Leading Crypto VCs See Collective AUM Shrink: a16z Crypto Fund Management Scale Plummets 40%, Multicoin Halved

marsbitPublicado a 2026-04-18Actualizado a 2026-04-18

Resumen

Headline: Top crypto funds including a16z Crypto and Multicoin saw their Assets Under Management (AUM) decline significantly in 2025, according to undisclosed SEC filings obtained by Fortune. Key Points: - a16z Crypto’s AUM dropped nearly 40% to $9.5 billion, partly due to returning capital to LPs at a market peak. Its first crypto fund achieved a 5.4x DPI. - Multicoin’s AUM was halved to approximately $2.7 billion, impacted by the crypto market downturn and its hybrid hedge/VC fund structure. - Pantera Capital also reduced AUM, influenced by successful exits including five portfolio company IPOs like Circle and BitGo. - Haun Ventures was the exception, with AUM growing over 30% to nearly $2.5 billion, driven by the successful acquisition of portfolio company BVNK by Mastercard. - Despite AUM contractions, major firms are raising new funds: Paradigm targeting up to $1.5 billion, a16z Crypto seeking up to $2 billion, and Dragonfly closing a $650 million fund. The data highlights the high volatility and cyclical nature of crypto VC investments, where AUM fluctuations are heavily tied to token prices and market cycles rather than solely reflecting fund performance.

Original Author: Ben Weiss

Original Compilation: Deep Tide TechFlow

Guide: Fortune reporters obtained a batch of previously undisclosed financial disclosure documents from the SEC for crypto VCs. The data shows that the Assets Under Management (AUM) of top firms like Paradigm, Pantera, a16z crypto, and Multicoin shrank across the board in 2025. However, the shrinkage isn't all bad news—a16z crypto returned money to LPs at the market peak, with its first fund's DPI reaching 5.4x. The only firm that bucked the trend was Haun Ventures, which capitalized on the stablecoin sector through BVNK's acquisition by Mastercard.

The leading players in crypto VC couldn't escape the market crash of 2025.

Fortune reporter Ben Weiss obtained a set of previously undisclosed investment advisor financial disclosure documents from the U.S. Securities and Exchange Commission (SEC). The data is straightforward: the AUM of top firms like Paradigm and Pantera Capital collectively shrunk in 2025.

But the 2025 downturn pushed it back down. From 2024 to 2025, Multicoin's AUM was more than halved, dropping to approximately $2.7 billion. Since BTC began its plunge in October 2025, crypto assets have retreated across the board, and a structure like Multicoin's, which operates both hedge funds and VC funds, was hit particularly hard.

Additional context: Multicoin co-founder Kyle Samani left the company in February of this year to focus on other areas of technology investment.

Pantera: Five Portfolio Companies IPO, Capital Returned to LPs

Pantera Capital's AUM also shrank, but similar to a16z, part of the reason was主动 distributions back to LPs.

This institution, founded by former a16z crypto partner Katie Haun, saw its AUM increase by over 30% year-over-year, approaching $2.5 billion. On one hand, it was due to betting on the right sector—its investment in the stablecoin company BVNK was acquired by Mastercard for up to $1.8 billion. On the other hand, Haun Ventures itself also raised a new $1 billion fund in 2025.

Although AUM has shrunk, leading institutions have not stopped:

Paradigm is raising a new fund of up to $1.5 billion. a16z crypto is raising up to $2 billion. Dragonfly just closed its fourth fund of $650 million. Post-publication correction from Fortune: A Dragonfly spokesperson actually responded, confirming the data is "accurate" and stating "we are actively deploying capital".

Spokespersons for Paradigm, Pantera, a16z crypto, Multicoin, and Haun Ventures all declined to comment.

The Cyclical Destiny of Crypto VCs

The original article ends here, but a few background points are worth adding.

Crypto VCs are fundamentally different from traditional tech VCs. Traditional VCs invest in equity, with exits via IPO or M&A. Many crypto startups have their own tokens, exposing the VC's holdings directly to token price volatility.

According to Pantera Capital's outlook report earlier this year, the total crypto market cap excluding BTC (and excluding ETH and stablecoins) fell by about 44% from its late 2024 high. But following historical patterns, bear markets are also windows of opportunity for buying the dip. The密集 fund-raising by several leading institutions at this moment is a bet on the next cycle.

According to a previous Fortune exclusive, a16z crypto's fifth fund is planned to complete fundraising in the first half of 2026, led by Chris Dixon, and will continue to fully bet on the blockchain direction. Paradigm's new fund, according to The Wall Street Journal, will expand into AI and robotics technology. The divergence in the two strategies is clear: a16z continues all-in on crypto, while Paradigm chooses to hedge across sectors.

Preguntas relacionadas

QAccording to the SEC filings obtained by Fortune, which major crypto VC firms experienced a significant decline in AUM in 2025?

AAccording to the SEC filings, major crypto VC firms including Paradigm, Pantera Capital, a16z crypto, and Multicoin all experienced significant declines in their Assets Under Management (AUM) in 2025.

QWhy did a16z crypto's AUM drop by nearly 40%, and what positive outcome did this represent?

Aa16z crypto's AUM dropped by nearly 40% partly because the firm began distributing returns from its first three funds back to its Limited Partners (LPs), intentionally doing so at a market high in 2025. The positive outcome was that its first crypto fund achieved a net DPI (Distributed to Paid-In capital) of 5.4x, representing strong returns for its investors.

QWhat was the primary reason for Haun Ventures being the only firm to see an increase in AUM, and by how much did it grow?

AHaun Ventures was the only firm to see an increase in AUM, growing by over 30% to nearly $2.5 billion. This growth was primarily due to a successful investment in the stablecoin company BVNK, which was acquired by Mastercard for up to $1.8 billion, and the firm's own fundraising of a new $1 billion fund in 2025.

QHow does the article explain the fundamental difference between crypto VC and traditional tech VC that leads to such volatile AUM changes?

AThe article explains that the fundamental difference is that traditional VCs invest in equity and exit via IPO or acquisition, while many crypto startups have their own tokens. This means a crypto VC's holdings are directly exposed to the extreme volatility of token prices, leading to massive swings in AUM that are unimaginable in traditional venture capital.

QDespite the AUM declines, what actions are firms like Paradigm and a16z crypto taking, indicating their continued belief in the market?

ADespite the AUM declines, firms like Paradigm and a16z crypto are actively raising new funds, indicating their continued belief in the market. Paradigm is raising a new fund of up to $1.5 billion, and a16z crypto is raising a new fund of up to $2 billion.

Lecturas Relacionadas

Upbit Rebalances 864B SHIB In Internal Wallet Move

South Korean cryptocurrency exchange Upbit moved a total of 864 billion SHIB (worth approximately $4 million) between its internal wallet addresses. The activity, involving transfers of 384 billion SHIB outbound and 480 billion SHIB inbound, occurred following a 36% SHIB price rally. On-chain data indicates this was a routine internal wallet rebalancing operation by the exchange, not a sell-off or liquidation event. The article cautions that while large on-chain transfers are highly visible, they require proper context and wallet labeling to interpret correctly. In this case, the movement between known Upbit addresses does not inherently signal a change in market supply or selling pressure for SHIB.

bitcoinistHace 10 min(s)

Upbit Rebalances 864B SHIB In Internal Wallet Move

bitcoinistHace 10 min(s)

Can Generative Models Finally Be Trained End-to-End? The Core Is a For Loop

This article introduces a novel training paradigm for generative models called Explorative Modeling (XM), which enables true end-to-end training. Traditionally, powerful generative models like autoregressive and diffusion models are not trained end-to-end. They are trained to predict a single small step but require iterative multi-step sampling for inference. This "exposure bias" leads to error accumulation and limits performance. The core challenge XM addresses is "mode blurring." In generative tasks, a single input (e.g., "generate a dog") corresponds to many valid outputs (multiple modes). Standard training objectives like reconstruction loss force the model to average these modes, producing unrealistic, blurry outputs. To avoid this, existing models break generation into many small, almost deterministic steps, sacrificing end-to-end training. XM tackles this by restructuring the training loop itself. Its key insight is to amplify "generative expressivity." For each training input, instead of generating one sample, the model generates K candidate outputs. Only the candidate closest to the real data is used for computing the loss and updating the model via backpropagation. This simple "best-of-K" mechanism is implemented as a short for-loop. By exploring multiple possibilities, the model learns to distribute its guesses across different modes rather than collapsing to their uninformative average. The paper demonstrates that "exploration" acts as a new, powerful scaling axis. Gains from XM increase with model size, data scale, and compute. Experiments show improvements in FID scores for image generation and significant efficiency gains, sometimes outperforming larger models without exploration. When pushed to the limit, XM enables fully single-step, end-to-end generative models. In robotics tasks, an "Explorative Policy" matched the performance of a 100-step Diffusion Policy with a single forward pass, drastically improving inference speed. While the best-of-K concept is not entirely new, the authors' contribution lies in formally understanding it as a direct method to boost generative expressivity without fragmenting the generation process. This work suggests that as models scale, enhancing exploration during training may become crucial for overcoming fundamental performance bottlenecks.

marsbitHace 10 min(s)

Can Generative Models Finally Be Trained End-to-End? The Core Is a For Loop

marsbitHace 10 min(s)

A Maximum of 20 Submissions Per Person? DeepMind Researcher Satirically 'Petitions' Against ICLR's New Rule

The International Conference on Learning Representations (ICLR) 2027 has introduced a new submission rule limiting authors to a maximum of 20 papers per conference. This policy, announced due to a 68% surge in submissions for ICLR 2026 that strained the review system, aims to ensure fair and informative peer review. Google DeepMind researcher Dan Roy responded with a sarcastic petition on X, calling the 20-paper cap "absurd" and claiming it would slow AI progress. His post mockingly argued that AI agents can now produce incremental research and that large language models (LLMs) already handle all reviewing, so output should be maximized. This satire critiques the broader trend of AI's role in academia, referencing a 2025 study that found about 21% of ICLR 2026 reviews were likely fully AI-generated. Roy's underlying point questions whether simply capping submissions addresses the systemic issue of AI potentially flooding conferences with agent-written papers and AI-assisted reviews. The debate highlights tensions between managing submission volume and maintaining research quality as AI tools become pervasive.

marsbitHace 14 min(s)

A Maximum of 20 Submissions Per Person? DeepMind Researcher Satirically 'Petitions' Against ICLR's New Rule

marsbitHace 14 min(s)

Coldcard Vulnerability Leads to $89 Million Theft, Sparking Largest On-Chain Migration Since FTX Collapse

A security vulnerability in Coldcard hardware wallets led to the theft of 1,367.05 BTC (approx. $89 million) from 4,585 addresses, according to Galaxy Research. The flaw, caused by insufficient randomness in seed phrase generation for specific firmware versions, triggered a massive migration of funds as users moved assets to new, secure wallets. This event caused the largest on-chain movement of small Bitcoin transactions since the FTX collapse, distorting key market indicators like active addresses and exchange deposits, and significantly dampening market sentiment. The incident also highlighted challenges in cybersecurity investigations, as researchers reported that safety guardrails on mainstream U.S. AI models hindered the analysis of attack patterns, forcing them to use an open-source Chinese AI model for forensic work. This underscores a dilemma in balancing AI security restrictions with the need for rapid response in tracing stolen crypto assets.

marsbitHace 1 hora(s)

Coldcard Vulnerability Leads to $89 Million Theft, Sparking Largest On-Chain Migration Since FTX Collapse

marsbitHace 1 hora(s)

Annual Salary of Millions Competing for Electricians, Meta Rushes to Open Its Own Technical School

The AI boom is facing an unexpected bottleneck: a severe shortage of skilled construction workers and electricians. As tech giants like Meta, OpenAI, and Alphabet race to build massive data centers—such as OpenAI's $16 billion "Stargate" project—they are hitting a critical labor wall. The U.S. needs an estimated 130,000 more electricians, 240,000 construction workers, and 150,000 supervisors by 2030 for AI infrastructure alone, but tens of thousands of electrician jobs go unfilled each year. While AI companies offer high premiums, with electricians earning up to $280,000 annually, worker scarcity still causes massive losses—delays on a single project can cost $14.2 million per month. The complexity of building AI data centers, which require immense power (equivalent to powering hundreds of thousands of homes), sophisticated electrical systems, and advanced liquid cooling solutions, demands highly skilled technicians who are in short supply. To combat this, companies are investing heavily in training. Meta has committed $115 million to a free training school offering tuition, housing, and stipends, targeting 5,000 new workers. OpenAI is partnering with unions to secure skilled labor. These efforts are paying off, with a significant rise in Gen Z interest in trade schools over college. However, the power demands are staggering. AI data centers are driving a rapid surge in electricity consumption, projected to account for up to 12% of U.S. power use by 2028 and raising costs for consumers. Furthermore, the construction boom is project-based, leading to a potential future glut of trained workers once building peaks, which could depress wages industry-wide. The race for AI supremacy now depends as much on skilled hands as on advanced chips.

marsbitHace 1 hora(s)

Annual Salary of Millions Competing for Electricians, Meta Rushes to Open Its Own Technical School

marsbitHace 1 hora(s)

Trading

Spot