Will the monthly average compute price of NVIDIA's B200 be above $6.00 in December 2026?

- [2026-09-27] The Ornn Compute Price Index (OCPI-B200) — the resolution source named in the market rules — settled at **$8.08/GPU-hour** for NVIDIA B200, up 34.4% over the trailing 30 days from $6.01 and up 4.8% over the trailing 7 days. https://data.ornn.com/markets/b200 - [2026-09-27] The OCPI-B200 has moved +4.8% over the last seven days and +34.4% over thirty days (from $6.01 per GPU-hour). The three-month public window spans $4.38 (low) to $8.08 (high). https://data.ornn.com/markets/b200 - [2026-09-14 to 2026-09-27] Daily OCPI-B200 settlements over the last two weeks show a steady climb: $7.19 (Sep 14) → $7.21 (Sep 15) → $7.03 (Sep 16) → $7.10 (Sep 17) → $7.26 (Sep 18) → $7.43 (Sep 19) → $7.71 (Sep 20) → $7.71 (Sep 21) → $7.77 (Sep 22) → $7.95 (Sep 23) → $8.01 (Sep 24) → $8.04 (Sep 25) → $8.04 (Sep 26) → $8.08 (Sep 27), i.e. the index has been continuously above $6.00 through the whole visible window and is now trending upward. https://data.ornn.com/markets/b200 - [2026-09-27] Across Ornn's free-tier comparison table, B200 remains the highest-priced GPU tracked, well above H200 ($4.62) and H100 SXM ($2.49). https://data.ornn.com/markets - [2026-06-30] Ornn separately publishes compute forward curves for H100, H200 and B200 built on OCPI; a compute forward curve is the set of prices the market will pay today for GPU capacity delivered at some future date, whether that is at the end of the month or several years out , meaning a December 2026 forward mark exists in principle, though the specific December level was not accessible without a paid subscription. https://x.com/OrnnExchange/article/2071933704163745971 - [2026-04-27] Historical context: the OCPI-B200 spiking from $2.31 to $4.95 per hour since early March 2026, illustrating that B200 pricing has been on a sustained multi-month uptrend tied to frontier-model release cycles and Blackwell supply constraints, well before the current $8 level. https://tomtunguz.com/b200-gpu-pricing-spot-market-model-releases/ - [2026-03-30] A rival index (Silicon Data SDB200RT) showed B200 briefly touching the $6/hr range in late March 2026 before pulling back to $5.48, i.e. earlier in the year the price oscillated around/below $6, underscoring the volatility of this GPU category, though the metric that resolves this market is specifically Ornn's, not Silicon Data's. https://www.silicondata.com/blog/b200-rental-price-march-2026-update - [2026-09-29] Independent list-price aggregators corroborate an elevated B200 market in September 2026: Hyperbolic lists the lowest tracked fixed B200 rate at $5.99/GPU-hr as of September 2026. Vast.ai now has a $6.82/GPU-hr marketplace median, while AWS starts at $14.24/GPU-hr . https://www.thundercompute.com/blog/nvidia-b200-pricing - [2026-09-28] Another aggregator (GetDeploying) reports the median on-demand price is $6.49 per GPU per hour across 21 providers... having risen about 4% over the past 90 days to $6.49 per GPU per hour, and about 16% over the past 12 months , broadly consistent with Ornn's transaction-weighted figure being higher due to including premium hyperscaler-heavy volume. https://getdeploying.com/gpus/nvidia-b200 - [2026-09-27] Ornn's site states the index is a volume-weighted, winsorized average of transacted B200 rental prices across the neocloud market — real trades between buyers and sellers, not scraped listings, surveys, or provider rate cards , and it is the exact index cited in the Kalshi rules ("Ornn"). https://data.ornn.com/markets/b200 BOTTOM LINE: As of the most recent available settlement (Sep 27, 2026), the Ornn OCPI-B200 — the specific series this contract resolves against — sits at $8.08/GPU-hour, more than $2 above the $6.00 strike and rising sharply (+34% in 30 days), with no sign of a near-term reversal; independent list-price trackers corroborate a market solidly in the $6–8+ range. Absent a sharp reversal in Blackwell demand/supply dynamics between now and December 2026, the current-state evidence strongly favors the price staying above $6.00. NOT FOUND: The specific Ornn forward-curve price/mark for December 2026 B200 delivery (page requires a paid subscription); any official announcement of a supply glut, price collapse, or major Blackwell oversupply event scheduled before December 2026.

- [2026-04-27] An independent tracker citing Ornn's B200 spot index reported the price hitting $4.95 per hour that week, up from $2.31 in early March, a 114% surge in six weeks , with the author projecting "B200 likely settles above $5.00 for the summer" (tomtunguz.com). - [2026-07-14] Kalshi/press coverage reported NVIDIA's B200 per-hour compute rate peaking at $6.11 on May 30, 2026, then dropping to $4.22 by June 21 — a roughly 31% decline in less than a month , illustrating that the index has already crossed the $6 threshold at least once, but has also fallen well below it within weeks (KuCoin/CryptoBriefing). - [2026-07-14] A raw Ornn API sample dated July 14, 2026 shows the B200 index_value at 7.27625 USD/hour, i.e. above $7, versus H100 at ~$2.51 — evidence the index has spent time well above $6 mid-2026 (dashboard.ornnai.com/docs). - [2026-09-13] A trader/analyst who pulled Kalshi's public API on Aug 12, 2026 and integrated strikes found the B200 curve backwardated: "From September 2026 to March 2027, the market expects B200 rental prices to fall by about 9%, then trade around $5.30 to $5.40" , i.e. the crowd-implied forward price was already below the $6 strike (billhsu.substack.com). - [2026-09-13] The same analysis notes the current (mid-Sept 2026) implied B200 spot on Kalshi was roughly "$5.83 today" against an implied ~$5.30 six months out, and flags that liquidity/strike-range limitations mean the $3–$7 strike grid used for the curve was already "close to the upper bound" of the implied price — suggesting the market sees a real but not central probability of readings above $6 (billhsu.substack.com). - [2026-09-24] Ornn's own settled index shows OCPI-H200 at $5.16/hr (September 24, 2026) ; historically B200 has traded at roughly a $0.28–$1.80 premium to H200 depending on demand cycles (tomtunguz.com), implying an Ornn B200 reading anywhere from about $5.4 to $7.0 in the current regime. - [2026-09-28] Separate public-listing trackers (not Ornn, but a useful sanity check on market direction) show B200 median on-demand cloud pricing around $6.49 per GPU-hour across 21 providers, up about 4% over the past 90 days and about 16% over the past 12 months (getdeploying.com), and a competing index puts B200 and B300 medians at $6.52 and $7[+] (aimultiple.com) — both above the $6 line, though these track list/marketplace prices rather than Ornn's transaction-weighted OCPI. - [2026-08-18] Bank of America's compute-price tracking (a separate methodology from Ornn) put B200 spot near $5.66 an hour as of mid-August 2026... and $5.63 an hour a week later , i.e. modestly below $6 and roughly flat/slightly declining (shattered.io summary of BofA notes). - [2026-05-31] A prior Kalshi B200 monthly contract used a strike of "above $4.82 on May 31, 2026" , showing Kalshi calibrates strikes near the prevailing spot level each cycle rather than using a fixed $6 line — the $6 strike for December 2026 is meaningfully higher than earlier-year strikes, consistent with the underlying having trended up over 2026. - [2026-07-14] Kalshi's compute markets and forward curves for B200 launched in July 2026, publishing forward curves live for Nvidia B200, H200, and A100 chips , meaning there is no multi-year track record of these specific contracts resolving Yes/No — the reference class is thin (roughly two-three quarters of live data) and the underlying index itself only went live in around October 2025 (reaching six months as a live index by April 2026) . BOTTOM LINE: The base rate here is built on a very young (under one year) and highly volatile index: B200 per-hour pricing has swung from roughly $2.31 to above $7 within 2026 alone, and has closed above $6 on at least a few occasions but has also fallen sharply (e.g., ~31% in three weeks) shortly after peaks. The most forward-looking evidence — Kalshi's own market-implied forward curve — was backwardated as of mid-2026, pointing toward B200 settling in the ~$5.30–$5.90 range by early 2027, i.e. modestly below the $6.00 December 2026 threshold, though within the historical volatility band that could easily push the actual print above or below that level. NOT FOUND: A precise, dated Ornn OCPI-B200 settlement value for late September 2026 (only H200's $5.16 settlement and a July B200 data point of $7.28 were available); any official Kalshi resolution history showing how many prior monthly B200 strike contracts settled Yes vs. No; a full time series of December-dated B200 contracts' current traded prices (excluded per instructions).

YES means: The Ornn OCPI-B200 index prints a December 2026 monthly average above $6.00/GPU-hour. - [2026-09-17] Neocloud provider Nebius notified customers it will raise on-demand GPU pricing effective October 1, 2026, with a customer pricing notice indicating that Nebius plans to raise on-demand GPU rates from October 1, including roughly 17% for H100, 20% for H200, 19% for B200 and 21% for B300 — a concrete, scheduled, named-provider price increase for the exact chip in question that lands squarely inside the fourth quarter runway to December. - [2026-09-17] This is Nebius's second hike in a single quarter, and independent commentary frames it as capacity rationing rather than simple cost pass-through: it's closer to basic capacity rationing, the same mechanism that pushed cloud GPU rental rates up during the original 2023 H100 shortage, except this time it's spreading into CPU and memory tiers that used to be treated as commodity add-ons , and the report notes when retail-facing prices move twice in a single quarter, it usually means the wholesale side of that equation, chip allocation, power contracts, or memory supply agreements, is moving even faster underneath . - [2026-09-27] A separate market-structure analysis explains why elevated Blackwell pricing can persist even as headline capacity grows: forward orders remove supply before it reaches the open market — hyperscaler and neocloud commitments take inventory out of circulation months before it is racked, so what reaches spot and on-demand is the residual. A market clearing on residual supply prices higher than one clearing on total supply, even when total supply grows . - [2026-08 approx.] A memory-cost pass-through mechanism specific to Blackwell is documented: H200 and B200 both require HBM3e, so as Blackwell ramps, it compounds the same bottleneck that is already constraining H100 H200 supply , and the downstream effect on cloud pricing is less obvious but real: HBM shortages raise GPU memory subsystem costs, which flows into lease and rental prices even when a cloud provider does have inventory... The hardware itself costs more to source . - [2026-08-16] Contract/premium-tier pricing for the newest silicon has been firming through 2026 even as commodity-tier (older GPU) pricing softens: at the reserved / contract / newest-silicon tier, prices firmed and in several documented cases rose during the first half of 2026, because premium capacity stayed effectively sold out. But the 2026 story is bifurcated — spot softened while contract and Blackwell prices rose , consistent with B200 (the newest widely-deployed silicon) staying on the high side of the pricing spectrum into Q4. - [2025-11-25 / most-recent qtr.] NVIDIA's own reporting shows demand still running ahead of what it can ship: Blackwell/GB300 mix has been strong and management has repeatedly stated Blackwell demand exceeds supply, with CEO Jensen Huang stating "Blackwell sales are off the charts, and cloud GPUs are sold out. Compute demand keeps accelerating and compounding across training and inference — each growing exponentially" . - [2026-Q3 filing/analysis] Supply-side ordering has surged even as NVIDIA races to keep up: Nvidia's total supply-related commitments, such as for CoWoS wafers, HBM memory, or other components, surged nearly 52% QoQ to $50.3 billion in Q3... This is a notable increase from the prior five-quarter average of ~$30 billion , indicating demand visibility strong enough that NVIDIA itself is not forecasting an imminent supply glut before year-end. - [2026-03] Hyperscaler capex, the demand engine behind B200 rentals, remains historically large: hyperscaler capex is accelerating. The four largest cloud companies (Alphabet, Amazon, Meta, Microsoft) are projected to spend a combined $700 billion on capital expenditures in 2026, the vast majority on AI infrastructure. This spending flows directly through NVIDIA's data center business , sustaining the bidding pressure on scarce Blackwell capacity through Q4. - [2026-09-27] Even TSMC capacity-easing narratives concede relief is only gradual and partial by Q4: TSMC projections indicate 120K+ units/month by Q4 2026, vs. 70K-80K in Q2 2026 — roughly 50-70% more supply, which is a large increase but still needs to catch up to demand described elsewhere as exceeding supply by 1.4-1.6x, meaning a full price collapse to sub-$6 by December is not guaranteed even under this bullish supply scenario. - [2026-09-27] Independent cross-checks corroborate that the direction of travel into autumn 2026 is up, not down: Ornn's own published series shows the index rising from $6.01 thirty days prior to $8.08 on Sep 27 with no down-day in the trailing two weeks, and Ornn explicitly frames this as an accelerating trend (+4.8% in 7 days vs +34.4% in 30 days, i.e., the pace has not decayed). BOTTOM LINE: The single most likely path to YES is simple trend continuation plus a mechanical floor from scheduled neocloud price hikes (Nebius's ~19% B200 increase effective Oct 1) landing directly in the run-up to December, reinforced by persistent hyperscaler-driven demand and NVIDIA's own supply-commitment data showing no near-term glut; for the index to fall from $8.08 back under $6.00 by December would require roughly a 26%+ reversal, which is larger than any pullback seen in the trailing three-month window's high-to-close range. NOT FOUND: An explicit December 2026 Ornn forward-curve print for B200 (paywalled); any hyperscaler or NVIDIA statement specifically forecasting a Q4 2026 B200 rental-price decline; confirmation of whether Nebius's ~19% B200 hike is representative of the broader neocloud market Ornn samples from.

YES means: the Ornn OCPI-B200 index's arithmetic mean of hourly values in December 2026 is above $6.00/GPU-hour (contract resolves Yes). This report argues the NO case — that despite the current elevated reading, the December 2026 average lands at or below $6.00. - [2026-07-14] The index has already demonstrated it can swing from above $6 to well below within weeks: press coverage found NVIDIA's B200 rate peaking at $6.11 on May 30, 2026 then falling to $4.22 by June 21, a ~31% drop in under a month — proof the current $8.08 print (Sep 27) is not a floor and mean-reversion of this magnitude would put a monthly average for December comfortably under $6 even from a much higher starting point (KuCoin/CryptoBriefing, cited in prior team report). - [2026-06-15] TrendForce/Economic Daily reporting shows the TSMC CoWoS supply-demand gap — the binding constraint behind Blackwell/B200 scarcity pricing — is expected to narrow significantly from around 20% currently to about 10% by the end of 2026, with conditions likely to improve further in 2027 , directly targeting the Nov-Dec 2026 window this contract covers. - [2026-06-15] The same report quantifies the capacity unlock: TSMC's monthly CoWoS capacity could reach a record 120,000 to 140,000 wafers in 2026 , which independent industry trackers tie explicitly to Blackwell/B200 output ramping through Q4. - [2026-07-29 update / 2026-07-03 published] A supply-chain analysis lays out an explicit month-by-month path to sub-$6 pricing: it projects that by TSMC projections indicate 120K+ units/month by Q4 2026, vs. 70K-80K in Q2 2026 , and forecasts Higher B200 supply doesn't guarantee margin expansion... Competitive pricing pressure from AMD, custom alternatives, and customer negotiating power—now that scarcity recedes—will compress gross margins on Blackwell in H2 2026 — its own explicit "Q4 2026" outlook states B200 approaches commoditization with pricing discounts beginning. - [2026-09-27] Spheron's live tracker — an independent, non-Ornn spot-price source — shows Spheron B200 instances list at $4.33/hr spot as of 27 Sep 2026 , more than $3.75 below Ornn's simultaneous $8.08 print, underscoring how wide the dispersion is across venues and how a shift in Ornn's transaction mix toward lower-priced spot deals could pull its volume-weighted average down sharply. - [2026-09-27] The same source frames the shortage as narrowing and segment-specific: the shortage is real but it is no longer general... The constraint now lives almost entirely in the top-end SXM parts , implying the acute scarcity driving September's B200 spike is not guaranteed to persist through year-end as more capacity comes online. - [2026-02-28] Independent analysis of GPU-generation pricing cycles argues the standard pattern is Current-gen supply ramp → 40–50% price decline over 12–18 months... Applied to Blackwell, this means B200 prices will likely decline gradually through 2026 as supply matures , consistent with a structural downward path even if short-term spikes (like Sep 2026's) occur. - [2026-04-16] A parallel piece on B300 substitution notes the historical playbook — The H100 went from $8/hr in early 2024 to under $3/hr in 2026 — and that B300 (which competes for the same demand pool and can substitute for B200 in many workloads) is itself expected to compress once next-gen Rubin volume arrives, which would pull demand and pricing pressure off B200 specifically. - [2026-09-24 via prior team report] Ornn's own H200 settlement of $5.16/hr — a GPU one tier below B200 — sits well under the $6 line; if B200's current unusually wide premium over H200 (roughly $3/hr vs. a historical $0.28–$1.80 band per prior reporting) normalizes even partially by December, B200 could fall back toward the $5.40–$6.00 range implied by that historical spread. - [2026-09-13, cited in prior team base-rate report] The market's own forward-curve read (via Kalshi's traded strikes) was backwardated as of mid-September, with the crowd pricing B200 falling toward roughly $5.30–$5.90 over the following two quarters — i.e., sophisticated positioning as of just two weeks before today's data pull still leaned toward sub-$6 by early 2027, despite the concurrent Ornn spot spike. - [2026-09-29] Aggregate industry commentary continues to frame the Blackwell shortage as easing rather than worsening into Q4, with capacity phase-ins (CoWoS-L Zhubei additions, SK Hynix M15X HBM3e ramp) explicitly dated for Q3–Q4 2026, i.e., landing inside the contract's resolution month. BOTTOM LINE: The NO case rests on (1) documented precedent that this specific index has fallen >30% in under a month before, (2) a broad, multi-source consensus that CoWoS/HBM supply materially expands in Q4 2026 specifically, easing the scarcity premium behind the current $8.08 spike, and (3) the market's own forward curve (as of mid-September) still implying sub-$6 levels for early 2027 — meaning December's monthly average could plausibly mean-revert below $6.00 even after a strong September, though this requires a reversal from currently very elevated and still-rising levels. NOT FOUND: A precise Ornn OCPI-B200 forward mark specifically for the December 2026 contract period (paywalled); confirmation of the magnitude/timing of any December price decline actually realized (data doesn't exist yet, event is in the future); the full Kalshi B200 resolution history for prior monthly contracts (Yes/No count).