Data & Insights: Crypto AI Is Back? Monad too.

Moving Volume

• Bitcoin spot ETF combined volume has fluctuated significantly throughout 2026. Volumes have pulled back considerably from an early 2026 high, with the 7-day moving average currently sitting around $2.8B, well below the peak of nearly $14B observed in January.
  • The decline tracks with Bitcoin's broader price consolidation, which has seen BTC range-bound between $65K and $70K since February. Eyes remain on Bitcoin as the market leader, and participants appear to be waiting for a directional move before re-engaging.
• Spot Bitcoin ETFs have become one of the most closely watched barometers of both retail and institutional interest in crypto, given how accessible they have made Bitcoin exposure through traditional brokerage accounts. When volumes compress like this, it tends to reflect a cooling in near-term conviction rather than a structural shift in demand.
  • Notably, the infrastructure buildout around these products continues to accelerate regardless of sentiment. Morgan Stanley recently announced it is launching a spot Bitcoin ETF with a fee of just 0.14%, undercutting all existing rivals. This is likely among the first signs of fee compression across issuers as traditional asset managers compete for dominance in a product category that barely existed two years ago.
  • Fee wars at this stage are a signal worth paying attention to. They suggest that major institutions view spot Bitcoin ETFs as a long-term product line worth subsidizing today to capture market share, rather than a cyclical offering tied to near-term price action.
• The current lull in volume may prove temporary, but the competitive dynamics among issuers point to a market that is maturing structurally even during periods of reduced trading activity.

Hyperscaler

• Centralized exchanges have long been the crown jewel of the crypto industry, processing over $3T in monthly perpetual futures volume and providing the core use case that underpins the sector: trading and speculation.
  • Hyperliquid is steadily chipping away at that dominance. The platform's share of total perpetual futures volume has climbed to just under 6% in March, up from roughly 3.5% a year ago, with monthly volumes approaching $200B.
• What makes the trend particularly notable is that the ratio has continued to climb even as overall exchange volumes have compressed from their August 2025 peak. This suggests Hyperliquid is genuinely pulling market share rather than simply riding broader volume.
  • Onchain competition remains limited. While platforms like dYdX and GMX exist, neither has matched Hyperliquid's trajectory in terms of volume growth or product expansion, leaving it as the clear frontrunner in decentralized perpetual futures.
• The platform's expansion into non-crypto assets is an increasingly relevant part of the story. Commodities like oil now trade 24/7 on Hyperliquid, and non-crypto volume is making up a growing share of overall activity on the platform.
  • This points to a structural advantage that decentralized venues hold over traditional markets. A trading firm that waits for CME to open Sunday evening to hedge an oil position is carrying weekend gap risk that a 24/7 venue eliminates entirely. 
• If decentralized perps platforms can continue scaling liquidity and asset coverage, the addressable market extends well beyond crypto-native volume and into the multi-trillion dollar universe of traditional derivatives, where settlement delays and market hours remain structural inefficiencies.

Predicting More Raises

• Prediction markets have emerged as one of the most heavily funded verticals within DeFi, accounting for a significant portion of the category's venture capital inflows over the past year.
  • Cumulative venture funding into prediction markets totaled roughly $3.4B in 2025 alone, with Polymarket and Kalshi capturing the bulk of that capital. For context, the sector raised just $80M across all of 2024 and barely registered prior to 2021.
  • That trajectory has carried into 2026, with approximately $1.1B deployed into prediction market ventures year-to-date, suggesting investor appetite has not meaningfully cooled despite the rapid scaling of commitments.
• The latest deal came from Intercontinental Exchange, the parent company of the New York Stock Exchange, which doubled down on Polymarket with a reported $600M in new cash investment and up to $40M to buy out existing holders. This follows a prior $1B investment, bringing ICE's total commitment to roughly $1.6B.
  • ICE's continued allocation at this scale is notable given its position as one of the most established players in traditional financial infrastructure. The size of the commitment suggests a long-term thesis on prediction markets as a distinct asset class rather than a speculative cycle play.
• Meanwhile, Polymarket continues to hold the top position in crypto mindshare, according to Kaito, maintaining its lead over competitors despite growing interest in the space from both crypto-native and traditional finance participants.
  • The broader bull case for the sector extends well beyond trading. Proponents argue that liquid, incentive-aligned prediction markets can serve as real-time information discovery tools, producing more accurate probability estimates on geopolitical events, economic outcomes, and policy decisions than traditional polling or expert forecasting.

GMonad

• Total value locked in the Monad blockchain currently stands at ~$338 million, up ~70% since the beginning of February.
  • This figure makes Monad the fastest layer 1 blockchain to reach $300 million TVL in the last few years, crossing the threshold roughly four months after its November 2025 mainnet launch
  • Despite the recent growth, Monad still accounts for less than 0.3% of the ~$90 billion total TVL tracked across all chains, leaving significant room for share gains.
• However, while the speed is relatively impressive, the economics behind it are not, at least for now.
  • Monad’s on-chain fee generation currently averages under $3,000 a day.
  • That means $334 million in capital is parked on a chain producing annual revenue in the low six figures, making Monad’s fee-to-TVL ratio among the lowest of any chain with meaningful TVL.
  • While prominent DeFi protocols such as Uniswap, Curve and Morpho are present in the network, the presence of protocols does not guarantee the presence of actual users.
  • If fee revenue doesn’t scale with TVL, Monad risks joining the growing list of L1s and L2s where capital sits idle chasing token incentives rather than productive yield.
• Monad’s token, MON, has likely priced in this skepticism, as its currently valued at a fully diluted valuation (FDV) of ~$2.2 billion
  MON is currently down ~50% from its post-launch peak of $4.7 billion FDV from over four months ago.

Dreaming of Decentralized Compute

• The GMCI AI Index recorded a value of 51.26 on Sunday, up 48% since the beginning of February.
  • However, it is worth noting that the index remains 84% below its all-time high from Q1 2024.
  • The GMAI index comprises 9 tokens with extreme top-heaviness, with Bittensor (TAO), Render (RNDR) and Artificial Superintelligence Alliance (ASI) making up over 71% of the index
  • Despite being called a crypto AI index, it is more so a measure of the performance of three large-cap AI infrastructure tokens rather than broader AI crypto sentiment.
  • This composition matters for interpreting the March rally, as the price of TAO nearly doubled in March, and combined with its 24.89% weight, contributed the majority of the index’s appreciation.
• The reason for the rise in the price of TAO is the market catching up to a fundamental realization of what the Bittensor network can demonstrably do.
  • Bittensor subnet 3’s team announced the Covenant-72B model,  a 72-billion-parameter language model trained permissionlessly across more than 70 globally distributed nodes.
  • The model scored 67.1 on MMLU, a popular benchmark for evaluating the capabilities of LLMs, putting it in a competitive range with Meta's Llama 2 70B.
  • This showcased the first credible proof that distributed training, which has often been dismissed as too slow and fragmented, can produce models that compete with centralized incumbents.
  • The top subnet token, τemplar (SN3), rallied by over 400% in the last month to a ~$130 million market cap
  • This development in the Bittensor ecosystem gained the attention and acknowledgement of notable public figures such as Jensen Huang and Chamath Palihapitiya, which brought further attention and narrative fuel to the ecosystem and, in turn, the TAO token.
  • The subnet ecosystem beyond Covenant-72B is also maturing, with Targon (SN4), a decentralized GPU compute marketplace operated by Manifold Labs, securing a six-figure deal to power Dippy AI's backend of 8.6 million users.
    At the time of writing, there have been over 10.7 million TAO tokens issued, with over 68% staked.
    The current price of TAO stands at ~$310, with a market cap of just over $3 billion.