Dragonfly · 806

TOKEN2049 Singapore will host October 7-8

Comparatively, the well-known crypto event TOKEN2049 will be held from October 7 to 8 at Marina Bay Sands in Singapore. It is expected to attract 25,000 participants from over 7,000 companies in 160 countries, with 500 booths, 300 speakers, and more than 1,000 surrounding events. Alex Fiskum, co-founder of TOKEN2049, said that the Dubai event was postponed to 2027 this year, and the team will fully focus on Singapore. Additionally, the 36-hour Hackathon TOKEN2049 Origins and NEXUS Startup Competition will be held during the event week, and well-known venture capitalists such as Dragonfly, Multicoin, and Maelstrom will be judged.          

16d ago

AI proxy infrastructure Sapiom closes $35 million Series A funding round led by Dragonfly

Comparatively, according to Business Wire, AI proxy infrastructure platform Sapiom completed a Series A round of funding led by Dragonfly, with Accel, Gradient, Coinbase Ventures, Operator Collective, Formus Capital, and VanEck Ventures. Existing investors Okta Ventures and Menlo Ventures, Anthropic, and Array Ventures continued to follow suit. The company was founded just 11 months after completing a $15 million seed round with a cumulative financing of $50 million. Sapiom aims to remove infrastructure barriers for AI agents from presentation to production environments. Sapiom simultaneously launched three products: Router, Agent Studio, and Runtime. Dragonfly managing partner Haseeb Qureshi will join Sapiom's board of directors.

16d ago#financing

Coldcard vulnerability investigation upgraded: at least 15 attackers discovered, single victim lead finds 12 BTC stolen

In comparison, Alex Thorn, head of research at Galaxy Digital, said that according to new victim reports received after the incident, the number of Coldcard vulnerability attackers had reached at least 15. According to Thorn, the information provided by the victims helped the research team uncover previously unidentified attacks. Since this vulnerability attack method is different from the centralized exchange theft incident, the connection between attackers needs to be confirmed by on-chain analysis and victim feedback. He said that just one victim reported less than 1 BTC stolen helped the team uncover a previously unknown attack that stole around 12 BTC from 126 addresses. According to Galaxy Research's previous estimates, the Coldcard vulnerability has led to at least three rounds of attacks, which cost about $100 million in BTC; in addition, Galaxy also discovered a suspected fourth round of attacks, which could raise the total losses to about $130 million. At the same time, the incident also sparked discussions about the security of Bitcoin's self-custody. Dragonfly managing partner Haseeb Qureshi said that the “$2 cost AI security hardening” may be able to avoid this vulnerability, adding that some AI models can rediscover related vulnerabilities in a relatively short period of time. However, industry insiders pointed out that current claims about how quickly AI finds vulnerabilities lack strict blind testing and verification. Researchers believe that as AI model capabilities improve, vulnerability detection and attack costs in the cryptographic industry may continue to decline, and wallet developers need to further strengthen code audits and security protection.

18d ago

Coldcard vulnerability AI discovery costs just $2, Dragonfly partner says security hardening must be invested

Comparing news, Dragonfly managing partner Haseeb Qureshi wrote on the X platform that the key in the security field has become a cost competition, that is, a comparison between the cost of attackers using cutting-edge AI to scan for vulnerabilities and the cost of the attacked party investing in AI security hardening. Claude discovered the vulnerability within 8 minutes, and GLM reproduced it within 20 minutes, indicating that Coldcard invested almost zero in AI security hardening. Qureshi estimates that GLM 5.2 would only cost around $2 to reproduce the attack, meaning that it would only cost $2 to discover the vulnerability with AI. He warned that it would be difficult for small companies to survive in the field of security-sensitive products, and said any security-sensitive product should undergo an AI security scan every time it is released.

18d ago

Dragonfly Partner: Today's AI boom might be difficult to emerge without the development of the crypto industry

Comparing news, Dragonfly partner Haseeb posted an article on the X platform saying that without the development of the crypto industry, today's AI boom might be difficult to emerge. He said that Bitcoin miners have accumulated experience in power acquisition, site selection, and large-scale infrastructure construction in remote regions over the past 10 years, providing an important foundation for the current expansion of AI data centers; at the same time, crypto mining requirements have also helped Nvidia complete the transformation from a gaming GPU manufacturer to an AI computing chip supplier. Furthermore, the core teams of many emerging AI cloud service providers such as CoreWeave, Crusoe, Together AI, and Nscale all have a background in the crypto industry.

23d ago

How can crypto VCs find their next opportunity? Haseeb says “some things never come back”

Author: Wu Says Blockchain Original title: Haseeb on Crypto VC: Sorry, Some Things Will Never Come Back In an interview with MAD Society on July 15, 2026, Dragonfly managing partner Haseeb Qureshi discussed crypto venture capital, founder judgments, and long-term trends in the industry. He believes that the key to venture capital is to seize a few non-consensus opportunities. Excellent founders should have outstanding “peak ability,” but lack of integrity and inconsistent words and actions are clear danger signs. Haseeb also said that it is difficult to form long-term enterprises in the direction of some structured products and the tokenization of individual assets, while the DeFi, stablecoin, payment and prediction markets will continue to exist; in the long run, cryptographic technology will eventually be incorporated into various financial and technology products, and the “crypto company” label may gradually disappear. The audio transcription was done by GPT, there may be errors, please watch the original video at YT. Poker and venture capital: How to establish judgment discipline in a long feedback cycle Haseeb Qureshi: There really isn't much compatibility between poker and venture capital. Poker is very similar to trading because they all have very fast feedback loops that can be iterated very closely and quickly. As soon as you play a hand, you'll know whether you won or lost, and whether your decision was right. But in venture capital, the feedback cycle is very slow. If you invest in a founder, it may take many years before you know if your original judgment was correct. In the first year, you may see some initial signs, such as the company is growing and seems to be starting to gain some market recognition. Even if a company has completed Series A or even Series B financing, it can still suddenly go awry. It may have looked like it was going well for several years, but the founders had a fatal flaw that eventually led them to lose the ball in their final offense in the final game of the season. So the reality is, it's hard to quickly judge whether you're doing a good enough job as a venture capitalist. Many funds raised funds by relying on the early book valuation of their portfolios, but it was only discovered in the end that there were no real winners in the entire portfolio. Let's say you invested in Axie Infinity or OpenSea early on, and you probably thought, “Wow, I'm an amazing investor, I did such a great job.” There are also several funds that have invested in FTX in the early stages. At the time, people would say, “My God, this guy is simply the son of choice in the investment world. Can you believe he participated in the FTX seed round?” But just a few years later, the situation became: “OK, this fund doesn't seem to be anything special now.” Because its brightest star project has already exploded. Venture capital is unique in this regard. This means, first, you must take the initiative to establish a feedback mechanism for yourself, rather than expect the world to give you direct feedback. Because as a venture capitalist, you have to keep learning and improving, but it often takes many years to know whether an investment is successful or not. Therefore, feedback must come more from your judgments about your own performance rather than from external results. For a lot of people, this is very difficult. Another difference between venture capital and poker is that venture capital is a team sport, while poker is a single player game. Of course you're playing cards with other people, but essentially you're facing the entire table alone. That's not the case with venture capital. You can only be successful if the founder you invest in is successful; you can only really win if your fund is successful and the projects carried out by the other partners in the fund are also successful. As a result, venture capital relies heavily on collaboration and interpersonal relationships. But if you're a poker player, you hardly need to care about anyone else in the world. As long as you sit at the table, play properly, and continue to make a profit, you can still be a successful poker player even if you don't have any friends. This is also a very different point between the two. Most really good venture capitalists are really good at dealing with relationships. I don't think I'm particularly good at this, but I'm definitely a lot better than the past and better at building relationships than most traders I know. Most traders don't need that. Just like poker players, they don't need to be friendly, be good at handling relationships, and don't need to have a large network of people. Therefore, the ability to really help you make good venture investments in poker is mainly the ability to think clearly about risk and the ability to control emotions well. I found that a lot of venture capitalists aren't really good at this. They can be very emotional, and it's hard to handle conflict. These two aspects are just right...

26d agoburnking

Is crypto venture capital dying out?

Source: Token Dispatch Author: Vaidik Mandloi Compiled and collated by: BitPushNews created an era's top investor and began leaving it as one of the largest cryptocurrency exclusive funds ever formed. Paradigm recently raised $1.2 billion to begin investing in startups in the fields of artificial intelligence (AI), robotics, and aerospace. They've even completely removed the word “crypto” (crypto) from their website! Their investment logic is: Cryptocurrency was only their first frontier, but there are so many other new things happening right now that they must not turn a blind eye. Coincidentally, Framework Ventures also closed a $400 million fund in June and began expanding their investment reach beyond the crypto sector, and they are no exception. Over the past year, almost every leading crypto specialty fund has begun to drift towards broader topics and investment licensing. In the first quarter of 2026, only 8 new crypto-specific venture capital funds were established globally, the lowest since 2020. This article will explore in depth whether crypto-specialty venture capital is actually dying out as a fund category. If so, how does this shuffle map into the life cycle of these funds, and what does it mean for crypto startups — they will now have to compete for attention in multi-industry portfolios. The life cycle of professional funds Crypto professional funds came into being because they were willing to take the time to build a competitive advantage and were the only ones willing to take and underwrite this risk at the time. Understanding how Solidity contracts actually work and connecting with anonymous developers on the Discord channel—these aren't things Tiger Global's growth equity partners were able to touch in 2017. To understand whether crypto VC is coming to an end as an investment category, it would be beneficial to see how the specialty fund category has evolved in history, as this phenomenon has happened more than once in the past. Between 2006 and 2011, Climate Tech (Climate Tech) became mainstream as an investment logic. VCs have set up clean energy exclusive funds for the same reason that crypto VCs set up exclusive blockchain funds: they think they have keenly captured an epoch-making technological shift before generalists (generalists) reacted, and wanted to build a new investment institution around this firm belief. They poured more than $250 billion into clean energy startups, but lost more than half of their capital. Interestingly, the technology itself actually worked, and today's clean energy market is extremely large — which has caused the cost of solar energy in this sector to drop dramatically by 85% over the same period. What the VCs misunderstood, however, was that they hardwired the same model applied to software companies and threw $5 million in seed round checks to companies that actually needed $200 million in project financing and took 15 years to make a profit. The Energy Initiative (Energy Initiative) of the Massachusetts Institute of Technology (MIT) conducted an ex post facto review and found that the venture capital model was fundamentally flawed in the field. Professional VCs completed the experimental phase by taking technical risk funding, funded early R&D, and gave the field credibility to attract larger capital; however, once the technology matured enough to allow infrastructure lenders and project finance facilities (project finance facilities) to underwrite, the information advantage of professional investors disappeared. Data source: MIT Energy InitiativeSpac (a special absorption merger and acquisition company) has also evolved a similar trajectory. To add background, SPAC is a “blank check company” with no actual business, raising capital through an IPO and then merging with a private company to help it go public faster than a traditional IPO. In 2020 and 2021, some investors saw it as a replicable vehicle and built entire companies around them. Chamath Palihapitiya raised $1.6 billion in SPAC exclusive capital. But by 2022, SP established in 2021...

32d agoWendy#AI #Framework Ventures #Paradigm #VC #Investment funds

Dragonfly partner: Meme coin trading is mainly for retail investors and should not show the characteristics of weekend lows in the professional trading market

Comparing the news, Dragonfly partner Haseeb posted an article on the X platform saying that in response to questions about the long-term stability of Pump.fun's revenue curve, he believes that this phenomenon is not abnormal; from a perspective, it is quite similar to the overall trend of crypto spot trading volume. Haseeb said that some users believe that the daily trading volume is too stable and lacks weekend lows, so they doubt that the data is abnormal, but he believes that this opinion ignores the structure of participants in the Meme coin market. Since Meme coin trading is mainly driven by retail investors, there will be no significant drop in trading volume over the weekend like Hyperliquid and other markets dominated by professional traders and hedge funds. Haseeb said that users can verify this by checking the revenue data before the Pump.fun token was issued. The revenue stability was similar before, which does not mean there are any abnormalities. Furthermore, he stated that trading terminals and Telegram bots have a high trading volume, so their revenue is not necessarily highly correlated with Pump.fun itself.

32d ago

Stablecoin payment infrastructure Velocity closes $38 million Series A round led by Dragonfly

Comparatively, stablecoin payment infrastructure startup Velocity announced the completion of a Series A round of financing of 38 million US dollars, led by crypto investment firm Dragonfly, with the participation of institutions such as Coinbase, Capital One Ventures, and Wintermute. Velocity CEO Eric Queathem did not disclose the latest funding estimates. Established in 2025, Velocity focuses on providing stablecoin payment solutions for enterprises, payment service providers, fintech companies, and financial institutions to help them optimize cross-border payments, fund settlement, and fund management processes using USD-linked tokens. Currently, Velocity has covered the US, parts of Europe, and the Australian market. In the future, it plans to use the new capital to apply for licenses, expand its business in Africa and Latin America, invest in building more secure asset custody infrastructure, and develop new features including stablecoin yield products.

39d ago#financing

Opinion: AI has yet to trigger the end of DeFi hacking, and overall attack losses have declined this year

Comparing news, Dragonfly managing partner Haseeb Qureshi said that there was no situation where the market previously feared that AI would trigger the end of DeFi hacking. According to the data, although the number of hacking incidents reached a new high in 2026, the median loss from a single attack has fallen below $500,000, compared to over $2 million in 2025. Haseeb believes that AI-enabled attackers are currently mainly targeting small protocols with weak security protection, while larger DeFi protocols have gradually strengthened their defenses. Excluding the Bybit theft incident in 2025 and large-scale attacks such as KelpDAO and Drift Protocol in 2026, the average amount stolen this year is still lower than last year. However, blockchain security company CertiK notes that the crypto industry's losses due to hacking fell 46.8% year-on-year to $1.32 billion in the first half of 2026, which does not mean that the industry's security has improved significantly. It said that last year's data was greatly affected by Bybit's $1.4 billion theft, and more than 70% of losses in the second quarter of this year still came from KelpDAO and Drift Protocol attacks, and related attacks are generally thought to be related to North Korean hacker groups.

39d ago