Open Source market share on Sep 30th
Short Answer
Open source market share is not expected to exceed 69% by September 30, 2026; the market anticipates only a modest upward shift driven by general open-weight AI growth and the mid-September 2026 DeepSeek V4.1-Flash release (20.0%).
1. Market Behavior & Drivers
The primary explained price move in this market was a spike on September 27, driven by reports of significant gains for Chinese open-source AI models. News that these models had reached 57% to 67% of developer usage on OpenRouter brought the share close to the contract's 66% resolution threshold, increasing the implied probability of a "Yes" outcome.
Subsequent major price movements lack clear drivers. No specific event or report was found to coincide with the large spike on September 28, the sharp drop on September 29, or the final spike on September 30. The research notes that social media and news activity around September 29 indicated a trend that contradicted the market's price decline. This contract settles based on data from Vercel AI Gateway's Open vs. Closed Token Volume chart.
- The 66% threshold appears plausible, boosted by DeepSeek V4.1-Flash traffic redirection.
- Open-source momentum and DeepSeek V4.1-Flash also support the 69% threshold.
- Above 84% faces headwinds from new closed-source models and Vercel's user base.
Who Wins and Why
| Outcome | Market | Model | Why |
|---|---|---|---|
| Above 84% | 1.0% | 0.9% | Substantial increase is difficult due to new closed-source models and Vercel's specific user base. |
| Above 87% | 2.0% | 0.5% | An even greater leap faces headwinds from growth magnitude and influence of new closed-source offerings. |
| Above 78% | 1.0% | 2.5% | Reaching this threshold demands significant gains amid potential competition from new closed-source models. |
| Above 75% | 1.0% | 3.9% | Achieving this level requires sustained growth beyond the general positive trajectory of open-weight AI. |
| Above 81% | 1.0% | 1.5% | This threshold demands significant market share growth given new closed-source model impacts and Vercel's user base. |
Current Context
Open-weight AI is likely to exceed 66% market share by late 2026. The Kalshi prediction market on open-weight AI token volume for September 30, 2026, settles based on Vercel AI Gateway's Open vs. Closed Token Volume chart on October 1, 2026 [1]. The key threshold for this market is open-weight share above 66% [1]. As of September 28, 2026, CoinRithm reported the Kalshi contract for “Above 84%” at a 5% implied probability; this is a market price, not a final measurement [1]. Based on this market framing and recent adoption trends, open-weight AI is more likely than not to exceed the 66% threshold, though confidence is limited until the final Vercel observation occurs [1][2][3][4].
Open models show rapid adoption, driving recent market share gains. The September 2026 State of Open Source AI materials report that August 2026 marked the first month an open model led OpenRouter in requests [2][3]. Open models powered approximately one-third of tokens in production, while U.S. closed providers' combined request share decreased to 44% in late August from 54% in May [2][3]. This OpenRouter evidence is limited to routed traffic and excludes first-party usage from platforms like ChatGPT, Gemini, and Doubao, so it should not be treated as the entire global AI market [3]. The 2026 open-source AI trend is driven by Chinese and other open-weight model families, including Qwen, DeepSeek, GLM, and Kimi [2][3]. Chinese models surpassed 45% of weekly traffic and represented 61% of OpenRouter's ten most-used models [2][3]. Secondary commentary suggests continued share gains due to narrowing benchmark gaps, lower inference costs, and growing downloads, though numerical claims lack independent validation and should be considered directional [5][6].
Broader commercial market forecasts for open-source software vary materially by definition. The Business Research Company estimates open-source software revenue at $56.57 billion in 2026, projected to reach $95.38 billion in 2030 with a 14% CAGR [7]. Another forecast estimates $51 billion in 2026 and $110.39 billion in 2035 [8]. Open-source services are forecast separately at $40.79 billion in 2025 and $153.50 billion in 2034, growing at a 15.86% CAGR [9]. Expert and industry commentary identifies enterprise adoption of agentic AI, EU Cyber Resilience Act/DORA compliance, security, sustainability, digital sovereignty, and reinvestment in open infrastructure as major developments for 2026 [4][10].
Sources (10)
- 1Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 2The State of Open Source AI — v1.1 · September 2026stateofopensource.ai
- 3State of Open Source AI 2026stateofopensource.ai
- 42026 State of Open Source Report | OpenLogicopenlogic.com
- 5How Open Source Quietly Took Over AI: From Cheap Alternative to the Only Way Forwardindex.dev
- 6Open Source vs Proprietary AI 2026: The Data Says It All | AI Forward — Practical AI Guides for Business & HRaifwd.net
- 7Open Source Software Market Size, Trends Report 2026-2030thebusinessresearchcompany.com
- 8Open Source Software Market Share | Industry Report 2035marketresearchintellect.com
- 9Open Source Services Market Size, Share & Demand by 2034theinsightpartners.com
- 10What's in Store for Open Source in 2026appdevelopermagazine.com
2. Price Chart
Historical Price (Probability)
Outcome probability
Date
3. Significant Price Movements
Notable price changes detected in the chart, along with research into what caused each movement.
Outcome: Above 69%
📈 September 30, 2026: 15.0pp spike
Price increased from 5.0% to 20.0%
What happened: The provided research does not identify a primary driver for a 15.0 percentage point spike in the "Open Source market share on Sep 30th" prediction market. While a similar market for September 26, 2026, resolved with "Above 69%" at 99% implied probability, no specific social media activity or traditional news announcement is linked to a sudden 15.0 percentage point price movement for the September 30th date [1]. Although a September 21, 2026, commentary discussed significant open-weight model usage, this was an opinion piece and not a verified market data release or a direct cause for a spike on the specified date [2]. Therefore, social media activity appears to be irrelevant as a primary driver for this specific, unevidenced market movement.
Outcome: Above 75%
📉 September 29, 2026: 74.0pp drop
Price decreased from 78.0% to 4.0%
What happened: The provided web research does not identify a primary driver for a 74.0 percentage point drop in the "Above 75%" outcome for the "Open Source market share on Sep 30th" market on Sep 29, 2026. The most relevant social media activity and news around that time indicated a contradictory trend. Specifically, Vercel announced on September 19, 2026, a record 78.4% open-weight share of AI Gateway token volume, which was amplified on X as evidence of open models gaining significant developer traffic [3][4]. This news, if influential, would typically cause the "Above 75%" outcome to rise in price. Therefore, social media, based on the available information, was not a primary driver of this specific market price drop; rather, the existing social activity pointed toward an increase in open-source market share [3][4].
Outcome: Above 72%
📉 September 28, 2026: 56.0pp drop
Price decreased from 70.0% to 14.0%
What happened: The provided web research does not identify a prediction market event on September 28, 2026, for "Open Source market share on Sep 30th" experiencing a 56.0 percentage point price drop from above 72%. No social media activity, traditional news, or market structure factors were found that would correspond to this specific market movement [5]. While some reports discuss open-source market share for 2026, they do not describe a contemporaneous 56.0 percentage point decline or a related social media catalyst [5]. Therefore, social media cannot be identified as a primary driver for this unevidenced event.
Outcome: Above 66%
📈 September 27, 2026: 67.0pp spike
Price increased from 2.0% to 69.0%
What happened: The 67.0 percentage point spike in the "Open Source market share on Sep 30th" prediction market was primarily driven by traditional news and data reporting on September 27, 2026. Reports indicated that Chinese open-weight/open-source AI models had reached approximately 57%–67% of OpenRouter developer traffic by mid-September, with contemporaneous briefings describing them as capturing over half of traffic on major developer platforms [6]. This trend was reinforced by Financial Times-reported data showing a sixfold year-over-year increase in mentions of open-weight models in earnings calls, reflecting growing enterprise adoption of lower-cost solutions [7]. Social media was largely irrelevant, as the movement was attributed to economic and operational shifts rather than a single confirmed viral post [6].
Sources (7)
- 1Open Source market share on Sep 26th - Kalshi Odds | CoinRithmcoinrithm.com
- 2Open Weights Are Now Statecraft | The Arc of Powerthearcofpower.com
- 3Vercel gateway data shows open models at a record 78.4% of tokensdeai.org
- 4Vercel sees low-cost AI take 78% of developer traffic — NRBnrb.bg
- 5Open Models Run 56% of Production Tokens: Latest AI Newsblog.buildfastwithai.com
- 6Chinese AI Models Went From Niche to Dominant in Under a Year, Here's the Full Story Nobody's Fully Connected Yet | AdvisioTechadvisiotech.com
- 7Corporate America shifts to lower-cost open AI models as spending pressure risestradersunion.com
4. Market Data
Contract Snapshot
This market resolves based on the open-source market share observed on September 30th. A "Yes" resolution occurs if the open-source market share is above the specific percentage threshold for the contract, while a "No" resolution occurs if it is at or below that threshold. The maximum payout deadline for this market is October 1, 2026.
Available Contracts
Market options and current pricing
| Outcome bucket | Yes (price) | No (price) | Last trade probability |
|---|---|---|---|
| Above 69% | $0.13 | $0.97 | 20% |
| Above 66% | $0.33 | $0.91 | 10% |
| Above 72% | $0.14 | $0.98 | 2% |
| Above 87% | $0.01 | $1.00 | 2% |
| Above 75% | $0.03 | $1.00 | 1% |
| Above 78% | $0.02 | $1.00 | 1% |
| Above 81% | $0.01 | $1.00 | 1% |
| Above 84% | $0.01 | $1.00 | 1% |
| Above 90% | $0.02 | $1.00 | 1% |
Market Discussion
The "Open Source market share on Sep 30th" market forecasts the open-weight share of Vercel AI Gateway token volume for September 30, 2026 [1]. As of September 28, 2026, the leading outcome of "Above 84%" was priced at a 5% implied probability, despite a prior market for September 14 indicating high confidence for clearing a lower 67.5% threshold [1][2]. While a general narrative suggests open-source models are gaining token-throughput share, especially in programming and agentic workloads [3][4], some June 2026 commentary reported skepticism toward open-source dominance in production AI applications [5].
Sources (5)
- 1Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 2Open Source market share on Sep 14th - Kalshi Odds | CoinRithmcoinrithm.com
- 3State of AI: An Empirical 100 Trillion Token Study with OpenRouter | alphaXivalphaxiv.org
- 4Will open-source AI win (through 2025)? | Manifoldmanifold.markets
- 5Polymarket Pulse: June 23, 2026 – The AI Race Gets Weird - DEV Communitydev.to
5. Trust Index
Order book is critically thin.
Primary risk· Trade quality
How it adds up
Integrity80% of score82Good
Trade quality20% of score28Avoid
Includes the cost to trade: a $1,000 order can't be filled here because the order book is too thin.
Trust score71Good
Weighted blend with hard caps — a critically weak safety pillar, or a severe trading anomaly, caps the total regardless of the rest. Full methodology · About the Trust Index
Trust profile
Integrity4 screens run · 6 don't apply
6. What major open-source or closed-source model releases scheduled before September 30, 2026, could significantly alter Vercel's token volume metrics?
| DeepSeek V4.1-Flash Launch | September 10, 2026 [1] |
|---|---|
| OpenAI GPT-6.1 Sol Launch | September 29, 2026 [2] |
| Google Antigravity Input Context | 1,048,576 tokens [3][4][5] |
Several major AI model releases before September 30, 2026, could significantly alter Vercel's token volume metrics. These include key open-weight releases such as DeepSeek V4.1-Flash, alongside substantial closed-source developments. Closed-source models feature multiple Google Gemini Flash models, Google's Antigravity preview, OpenAI's GPT-5.6 family, and the notably potent GPT-6.1 Sol [1][6][7][3][4][5][2][8][9][10][11].
Specific open-weight model releases are poised to affect Vercel's open-model token volume. DeepSeek V4.1-Flash, a multimodal model launched on September 10, 2026, outperformed its predecessor V4-Pro, leading to V4-Pro requests being rerouted to it from September 14, indicating a potential shift in token usage [1][6]. Earlier in 2026, DeepSeek V4 Preview, introduced on April 24, also contributed to open-weight volume with its support for a 1M-token context [12]. Furthermore, Qwen's open-weight releases from 2025, particularly Qwen3-Coder on July 22, 2025, remain relevant due to its 480B total parameters, 1M-token extrapolation, and performance comparable to Claude Sonnet 4, which is particularly pertinent for Vercel's coding-heavy applications [13][14].
Closed-source developments are also expected to drive considerable changes in Vercel's token volume. Google introduced several Gemini Flash models, including Gemini 3.5 Flash on May 19, 3.6 Flash and 3.5 Flash-Lite on July 21, 3.7 Flash on August 13, and 3.8 Flash on September 2, with Flash-Lite specifically designed for high-volume automation [7]. Google's Antigravity preview, a managed autonomous agent with a 1,048,576-token input context, launched in September 2026, holds potential to further increase closed-model token volume [3][4][5]. OpenAI's GPT-5.6 family, which debuted on July 9, 2026, and saw expanded availability and reduced pricing in August, could lead to a substantial shift in traffic [8][9][10][11]. Most significantly, OpenAI's GPT-6.1 Sol, released on September 29, 2026, offers near-Astra intelligence for coding and professional tasks at one-fifth of Astra's token prices, positioning it as the strongest immediate catalyst for altering Vercel token volumes [2].
Sources (14)
- 1DeepSeek | Introducing DeepSeek-V4.1-Flash: smarter, faster, more efficient.deepseek.com
- 2Introducing GPT-6.1 Solopenai.com
- 3Changelogai.google.dev
- 4Release notes | Gemini API | Google AI for Developersai.google.dev
- 5Antigravity preview | Gemini API | Google AI for Developersai.google.dev
- 6DeepSeek | DeepSeek V4.1 Flash：更强、更快、更普惠deepseek.com
- 7Gemini deprecations | Gemini API | Google AI for Developersai.google.dev
- 8GPT-5.6: Frontier intelligence that scales with your ambition | OpenAIopenai.com
- 9Previewing GPT-5.6 Sol: a next-generation model | OpenAIopenai.com
- 10GPT-5.6 – August Updatescdn.openai.com
- 11Improving GPT‑5.6 Sol in ChatGPT—and expanding access to GPT-5.6 Luna for free users | OpenAIopenai.com
- 12DeepSeek | DeepSeek-V4 Preview: Entering the Era of Affordable Million-Token Contextdeepseek.com
- 13Qwen3: Think Deeper, Act Faster | Qwenqwenlm.github.io
- 14Qwen3-Coder: Agentic Coding in the World | Qwenqwenlm.github.io
7. How do leading open-weight models (e.g., Llama series, Qwen) and closed-weight models (e.g., GPT series, Gemini) compare on cost-per-token and performance benchmarks as of mid-2026?
| Closed Model API Input Cost (GPT-5.5) | $5 per 1M tokens (mid-2026) [1] |
|---|---|
| Open-Weight Hosted API Input Cost (DeepSeek V4 Flash) | $0.14 per 1M tokens (mid-2026) [2] |
| Hardest Task Performance (SWE-bench Verified) | Closed models ~95% vs ~80.5% for best open model [3][4] |
As of mid-2026, open-weight models generally provide lower cost-per-token. Performance benchmarks indicate a convergence in capabilities, though closed models retain a notable advantage on the most challenging tasks. It is important to distinguish that open-weight does not inherently mean fully open-source due to varied licensing agreements [5].
Closed frontier APIs generally exhibit higher cost-per-token than open-weight alternatives. Pricing data from mid-2026 highlights significant cost differences, with closed frontier APIs, such as GPT-5.5, priced at $5 per 1M input tokens and $30 per 1M output tokens. In contrast, open-weight hosted APIs presented considerably lower costs; for example, DeepSeek V4 Flash was listed at $0.14 per 1M input tokens and $0.28 per 1M output tokens [1][2]. The total cost of self-hosting open-weight models is dependent on factors like GPU price and utilization, which can alter the effective cost per token [5].
Performance gaps are narrowing, but closed models still lead advanced benchmarks. Benchmark comparisons from 2026 suggest a narrowing performance gap, with Kimi K3 reportedly within 0.5 points of GPT-5.6 Sol on Terminal-Bench (88.3% versus 88.8%). However, closed models maintained their lead on the hardest software-engineering and factual-recall tasks, achieving roughly 95% on SWE-bench Verified, compared to about 80.5% for the top open model [3][4]. Market participants in late September 2026 anticipated the open-weights share of token volume to be materially below 84%, with one major closed model provider's chances for the best AI model reportedly spiking by the end of 2026 [6][7].
Sources (7)
- 1Pricingai.google.dev
- 2Open Source LLM vs Paid LLM: Full 2026 Comparisonmayhemcode.com
- 3Best Open Source AI Models July 2026: Full Collectionbuildfastwithai.com
- 4Best Open Source AI Models & LLM Leaderboard (2026)lmmarketcap.com
- 5Open Source LLMs Compared: Best Models in 2026123ofai.com
- 6Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 7GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil Dealyoutube.com
8. Beyond OpenRouter, what do alternative data sources like Hugging Face downloads and enterprise adoption surveys indicate about the growth trajectory of open-weight AI through mid-2026?
| Hugging Face Total Downloads | 45.4 billion by October 2025 [1] |
|---|---|
| Open-Source AI Market Growth | 21.1% (from $19.05 billion in 2025 to $23.08 billion in 2026) [2] |
| Companies' Share of Top Downloads | 63.2% (October 2025) [1][3] |
Open-weight AI shows robust growth through mid-2026, driven by expanding ecosystems. Alternative data sources, including Hugging Face downloads and market research estimates, indicate a strong upward trajectory for open-weight AI adoption into mid-2026 [4]. The Hugging Face Hub accumulated 45.4 billion total downloads by October 2025, demonstrating rapid ecosystem expansion [1]. By summer 2026, public model repositories grew from 2.43 million to 2.96 million, datasets expanded from 711,000 to 1 million, and Spaces increased from 1.00 million to 1.44 million [5][3]. This period also saw a material broadening of open-weight supply and capability, exemplified by the Qwen model family accumulating approximately 2.045 billion downloads over the reported year [5][3].
Enterprise adoption drives growth, with downloads concentrated among top entities, reflecting practical deployment needs. While download activity is substantial, it remains highly concentrated; the top 50 open-source entities accounted for 80.22% of Hub downloads (36.45 billion of 45.44 billion) in October 2025 [1]. Furthermore, summer 2026 data indicated that just 1.5% of repositories generated 99.2% of total downloads [5][3]. This growth is increasingly oriented towards enterprise and deployment applications rather than purely hobbyist use [1][3]. Companies represented 63.2% of downloads among the 50 most-downloaded entities in October 2025, and over 30% of Fortune 500 companies reportedly maintained verified Hugging Face accounts [1][3]. Smaller models consistently attracted more downloads due to their practical deployment constraints [1][3].
Market forecasts predict significant financial growth, but downloads have important measurement limitations. A 2026 market research estimate projects the open-source AI model market to grow by 21.1% in one year, from $19.05 billion in 2025 to $23.08 billion in 2026, with continued expansion to $50.03 billion by 2030 [2]. However, Hugging Face explicitly cautions that download figures do not fully capture API usage, private deployments, model quality, or overall market share [5][3]. The search results did not provide a directly comparable enterprise-adoption survey statistic for the Vercel AI Gateway's Open-vs-Closed Token Volume percentage for September 30, 2026 [4].
Sources (5)
- 1Model statistics of the 50 most downloaded entities on Hugging Facehuggingface.co
- 2Open-Source AI Model Market Research Report 2026finance.yahoo.com
- 3Three Million Models and Countinghuggingface.co
- 4Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 5State of Open Models: Summer 2026 Observationshuggingface.co
9. What are the known methodologies and limitations of the Vercel AI Gateway's 'Open vs. Closed Token Volume' chart used for market settlement?
| Chart Measurement | Share of AI Gateway token volume from open vs. closed models [1][2] |
|---|---|
| Open-weight models (Aug 2026) | 56% of AI Gateway tokens but 14% of spend [3] |
| Data Updates | Daily values can be revised; data updated on a delayed operational schedule [3][2][4] |
Vercel's AI Gateway chart categorizes token volume by model weight publication. This chart specifically measures the share of AI Gateway token volume from models that publish their weights, classified as 'open,' versus all other models, categorized as 'closed' [1][2]. It is important to note that this chart does not represent a count of models, users, requests, revenue, spend, or the entire AI market [1][2]. The classification of models as open-weight adheres to Vercel's methodology, which can be broader than earlier reports, and past months may undergo revisions due to changes in this methodology [3][5]. The underlying data consists of anonymized, aggregate AI Gateway usage, aggregated daily, and encompasses input, output, reasoning, cached-input, and cache-creation tokens [3][2].
The chart has significant limitations regarding coverage and economic representation. It only reflects traffic routed through the AI Gateway, thereby excluding any activity not processed via Vercel [5]. Its data can be influenced by Vercel’s specific customer mix, routing choices, model availability, and the use of Bring Your Own Key (BYOK) [5]. For example, in August 2026, open-weight models accounted for 56% of AI Gateway tokens but only 14% of spend, highlighting why token-volume share should not be misinterpreted as economic market share [3].
Chart values are dynamic and lack full transparency for settlement purposes. The daily values presented in the chart are subject to revision, as exports or chart data are updated on a delayed operational schedule [3][2][4]. This means a value viewed on a particular day, such as September 30, 2026, might differ from the value subsequently used for settlement on October 1, 2026 [3][2][4]. Furthermore, the chart's denominator and the exact aggregation rules beyond the documented token categories are not fully specified [1][3][2].
Sources (5)
- 1AI model leaderboardsvercel.com
- 2https://vercel.com/docs/ai-gateway/leaderboardsvercel.com
- 3Open-weight models take 56% of token volume, Astra doubles Fable 5.1 spendvercel.com
- 4Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 5https://vercel.com/blog/ai-gateway-production-index-july-2026vercel.com
10. How might the implementation of the EU's Cyber Resilience Act and DORA in 2026 influence enterprise choices between open-source and closed-source AI providers?
| CRA & DORA governance impact date | September 30, 2026 [1][2][3][4][5][6][7] |
|---|---|
| DORA application date for EU financial entities | January 17, 2025 [6][8][9] |
| CRA mandatory reporting start date | September 11, 2026 [3][4][10][5][11] |
EU regulations will strengthen AI governance, leading to market segmentation. The European Union's Cyber Resilience Act (CRA) and the Digital Operational Resilience Act (DORA) are projected to modestly strengthen demand for enterprise-grade governance across both open and closed AI solutions by September 30, 2026 [1][2][3][4][5][6][7]. This regulatory impetus is anticipated to foster market segmentation, compelling different types of enterprises to select AI options that align best with their specific compliance and operational resilience requirements [3][4][5][6][7].
DORA will influence financial entities' AI provider choices. DORA has been applicable to EU financial entities since January 17, 2025, with its enforcement and operationalization set to take effect in 2026 [6][8][9]. It mandates that financial entities develop ICT third-party risk strategies for services critical to their operations and establishes EU-level oversight for essential ICT third-party providers [7][12][8][9]. While closed-source AI APIs might appeal to regulated financial enterprises due to clear vendor accountability, this could heighten DORA concerns regarding concentration risk and substitutability [6][7][9]. Conversely, open-weight or self-hosted AI can offer improved transparency and reduce dependence, thereby supporting DORA's exit and concentration-risk objectives, though it shifts greater responsibility for patching and vulnerability response to the enterprise [5][11][6][7].
CRA will reshape open-source AI procurement by 2026. By September 30, 2026, the CRA will be in its initial operational phase, with mandatory reporting having commenced on September 11, 2026 [3][4][10][5][11]. The CRA primarily covers commercially supplied open-source products, even when non-monetized free software is generally excluded from its commercial scope, with obligations attaching when software is commercially provided or supported [5][11][6][7]. This implementation is likely to favor enterprise-ready open-source providers over informal projects, as enterprises will prioritize maintainers who can document support, vulnerability handling, and accountability [3][4][5]. The likely procurement outcome by 2026 is market segmentation: non-financial enterprises may combine closed-source managed services with open models on governed platforms, while financial institutions will favor options offering the strongest evidence of resilience, accountability, reporting, and recoverability [3][4][5][6][7].
Sources (12)
- 1Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
- 2Open Source market share on Sep 28th - Kalshi noteringen | CoinRithmcoinrithm.com
- 3Commission publishes new guidance to support timely Cyber Resilience Act implementation | Shaping Europe’s digital futuredigital-strategy.ec.europa.eu
- 4Cyber Resilience Act - Reporting obligations | Shaping Europe’s digital futuredigital-strategy.ec.europa.eu
- 5Cyber Resilience Act - Open source | Shaping Europe’s digital futuredigital-strategy.ec.europa.eu
- 6-ec.europa.eu
- 7JC 2025 29 Guide on DORA oversight activitieseba.europa.eu
- 8Preparing for the DORA Oversight Framework: CTPP Designation and Next Steps, 14 May 2025esma.europa.eu
- 9Digital Operational Resilience Act (DORA) - European Insurance and Occupational Pensions Authorityeiopa.europa.eu
- 10Frequently Asked Questions | ENISAenisa.europa.eu
- 11Regulation (EU) 2024/2847 of the European Parliament and of the Council of 23 October 2024 on horizontal cybersecurity requirements for products with digital elements and amending Regulations (EU) No 168/2013 and (EU) No 2019/1020 and Directive (EU) 2020/1828 (Cyber Resilience Act) (Text with EEA relevance)eur-lex.europa.eu
- 12JC 2025 29 Guide on DORA oversight activitiesesma.europa.eu
11. What Could Change the Odds
Key Catalysts
Recent market activity indicates substantial momentum for open-weight token share, with a Kalshi market snapshot showing Above 67.5% at 97% for September 28 and Above 65% at 99% for September 18 [1][2]. Fundamental drivers supporting this trend include open models powering approximately one-third of tokens, a narrowing capability gap, lower prices, increasing sovereign and enterprise adoption, and rapidly rising Chinese open-weight distribution [3][4][5]. For example, Qwen recorded 942 million Hugging Face downloads, exceeding Llama's 476 million [5]. Free, open-source models capable of running on desktop computers represent a significant breakthrough [6].
However, significant bearish counterpoints persist in monetization and deployment. Closed providers reportedly captured about 96% of model-layer revenue, indicating open models face challenges translating usage into revenue [7][3][8][5]. Open models reach production less often than closed models by 12 percentage points, and their usage can concentrate in low-cost traffic [7][3][8][5]. This suggests that while usage share may rise, revenue could remain dominated by closed providers [7][3][4][8].
The relevant Kalshi market measures Vercel AI Gateway's Open vs. Closed Token Volume, with settlement based on the October 1, 2026 10:00 AM ET update for a September 30, 2026 measurement date [9]. Nearer-term catalysts include daily Vercel/OpenRouter traffic updates, new open-weight releases, default-model changes in agentic applications, and enterprise or sovereign deployments [7][3][8].
Key Dates & Catalysts
- Expiration: October 08, 2026
- Closes: October 01, 2026
Sources (9)
- 1Open Source market share on Sep 28th - Kalshi noteringen | CoinRithmcoinrithm.com
- 2Open Source market share on Sep 18th - Kalshi Odds | CoinRithmcoinrithm.com
- 3State of Open Source AI 2026stateofopensource.ai
- 4#198 — Open source AI in 2026: The founder briefing — hillock.hillock.studio
- 5Chinese Open-Weight AI Handles Most Developer Tokens: Closed Models Capture 96% of Revenuetechtimes.com
- 6Anthropic IPO at Risk, Meta’s Muse Pop, Token Prices Fall, Open Source Gains Share, Alignment Failsyoutube.com
- 7The Great Model Migration — CodeSOTAort.fabryka.ai
- 8The State of Open Source AI — v1.1 · September 2026stateofopensource.ai
- 9Open Source market share on Sep 30th - Kalshi Odds | CoinRithmcoinrithm.com
13. Historical Resolutions
Historical Resolutions: 20 markets in this series
Outcomes: 4 resolved YES, 16 resolved NO
Recent resolutions:
- KXOPENSOURCESHARE-26SEP29-T87.5: NO (Sep 29, 2026)
- KXOPENSOURCESHARE-26SEP29-T85: NO (Sep 29, 2026)
- KXOPENSOURCESHARE-26SEP29-T82.5: NO (Sep 29, 2026)
- KXOPENSOURCESHARE-26SEP29-T80: NO (Sep 29, 2026)
- KXOPENSOURCESHARE-26SEP29-T77.5: NO (Sep 29, 2026)
