How the contract works
A contract on this market settles at $1 if the resolution conditions are met and at $0 if they are not, based on the rules tied to NVIDIA, SOXX, TSMC, ASML, Broadcom, Arista, Super Micro Computer, OpenAI, Anthropic and the SiliconData H100 rental index. The final determination covers the 90 days following 1 January 2027, so events in that window count even though the headline date is the end of 2026. A price of 0.30, for example, would mean buyers and sellers collectively judge the chance of a Yes outcome at roughly three in ten โ that is a hypothetical, not this market's current level. Anyone holding a position does not have to wait for settlement; it can be sold on the open market at whatever price is current at the time.
Yes9%
No91%
Probability
How the price has moved
The tracked consensus sits at 9% on total volume of $2,409,203, all of it on Polymarket, the only venue currently listing this market. With a single venue and no independent second price to compare against, there is no cross-market spread to read for disagreement, and no reported day-over-day or week-over-week swing data to point to a specific move. What the figure does show is a market pricing the compound, three-of-six resolution rule as a low-probability outcome given current stock levels for Nvidia and its chip-sector peers, the capital position of OpenAI and Anthropic, and the current trajectory of H100 rental pricing.
Analysis
Context
This market asks whether the current run-up in AI-related stocks and spending ends in a defined collapse by early 2027. The question was built around 2026, a year in which Nvidia, TSMC, Broadcom, ASML, Arista Networks and Super Micro Computer have all traded near or at record highs, driven by demand for AI training and inference hardware. OpenAI and Anthropic, the two leading AI labs, have raised capital at valuations in the hundreds of billions of dollars, and cloud providers have committed hundreds of billions more to AI data-center buildout.
The resolution criteria are unusually specific for a market of this kind. Rather than asking a vague question about sentiment, it lists six measurable events โ stock declines of 50% or 40% from all-time highs across named companies, a bankruptcy or acquisition of OpenAI or Anthropic, or GPU rental prices falling to $1 or below for five straight days on the SiliconData Silicon Index. At least three of these six must occur within 90 days of 1 January 2027 for the market to resolve Yes.
The debate over an AI bubble has run through 2025 and 2026 in business press and among investors, with comparisons frequently drawn to the dot-com bust of 2000. Skeptics point to the gap between AI capital spending and near-term revenue; defenders point to the scale of enterprise and consumer adoption already visible in usage data.
The consensus price across tracked venues sits at 9%, with all $2,409,203 of recorded volume concentrated on a single venue, Polymarket. That concentration matters: there is no second independent venue here to check the price against, so the 9% figure reflects one pool of traders rather than a cross-market consensus in the usual sense. A low single-digit-to-low-teens price on a compound, multi-trigger resolution rule is consistent with how these markets typically behave โ the more conditions that must co-occur, the lower the implied probability, because each additional required event multiplies down the combined chance.
The structure of the resolution rules explains much of the pricing. Nvidia would need to fall 50% from its all-time high, SOXX would need a 40% drop from its peak, or one of five other large chip and infrastructure names would need to halve โ and at least three of six total triggers, spanning stock prices, corporate status and GPU rental economics, would have to land within the same 90-day window after 1 January 2027. Single-company drawdowns of that size have happened before in tech, but three near-simultaneous ones tied to different mechanisms (equity prices, corporate solvency, and a spot-rental index) is a materially harder bar to clear than a single stock correction.
The OpenAI/Anthropic bankruptcy-or-acquisition trigger is worth separating out. Both firms have raised capital at valuations well above $100 billion in recent funding rounds backed by major technology companies and investment funds, which makes an abrupt insolvency or forced sale within a 90-day window a high bar absent a specific funding or legal shock not currently visible in public reporting. The H100 rental-price trigger is different in character: prices for older-generation GPUs have been drifting down for years as newer hardware from Nvidia (such as its Blackwell architecture) supersedes the H100, so this particular condition could plausibly move over time even without a broader crash โ but it alone is not sufficient, since the rule requires at least three of six conditions.
Taken together, the pricing reflects a market that sees the individual components of a bubble narrative โ stretched valuations, heavy capex, debt-financed data-center buildout โ as real talking points, but treats the specific, compound definition of "burst" used here as a high bar that has not yet been approached by current stock levels or corporate filings.
What moves the probability
Hyperscaler capital spending
Microsoft, Google, Amazon and Meta have committed capital expenditure guidance well into the hundreds of billions of dollars for AI data-center buildout through 2026 and beyond. Continued spending at that scale supports demand for Nvidia, TSMC, Broadcom and the rest of the chip supply chain, pushing against the conditions needed for this market to resolve Yes. A public pullback in guided capex from any of these firms would be the clearest early signal in the opposite direction.
Distance from all-time highs
Several of the named stocks โ Nvidia, TSMC, Broadcom, Arista, Super Micro โ have traded at or near record levels through 2026, meaning a 50% or 40% decline from those highs would require a very large drawdown, not a routine correction. The larger the cushion between current prices and the trigger threshold, the more the market treats the stock-price conditions as unlikely within the window.
OpenAI and Anthropic funding status
Recent capital raises for both labs at valuations above $100 billion, backed by large technology and investment partners, make a near-term bankruptcy or forced acquisition a high bar. This condition would most likely require a specific, disclosed funding or legal event rather than a general market slide.
H100 rental price trend
Rental prices for the H100 GPU, tracked via the SiliconData Silicon Index, have already been declining as newer chip generations enter supply, which makes this the single condition most plausibly reachable on its own. It cannot resolve the market by itself, however, since at least two other triggers are still required.
Single-venue price discovery
All $2,409,203 in recorded volume sits on one venue, Polymarket, with no second market to compare against. That concentration means the 9% consensus reflects the positioning of one trading pool rather than an average across independently priced markets, which is worth noting when reading the figure.
The case for
- Nvidia, TSMC, Broadcom, Arista or Super Micro would need to fall 50% (or SOXX 40%) from their all-time highs within 90 days of 1 January 2027, which a sharp risk-off move in AI-linked equities โ of the kind seen in past tech corrections such as 2000 or 2022 โ could produce if it hit multiple names at once.
- A pullback in hyperscaler AI capex guidance, if it materialized publicly before the window closes, would be read by markets as a demand shock across the chip supply chain and could accelerate stock declines toward the required thresholds.
- H100 rental prices, already trending down as newer GPU generations arrive, could plausibly reach the $1.00-for-five-days threshold on the SiliconData index without a broader crash, satisfying one of the three required triggers.
- A funding shortfall or failed round at OpenAI or Anthropic, while not currently signaled in public reporting, would immediately satisfy one of the six conditions and materially raise the odds of hitting the three-trigger threshold.
The case against
- The 9% consensus indicates the market currently sees the compound, three-of-six threshold as a high bar rather than a likely outcome.
- Nvidia, TSMC, Broadcom, Arista and Super Micro have all traded near record highs through 2026, meaning a 50% (or 40% for SOXX) decline would require a historically large reversal, not an ordinary pullback.
- OpenAI and Anthropic have both raised capital at valuations above $100 billion from major technology and investment backers, making bankruptcy or a forced acquisition within a narrow 90-day window unlikely absent a specific disclosed shock.
- The rule requires at least three of six distinct conditions โ spanning equities, corporate solvency and GPU rental pricing โ to occur within the same 90-day window, a much harder bar than any single metric moving on its own.
What to watch
Watch quarterly earnings and capital-expenditure guidance from Microsoft, Google, Amazon and Meta through late 2026 and into January 2027, since any public pullback in AI data-center spending would pressure the chip-stock triggers. Watch Nvidia, TSMC, Broadcom, Arista Networks and Super Micro Computer stock prices relative to their all-time highs as the 1 January 2027 marker approaches, since the 90-day resolution window runs from that date. Watch for any funding, restructuring or acquisition news involving OpenAI or Anthropic, and watch the SiliconData Silicon Index for H100 rental pricing, since a sustained drop to $1.00 or below for five consecutive days is one of the six specific triggers.
Trade this contract
Open on Polymarket
- gas covered