estimated useful life of servers or network equipment in a 10-K or 10-Q filed between 2026-11-15 and 2027-12-31?
At least one major hyperscaler again extends server useful lives, offsetting depreciation-driven margin compression through accounting choice rather than economics.
Next test: Oct 28, 2026 · 6 tests Call · 66% yes
Will a listed AI neocloud report a GPU impairment or depreciation charge of $100M+ by end of 2027?
GPU leasing deals assume chips hold value over their contract term, but faster hardware cycles could force write-downs. This page tracks whether a major US-listed neocloud discloses a large impairment or accelerated depreciation tied to aging GPUs.
Next test: Nov 9, 2026 · 6 tests Call · 78% yes
Will the average trailing-twelve-month ROIC of Microsoft, Alphabet, Amazon and Meta fall by the quarter nearest Q3 2027 versus the quarter nearest Q4 2024?
Massive AI capital spending is piling into balance sheets and depreciation schedules faster than AI revenue shows up, which could drag down returns on invested capital across the biggest cloud players.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will Oracle report, in any fiscal quarter reported between 2026-11-15 and 2027-12-31, a year-over-year decline in total gross margin percentage of 300 basis points or more alongside a capex guidance increase?
Oracle, the leveraged builder of AI capacity against concentrated counterparties, will show the telecom-builder signature: gross margin compressing through the utilization floor while capex guidance rises and RPO converts slowly.
Next test: Nov 9, 2026 · 6 tests Call · 70% yes
Will combined disclosed AI-attributed annualized revenue across the major US AI providers reach $300bn by the end of 2027 as reported in public filings or earnings disclosures?
AI revenue scales fast enough that the revenue-to-depreciation ratio on AI infrastructure rises toward the level where returns clear an 8-10% cost of capital, contradicting the value-destruction case that argues capex outruns the revenue needed to earn it back.
Next test: Oct 29, 2026 · 8 tests Call · 55% yes
Will combined 2027 capex from Microsoft, Alphabet, Amazon, and Meta grow less than 10% year over year?
The AI buildout may be outrunning real demand. This tracks whether the four biggest cloud spenders sharply slow their capital spending growth in 2027, signaling a digestion phase.
Next test: Nov 9, 2026 · 6 tests
Will Nvidia report a year-over-year decline in data center revenue in any fiscal quarter reported before December 31, 2027?
The capex wave pauses and Nvidia's data center revenue declines year over year at least once by end-2027 as the overshoot works through.
Next test: Nov 9, 2026 · 6 tests
6 theses
Data Centers & Physical Infrastructure
The buildings, land, racks, cooling and in-building electrical gear.
Will at least one of Galaxy Digital's Helios 830 MW approved expansion or Hut 8's Beacon Point capacity be announced as contracted under a long-term lease with a named tenant by 2027-06-30?
Approved, interconnected but unleased megawatts at energized-land owners convert into signed long-term leases, confirming that power position translates into contracted rent.
Next test: Oct 28, 2026 · 6 tests Call · 78% yes
Will LandBridge or Texas Pacific Land report surface/data-center-related revenue (surface use, easements, pore space, water) up at least 30% year over year in FY2026 results?
Owners of Permian surface acreage monetize data-center and power-generation demand into measurably rising fee-based surface revenue.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will a US hyperscaler or AI cloud provider disclose a $500M or more AI/GPU-related impairment charge in a period ending by 2027-09-30?
Older accelerator generations lose economic value faster than 5-6 year depreciation schedules assume, forcing at least one large operator to book a material write-down or shorten asset lives on AI hardware.
Next test: Nov 3, 2026 · 6 tests
Will CoreWeave's FY2026 Form 10-K retain a weighted-average or stated useful life of six years or more for its GPU computing equipment, with no AI-hardware impairment charge above $250M disclosed for FY2026?
Prior-generation accelerators keep recontracting at economic prices, so neocloud depreciation schedules hold and no material stranding charge appears.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will a major AI data center operator publicly disclose, by 2027-12-31, that installed AI compute capacity of 100 MW or more is idle or revenue-deferred because of power interconnection or energization delays?
Grid interconnection timelines run far longer than hardware deployment cycles, leaving purchased accelerators dark and aging before they can earn revenue.
Next test: Nov 3, 2026 · 6 tests Call · 68% yes
Will EIA report commercial or data-center-linked electricity sales growth above 5% year-over-year for calendar 2026, by late 2027?
Electricity and interconnection are the rivalrous complement whose scarcity converts AI demand into sustained rent for power owners rather than for model or app providers.
Next test: Oct 28, 2026 · 6 tests
9 theses
Power, Energy & Grid
The electricity that reaches the fence, who generates it, who moves it and who pays.
Will US electricity generation from natural gas in calendar year 2026 exceed the 2024 level by more than the combined increase from wind, solar, and nuclear over the same period, per EIA Electric Power Monthly?
Gas-fired generation supplies the largest share of incremental US electricity demand through 2026 because turbines can be sited and built faster than any alternative at AI-datacenter timelines.
Next test: Nov 9, 2026 · 15 tests
Will the Henry Hub front-month natural gas settlement price close above $7.00/MMBtu on any trading day before 2027-12-31?
Combined AI datacenter gas burn, LNG export ramp, and slowing shale supply growth push Henry Hub above $7 at least once, showing the gas bridge is supply-constrained rather than cheap.
Next test: Nov 6, 2026 · 22 tests Call · 71% yes
Will at least three major tech firms announce a new natural gas generation arrangement of 300 MW or more for US data centers between 2026-11-15 and 2027-12-31?
Hyperscalers themselves turn to dedicated gas generation at scale for near-term AI capacity, confirming gas as the practical bridge fuel despite stated clean-energy targets.
Next test: Oct 28, 2026 · 6 tests Call · 62% yes
Will a major US hyperscaler attribute an AI capacity delay to power or grid interconnection between 2026-11-15 and 2027-12-31?
Grid interconnection and electrical-chain timing, not capital or chip availability, is the binding constraint, and hyperscalers will say so on the record as power-on dates slip.
Next test: Nov 6, 2026 · 6 tests Call · 64% yes
Will NVIDIA report data-center revenue growth of less than 15% year-over-year in any fiscal quarter reported between 2026-11-15 and 2027-12-31 while simultaneously citing customer power/grid availability as a factor?
Power and interconnection scarcity may bind before chip supply does, potentially slowing GPU absorption in ways that show up as decelerating NVIDIA data-center revenue, if the company attributes any slowdown to siting and energization.
Next test: Nov 6, 2026 · 6 tests
Will US annual electricity sales to the commercial sector plus reported data center load growth, per EIA, show data-center-driven demand growth outpacing net generating capacity additions in 2027?
US net firm capacity additions will fall short of incremental data center load growth in 2027, confirming the grid as the system-level bottleneck on AI buildout.
Next test: Nov 6, 2026 · 6 tests Call · 64% yes
Will the EIA-reported US average residential electricity price for calendar year 2027 exceed 18.7 cents per kWh?
Data center load growth outpaces generation and transmission additions, pushing the national average residential electricity price above 18.7 c/kWh in 2027.
Next test: Nov 6, 2026 · 6 tests Call · 65% yes
Will at least three US state public utility commissions approve new large-load or data center tariffs with minimum-take or exit-fee provisions between 2026-11-15 and 2027-12-31?
Visible household bill increases attributed to data centers force state regulators to approve special large-load tariffs that shift costs away from residential ratepayers.
Next test: Nov 9, 2026 · 6 tests
Will the five US states with the fastest 2026-2027 electricity sales growth show lower average residential price growth than the five slowest-growth states?
Fixed infrastructure costs spread over higher data center volumes so high-load-growth states see slower residential price inflation than low-growth states, contradicting the ratepayer-burden claim.
Next test: Nov 6, 2026 · 6 tests Call · 69% yes
7 theses
Foundation Models & AI Intelligence
Who trains and serves the models, and the economics of doing it.
Will Microsoft report a company-wide gross margin percentage for the quarter ending June 30, 2027 that is at least 100 basis points above the same quarter of fiscal 2025, as disclosed in its SEC filings?
Falling cost per served token outpaces token price declines, so hyperscaler-reported gross margins expand despite surging AI inference volume and depreciation from the capex buildout.
Next test: Oct 28, 2026 · 6 tests Call · 78% yes
Will the median trailing gross margin of US-listed AI-native application software companies reporting AI inference as cost of revenue be at least 70% in their fiscal 2027 filings?
Cost-per-token declines and cohort maturation lift AI application gross margins back toward classic software levels (70%+) rather than leaving them structurally compressed by variable inference costs.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will US data center electricity consumption, as reported by EIA, grow at least 15% year-over-year in the most recent full-year or twelve-month figure published before 2027-12-31?
Elastic inference demand translates into physical compute expansion, so data center electricity consumption keeps growing double-digit percentages despite per-token cost declines.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will an open-weight model match or be within 2 points of the top closed model on a major coding benchmark before 2027-12-31?
Open-weight releases from Chinese and Western labs (DeepSeek, Qwen, GLM, Llama-class) will reach within ~2 points of the best closed frontier model on a headline public coding benchmark, demonstrating effective capability parity.
Next test: Nov 2, 2026 · 6 tests Call · 76% yes
Will the published list price per million output tokens of the flagship frontier models from OpenAI and Anthropic fall by 50% or more from their 2026-01-01 levels by 2027-12-31?
Open-weight substitutes hosted cheaply by third-party providers will force closed labs to cut flagship API list prices by at least half.
Next test: Nov 18, 2026 · 6 tests Call · 78% yes
Will the per-million-output-token list price of the flagship frontier models from OpenAI and Anthropic fall by at least 40% from their 2025 levels by the end of 2027?
Competition from near-frontier open-weight models (especially Chinese labs) and zero-switching-cost routing forces US frontier labs into repeated list-price cuts, so flagship output-token prices fall at least 40% versus 2025 levels.
Next test: Nov 18, 2026 · 6 tests
Will a major foundation model provider publicly disclose model/inference gross margin above 60% in a filing or audited disclosure before 2027-12-31?
Despite price competition, some argue frontier providers retain pricing power above serving cost; others expect commoditization, and the question is whether a provider discloses inference gross margin above 60% before 2027-12-31.
Next test: Nov 2, 2026 · 7 tests Call · 74% yes
4 theses
AI Software, Applications & Agents
Who turns model output into software people and firms pay for.
Will at least two of the five largest seat-priced software vendors report subscription revenue growth below 8% for two straight quarters before 2027-12-31?
AI agents compress seat counts and per-seat price enough that headline subscription growth at major seat-priced vendors decelerates below 8% for two straight quarters.
Next test: Oct 28, 2026 · 6 tests Call · 72% yes
Will at least three of five major software vendors launch a consumption or outcome-based agent pricing model as a primary option by 2027-12-31?
Seat-priced vendors defensively convert to consumption/outcome pricing for agent products, confirming that seat economics no longer support growth.
Next test: Nov 2, 2026 · 6 tests Call · 72% yes
Will at least one major per-employee-per-month or per-seat SaaS vendor report a year-over-year decline in total paid seats or customer headcount-based billings in a public filing or earnings release before 2027-12-31?
Agentic automation shrinks the customer seat base enough that a major per-seat vendor publicly reports declining seats or PEPM-linked billings, validating the seat-loss mechanism behind outcome pricing.
Next test: Oct 28, 2026 · 6 tests Call · 74% yes
Will at least three of five major seat-priced SaaS vendors report net revenue retention below 100% in a quarter before 2027-09-30?
AI-driven seat compression and build/substitution pressure push net dollar retention at major seat-priced SaaS incumbents below 100%, the clearest public evidence that pricing power inside the installed base has broken.
Next test: Oct 28, 2026 · 6 tests Call · 68% yes
3 theses
AI Security, Trust & Governance
The spend on protecting, identifying and governing AI systems and their users.
Will CrowdStrike report ending ARR of at least $5.5B with AI-native modules cited as a growth driver by 2027-06-30?
Enterprise AI deployment multiplies endpoints, identities and agents that require sensors, so CrowdStrike's ARR and per-customer module adoption accelerate rather than commoditize.
Next test: Oct 28, 2026 · 7 tests
Will at least two US-listed security or observability vendors name AI-agent governance as a specific revenue driver before 2027-09-30?
Rogue-agent risk converts AI governance from a compliance line item into funded demand for observability and counter-AI tooling, visible in vendor disclosures.
Next test: Nov 5, 2026 · 6 tests Call · 76% yes
8 theses
AI Adoption, Productivity & Industry Transformation
Where AI changes the costs and revenues of the rest of the economy.
Will BLS CPS data show the employment-population ratio for workers aged 20-24 fall by at least 1.5 percentage points from its 2025 average to the trailing-12-month average ending in a month on or before November 2027?
AI adoption removes entry-level tasks faster than new entry roles appear, so young-worker employment deteriorates measurably more than the overall labor market.
Next test: Dec 17, 2026 · 6 tests Call · 70% yes
Will Challenger, Gray & Christmas report cumulative US job cuts explicitly attributed to 'artificial intelligence' of at least 100,000 for calendar year 2027?
Employer-stated AI displacement scales into a six-figure annual job-cut category, confirming AI as a named rather than merely narrated cause of white-collar layoffs.
Next test: Dec 17, 2026 · 6 tests
Will the S&P 500 ex-financials aggregate trailing-twelve-month operating margin reported for Q4 2026 exceed the Q4 2025 level by at least 50 basis points?
Broad AI adoption in administrative, support, engineering and coordination work lowers opex per unit of revenue fast enough that aggregate index-level operating margins expand by at least half a point year over year, rather than being reinvested into headcount.
Next test: Oct 29, 2026 · 6 tests Call · 58% yes
Will at least three of MercadoLibre, Paycom, Salesforce and Lemonade report a fiscal 2026 operating margin at least 200 basis points above their fiscal 2025 level?
Companies that publicly attribute cost savings to internal AI deployment convert those savings into visible, audited operating margin expansion of at least two points within one fiscal year.
Next test: Oct 29, 2026 · 6 tests Call · 62% yes
Will US nonfarm business labor productivity growth, measured year-over-year, be at or above 2.5% in any quarter reported by BLS between 2026-11-15 and 2027-12-31?
AI adoption is already lifting measured output per hour, so nonfarm business productivity growth prints at or above 2.5% year-over-year in at least one quarter in the test window.
Next test: Oct 29, 2026 · 6 tests Call · 71% yes
Will the BLS nonfarm business labor share index fall below its 2025 average in the last reading published before 2027-12-31?
Whatever productivity AI delivers accrues to corporate margins rather than labor, so the measured labor share keeps declining even as output per hour rises.
Next test: Oct 29, 2026 · 8 tests Call · 58% yes
Will at least two of Accenture, Cognizant, Infosys, Globant and EPAM report negative year-over-year constant-currency revenue growth in any single reported quarter ending on or before 2027-09-30?
AI-assisted delivery and renewal repricing deflate the billable unit fast enough that major effort-billed IT services firms post outright organic revenue declines, not just deceleration.
Next test: Oct 28, 2026 · 8 tests
Will the largest listed customer-experience BPO operators (Teleperformance, Concentrix, TaskUs) collectively report declining organic revenue for a fiscal year covering 2026 or 2027?
Conversational and voice AI automates enough routine interactions that outsourced customer-service providers lose addressable volume and report organic revenue declines.
Next test: Oct 28, 2026 · 4 tests Call · 78% yes
11 theses
AI Financing, Credit & Capital Structure
Who lends to, invests in and insures the buildout, and on what terms.
Will Alphabet, Amazon, Microsoft, Meta and Oracle collectively issue more than $150 billion of new corporate bonds during calendar year 2026?
Because capex exceeds internal cash generation, the major AI buildout funders tap debt markets at record scale, issuing over $150 billion of bonds in 2026.
Next test: Nov 9, 2026 · 6 tests
Will the forward P/E multiple of the S&P 500 information technology sector contract by at least 15% from its 2026 peak level by 2027-12-31, while sector forward EPS estimates remain flat or higher?
AI-exposed equity valuations are priced for perfection, so multiples compress materially even as earnings estimates hold, producing a derating rather than an earnings collapse.
Next test: Nov 11, 2026 · 12 tests Call · 72% yes
Will the Philadelphia Semiconductor Index (SOX) experience a peak-to-trough drawdown of 25% or more at any point between 2026-11-15 and 2027-12-31?
Concentration, leverage and narrative dependence make the AI complex fragile enough that a growth-rate deceleration or a single disclosure shock produces a 25%+ semiconductor drawdown within the window.
Next test: Nov 23, 2026 · 6 tests Call · 76% yes
Will bond spreads for at least two major AI infrastructure borrowers widen 150+ basis points from Q4 2026 levels before 2027 ends?
AI buildouts increasingly lean on debt and vendor financing tied to demand from labs that are themselves not yet profitable. A funding stumble could hit credit markets before stocks or revenue show cracks.
Next test: Nov 23, 2026 · 6 tests
Will at least 5 US-listed IPOs of AI-linked companies each raise $500 million or more between Jan 2026 and Sept 2027?
Public markets could reopen as a major funding source for the AI buildout, producing a wave of large listings instead of companies staying private.
Next test: Nov 23, 2026 · 6 tests Call · 66% yes
Will a publicly reported payment default, covenant breach, restructuring, or rating downgrade to speculative-grade occur on a GPU-collateralized loan or securitization of at least $500 million before 2027-12-31?
The asset-life versus debt-tenor mismatch produces at least one public distress event on a large GPU-backed facility within the window.
Next test: Nov 24, 2026 · 6 tests Call · 55% yes
Will the median on-demand rental price of an Nvidia H100 80GB GPU remain at or above $1.50 per GPU-hour on a major public price index as of 2027-06-30?
Will H100 rental rates hold at or above $1.50/GPU-hour, or does collateral value decay faster than lenders assumed? Yes side argues recontracting and residual value support a price floor; no side argues depreciation and oversupply break it.
Next test: Nov 24, 2026 · 6 tests Call · 62% yes
Will Liberty Energy (LBRT) / Liberty Power Innovations disclose at least $1.5bn of closed non-recourse SPV project debt backed by fully executed energy services agreements by year-end 2027?
Will Liberty disclose at least $1.5bn of closed non-recourse SPV project debt backed by executed hyperscaler ESAs by year-end 2027? Yes side argues the structure is bankable and will close; no side argues it is a vehicle for hiding parent leverage and risk.
Next test: Nov 23, 2026 · 6 tests
Will a US-listed GPU cloud operator record a material GPU impairment or extend GPU useful life between 2026-11-15 and 2027-12-31?
At least one major listed neocloud will record a material impairment or accelerated depreciation charge on prior-generation GPU fleets, confirming that collateral value decays faster than the debt financing it.
Next test: Nov 18, 2026 · 6 tests Call · 62% yes
Will Microsoft, Alphabet, Amazon, Meta and Oracle's combined disclosed lease obligations exceed $800bn in filings for fiscal periods ending in 2026?
Hyperscalers are shifting data center capacity off their own balance sheets into leases and third-party SPVs, so disclosed lease commitments (and the not-yet-commenced line, the clearest marker of hidden buildout leverage) will grow faster than reported capex and cross $800bn.
Next test: Nov 23, 2026 · 6 tests Call · 66% yes
Will Nvidia consolidate a neocloud or data center SPV, or record at least $2bn in cumulative related impairments and credit losses, by fiscal year ending January 2028?
The risks and rewards of neocloud buildouts effectively sit with Nvidia, so accounting consolidation or a multi-billion-dollar loss recognition will reveal the retained exposure.
Next test: Nov 5, 2026 · 6 tests Call · 64% yes
No thesis page yet in Materials, Industrial Supply Chain & Construction, Physical AI, Robotics & Autonomy.
A thesis page states one question and both sides of it; DeepStack takes no side in words, and its view is the number it gives each test before the date. 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. The Call’s number is the chance, from 0 to 100, that the Call gives the yes answer to a test’s question, recorded before the test’s date; it is DeepStack’s own reading, not a market price and not advice. The readers’ mark on the same line is the readers’ median, the middle of the numbers signed-in readers gave the test, shown from 25 readers.