AI Predicts Market Pattern Evolution: How Gensyn, Delphi, and Reppo Are Reshaping Data Verification Infrastructure

In the same week of April 2026, two pieces of news rolled out one after another in the crypto AI space. First: the decentralized AI computing network Gensyn, supported by a16z crypto, officially launched its flagship product Delphi on the mainnet—an AI-settled information market platform. Creators can independently create prediction markets and earn 1.5% of the market’s trading volume as revenue. Second: the Reppo Foundation, a decentralized AI training data protocol, announced a strategic capital commitment of $20 million from Bolts Capital to advance its prediction-market-driven AI training data infrastructure.

The two announcements landed almost simultaneously, pointing to the same track—the intersection of AI and prediction markets. But on closer inspection, the angles they take, the logic behind their construction, and the ecosystem niches they aim to capture are fundamentally different. This is not only a status update on two projects, but also a reflection of the structural split forming in this emerging field of AI data verification: one side is a verifiable information market aimed at humans, while the other is a training data verification network aimed at machines.

Against the backdrop that monthly trading volume in the prediction market industry has already broken through the multi-billion-dollar mark, while traditional platforms face intense regulatory pressure, the arrival of Delphi and Reppo may be opening up a new space for a track. And the appearance of top-tier investment institutions behind both projects further upgrades the industry weight of this contest.

From a16z’s heavy bet to dual-project resonance

On a macro level, the AI training data market is in a phase of rapid expansion. According to a Slator report, the global Data-for-AI market is estimated to be about $9.3 billion in 2026 and is expected to grow to $21.5 billion by 2031, with an approximate compound annual growth rate of 18%. Another industry report shows that the market size for AI training datasets is expected to increase from about $3.2 billion in 2025 to $16.32 billion in 2033.

Demand for data is shifting from “quantity” to “quality” and “verifiability.” Traditional data labeling models rely on centralized providers and suffer from structural shortcomings such as uneven quality, high costs, and insufficient incentives. The introduction of blockchain and prediction market mechanisms provides a well-matched alternative—by using economic incentives to make participants “bet” their capital on data quality, producing higher-quality and more trustworthy data signals.

On the investment side, a16z has been continuously heavily investing in the crypto AI track since 2023. In June 2023, a16z led a $43 million Series A round for Gensyn, with participating investors including CoinFund and Protocol Labs. In 2025, a16z completed 31 investments in the crypto sector, focusing on prediction markets, the integration of AI and cryptocurrencies, privacy blockchains, stablecoins, and other areas, and also made two investments in the prediction market platform Kalshi.

In its 2026 outlook, a16z clearly stated that prediction markets will become bigger, broader, and more complex. By the end of 2025, the combined trading volume of Polymarket and Kalshi had reached $28 billion. This figure indicates that prediction markets have evolved from niche experiments into a track with macro-scale reach.

In December 2025, Gensyn launched Delphi on the testnet, and then used the Sonar platform to open an AI token public offering, selling 300 million tokens with a valuation cap of $1 billion fully diluted valuation, matching the valuation of its Series A round led by a16z. In April 2026, within the same week, both projects achieved key milestones—Gensyn’s mainnet launch and Reppo’s large-scale financing. The timing coincidence reflects the synchronized warming up of the track.

Architecture breakdown: structural divergence in positioning, tokens, and capital

Although Delphi and Reppo both claim to be “prediction markets,” both involve AI, and both try to solve information verification problems, their underlying logic differs fundamentally.

Gensyn’s Delphi is positioned as an “information market”—allowing anyone to create prediction markets for any verifiable public event, with the outcome determined by AI models. Creators independently choose the AI model used for settlement, and the model’s weights are frozen at the time the market is created and cannot be changed. External participants can use Gensyn’s “reproducible execution environment” technology to rerun the model inference process and verify the authenticity of the settlement results.

Reppo’s positioning is more focused: it is not a human-facing “event betting” platform, but an infrastructure for AI developers’ training data verification. Reppo builds a dedicated “data network” that turns human judgment into verifiable on-chain signals for AI model training. Its “events” are not election results or sports scores, but instead whether the labeling quality of a particular dataset is up to standard, or whether a segment of data can improve model performance.

The core difference between the two can be interpreted through the following framework:

In terms of the economic model, Delphi is built around its native AI token. The protocol charges a 0.5% fee on all trading volume, which is used to buy back AI tokens. 70% of protocol revenue is permanently removed from circulation via a buy-and-burn mechanism, 29% goes into the community treasury, and 1% is used for treasury executor rewards. Market creators earn 1.5% of trading volume as revenue, paid in stablecoins.

Reppo, by contrast, is centered on the REPPO token, with incentive mechanisms focused more on the accuracy of data verification than on trading volume. Participants earn incentives by predicting whether a dataset helps improve AI model performance; if their predictions match the actual results, they receive rewards. Economically, this design suppresses low-quality data submissions.

On the funding side, Gensyn’s total funding across three rounds exceeds $50 million, and the A round led by a16z provides top-tier credit enhancement. Reppo’s $20 million strategic capital commitment comes from Bolts Capital, which previously received support from Protocol Labs and others. Notably, a16z is also an investor in Kalshi, indicating that its positioning in this track is not a single bet.

Industry competition under the “information market” label

Gensyn has explicitly stated that its strategy is not to compete for the same markets as Polymarket or Kalshi, but to “open a brand-new niche market category owned by creators.” This narrative seeks to separate Delphi from traditional prediction markets, especially in a context where the U.S. faces strict regulatory scrutiny.

Reppo’s narrative focuses on “solving AI data bottlenecks,” emphasizing that the total addressable market for prediction markets is expected to reach $1 trillion in annual trading volume by the end of the decade. The market will go beyond sports and world events, extending into information and viewpoints.

Industry observers remain cautious. Edgen.tech points out that Delphi’s launch comes at a time when prediction markets are under regulatory pressure, and its AI settlement model may offer new ideas. a16z science advisor Andy Hall also emphasized that the key to future development lies not only in increasing the number of contracts, but in improving “methods for determining the truth”—centralized arbitration mechanisms are increasingly unable to meet the demands of scaling.

Can AI settlement truly be decentralized? Gensyn’s REE technology enables model inference to be verified externally, but issues such as model bias, the immutable nature of fixed weights, and the authority behind model selection remain potential points of controversy. For Reppo, verifying the security and reliability of decentralized networks also faces challenges—security vulnerabilities that persist in the DeFi space continue to hinder institutional investment, and the $292 million loss suffered by KelpDAO after a hacker attack serves as a warning.

Structural impact: the AI data value chain is being redefined

The synchronized advancement of Delphi and Reppo signals that “AI data verification” as an independent track is taking shape. Both approaches enter the same field from different directions and together form the foundational layer of infrastructure for decentralized data verification.

The economic basis of this track is that the stronger the AI model, the more rigid the demand for high-quality, verifiable data becomes. Traditional data labeling industries make “cost” the core dimension of competition, while decentralized verification mechanisms shift the competitive dimension to “trustworthiness”—using economic incentives to have verifiers back quality with their own capital. This shift could potentially reshape the value-distribution structure across the AI training data industry chain.

For the prediction market industry, traditional platforms focus on “events.” Delphi and Reppo expand the boundaries of “predictable events”: Delphi brings “any solvable question” into the market scope, while Reppo treats “data quality” itself as the prediction target. This expansion creates a new type of market rather than simply competing for existing share. The “broader and more complex” prediction market vision projected by a16z is being validated by these two practices.

The ripple effects on the crypto AI ecosystem are also worth noting: capital is accelerating into the data verification track; traditional AI data labeling faces structural competitive pressure; and the narrative of “data assetization” is deepening further.

Conclusion

In the week of April 2026, the AI prediction market track was illuminated by two pieces of news at the same time. One revealed a new paradigm for information market trading, and the other demonstrated a path for prediction market mechanisms to penetrate upstream of the AI training data value chain.

Both are redefining “trusted data” using economic incentives and crypto mechanisms—one aimed at human information consumption needs, the other aimed at machine data production needs. This coin reflects a broader judgment: AI systems’ reliance on high-quality data is continuously deepening, and data verification is becoming a foundational infrastructure layer of the AI economy.

As of April 24, 2026, Delphi has moved from the testnet to the mainnet, and Reppo has completed a new round of financing. Both projects are at a critical stage transitioning from proof of concept to large-scale operations. The next tests will be real user retention on the mainnet, building trust in AI settlement mechanisms, and finding sustainable compliance paths amid regulatory uncertainty.

Prediction markets predict everything—except their own fate. But what can be confirmed is that the AI data verification track has evolved from a vague concept into an industry direction backed by capital, technology, and product support.