How the contract works
A contract on this question settles at $1 if the outcome happens and at nothing if it does not. The price in between is simply what buyers and sellers currently agree the chance is, expressed as a number between zero and one: a contract trading at 0.30, for example, would mean the market treats the outcome as happening about three times in ten. Here the outcome is an agreement by any entity to acquire Nebius Group, confirmed by the company or by credible financial reporting, on or before 31 December 2026 at 11:59 PM ET. The agreement is what settles the market โ the deal does not have to close, or even survive, for the contract to resolve Yes. A position can normally be sold before the settlement date at whatever the price is at that moment, rather than held to the end.
Yes11%
No89%
Probability
How the price has moved
The contract was first recorded on 29 July 2026 at 35% and, within the same short window, traded as high as 79% and as low as 31%. That upper figure matters: at 79% the market was treating a signed acquisition agreement as the likeliest outcome. It then fell back hard, and now sits below the bottom of that recorded band. Across 115 observations and roughly $7.9 million of volume, the last 24 hours show no movement at all. No single publicly reported trigger accounts for the reversal, and it would be dishonest to assign one; what can be said is that the pattern โ a fast spike, a faster collapse, then a flat line โ is the shape a market makes when a takeover story circulates and is not confirmed. The flatness now is itself information. A market that has stopped moving after a violent repricing is a market that considers the question provisionally answered and is waiting for new facts rather than re-arguing the old ones.
Analysis
Context
Nebius Group is the Amsterdam-headquartered, Nasdaq-listed company that emerged from the break-up of Yandex. After Yandex N.V. sold its Russian businesses in July 2024, what remained was renamed Nebius Group and refocused on AI cloud infrastructure: GPU capacity rented to companies training and running large models. Founder Arkady Volozh returned as chief executive. The European Union removed him from its sanctions list in 2024, which was a precondition for the company operating normally in Western markets.
Since then Nebius has become one of the most closely watched of the so-called neoclouds โ the group of GPU landlords that includes CoreWeave and Crusoe. Nvidia took part in a private placement in the company in December 2024. In September 2025 Nebius announced a multi-year capacity agreement with Microsoft worth more than $17 billion, which transformed how the market valued the business and made it a name that large-cap technology investors track. Alongside the cloud business, the group still holds Avride in autonomous driving, the data business Toloka, the education company TripleTen and a stake in ClickHouse.
That combination โ scarce GPU capacity, contracted revenue from a hyperscaler, and a valuation that has moved violently โ is exactly what generates takeover speculation. The AI infrastructure sector has consolidated fast, and any company with power, data centre shells and Nvidia allocation is a theoretical target. This market asks a narrower question than the speculation does: whether an actual agreement to acquire the group is signed before the end of 2026.
The most informative fact about this market is not its current level but how far it has travelled. The first recorded price, on 29 July 2026, was 35% โ already a high figure for a takeover of a company this size, and a sign that participants were reacting to something specific rather than pricing background probability. Within the observation window the contract ranged from 31% up to 79%. A price near 79% means the market briefly treated a signed agreement as a strong favourite. The current consensus sits in low double digits, below the bottom of that recorded band. That is a full reversal, not a drift, and it happened in a matter of days across 115 recorded observations.
The last 24 hours show no change at all โ a flat reading of 0.0 percentage points. Flat after a collapse usually means the market has finished absorbing whatever it was reacting to and has settled on a new baseline. Total volume across venues is roughly $7.9 million, concentrated entirely on Polymarket, which is the only venue listing the question. Because there is a single venue, there is no cross-venue spread to read as a disagreement signal; the price is one order book's view, but a reasonably deep one for a single-company corporate event.
The structural case for a low probability is straightforward. Nebius is not a small target. After the Microsoft capacity agreement announced in September 2025 the company is valued in the tens of billions of dollars, which limits the set of plausible buyers to a handful of hyperscalers, chip vendors and infrastructure consortia โ most of which are already Nebius customers, suppliers or investors rather than natural acquirers. Nvidia, an investor since the December 2024 placement, has consistently backed neoclouds as customers rather than absorbing them. Microsoft is contracting for capacity, and buying the landlord would mean taking on the rest of the group as well.
Ownership is the second constraint. Nebius inherited the share structure and shareholder base of Yandex N.V., with Volozh as founder, chief executive and the most significant individual holder. A transaction that he does not want is not realistically available to a bidder, which removes the hostile route entirely and makes the question a matter of whether management chooses to sell. Management has spent two years rebuilding the company around a growth story; a sale in the middle of it is a decision, not an event that happens to them.
The third constraint is regulatory. A company with Nebius's history would draw scrutiny from Dutch and EU authorities on any change of control, and from US reviewers if the buyer were foreign. That does not block an agreement โ and the market settles on the agreement, not the closing โ but it lengthens the path from talks to a signed deal, which matters when the deadline is 31 December 2026. Roughly five months remain. For the contract to resolve Yes, negotiations would in practice need to be underway now.
What moves the probability
Whether a real process exists
Everything turns on whether the group is running a sale process or fielding an approach. The swing from 35% to 79% and back suggests the market briefly believed one existed and then stopped believing it. Confirmed reporting of an approach would push the probability up sharply; continued silence lets it decay toward the base rate for a large listed company being acquired in any five-month window.
Founder and board control
Volozh's position as founder, chief executive and major shareholder means no bidder can go around the board. This is the single largest structural drag on the probability, because it converts the question from 'is Nebius attractive?' to 'does its founder want to sell?'. It also means any credible signal would most likely come as a negotiated announcement rather than a leaked hostile approach.
The Microsoft contract
The multi-year capacity agreement announced in September 2025 cuts both ways. It makes Nebius more valuable and more strategically interesting, which supports takeover speculation. But it also gives management a reason to stay independent and creates change-of-control complexity that a buyer would have to negotiate through, which pushes the probability down.
AI infrastructure consolidation
The broader sector is consolidating, and capital is chasing power, land and GPU allocation. A large transaction elsewhere in the neocloud space would raise the perceived probability here by demonstrating that buyers exist at this scale. A funding squeeze or a sharp repricing of AI infrastructure equities would do the same, for the opposite reason โ distressed sellers agree to deals that confident ones refuse.
The five-month clock
Settlement is 31 December 2026. Large cross-border technology deals with sanctions history and multiple regulators rarely go from first contact to signed agreement in under a quarter. Every week without reported talks mechanically lowers the probability, which is why a flat, low price this late in the year is consistent with the market treating the question as close to answered.
Subsidiary sales are not the group
Nebius still holds Avride, Toloka, TripleTen and a ClickHouse stake. Reported interest in any of these could be misread as takeover news and move the price temporarily. Under the settlement rules none of it counts, so such moves tend to reverse once the detail is clear.
The case for
- A buyer would have to enter into an agreement for Nebius Group as a whole, confirmed by the company or by consensus reporting, on or before 31 December 2026 โ talks alone are not enough, but a signed deal that later collapses would still count.
- The most plausible route is a negotiated approach from a hyperscaler or chip vendor seeking guaranteed GPU capacity and contracted revenue, agreed with Volozh and the board rather than pursued against them.
- A sharp fall in AI infrastructure valuations or a financing squeeze during the autumn of 2026 could turn a reluctant seller into a willing one, which is the historical pattern for how large technology deals actually get signed.
- The market's own history shows participants once priced this near 79%, which means the scenario was recently considered plausible by people committing money to it.
The case against
- No acquisition has been confirmed, and the price collapsed from the high end of its recorded range back into low double digits without a Yes resolution, which is what a rumour looks like when it fails to firm up.
- Nebius's size after the September 2025 Microsoft agreement narrows the buyer list to a handful of firms, most of which are already partners, customers or investors and have shown no appetite for owning neocloud operators outright.
- Founder and management control means a bidder cannot force the outcome; a deal requires an active decision to sell from people currently executing a growth plan.
- Only about five months remain to settlement, and a change of control involving a company with Nebius's regulatory history would need Dutch, EU and probably US review โ a timeline that discourages bidders from starting now.
What to watch
Nebius's quarterly results and the accompanying calls are the main scheduled events before settlement; management commentary on strategic alternatives, or a refusal to comment on speculation, moves this more than anything else. Watch for regulatory filings that reveal an approach โ under Dutch and US disclosure rules, a material agreement must be announced promptly, so the resolution here would arrive as a press release rather than as a rumour. Watch also for changes in the shareholder register and any new large capacity contracts, since a company signing multi-year customer deals is a company planning to stay independent. Sector-wide, any announced consolidation among AI infrastructure providers, or a sharp drawdown in neocloud equity valuations during the autumn, would reprice the whole category and this contract with it. The deadline itself is 31 December 2026 at 11:59 PM ET, and the probability should decay steadily toward zero as that date approaches without news.
Trade this contract
Open on Polymarket
- gas covered
