Crypto Firm Entropy Calls It Quits, Plans Full Investor Refunds

bitcoinist发布于2026-01-27更新于2026-01-27

文章摘要

Entropy, a crypto startup focused on decentralized custody and automation tools, is shutting down after four years. The company, backed by investors like Andreessen Horowitz and Coinbase Ventures, failed to achieve the growth and scalable business model required by its backers. CEO Tux Pacific announced the decision to wind down operations and return approximately $25–$27 million to investors. Despite multiple pivots and product changes, Entropy struggled to gain sufficient customer traction and revenue. The refund process will be handled formally, making this a relatively clean shutdown compared to other crypto failures. The founder may transition to a new field like medical research.

Entropy, a startup that tried to build a safer way to hold and move crypto, is shutting down and sending most money back to investors.

The company’s leader said the business could not reach the size investors wanted. Reports say the team will return roughly $25–$27 million that had been put into the project.

What Happened To Entropy

According to reports, Entropy began with tools for decentralized custody aimed at big holders who wanted more control.

Over time the group changed course and tried to build automation features that would make crypto workflows easier.

The company raised capital from well-known backers, including Andreessen Horowitz and Coinbase Ventures. It ran for about four years and weathered two rounds of layoffs as the team tested different ideas.

In a Saturday post on X, Entropy founder and CEO Tux Pacific said the crypto automation platform has reached the end of the road after years of trying to find a workable future.

Decision To Return Capital

Two clear facts pushed the move. First, buyers and customers did not grow fast enough for the kind of return venture backers expect.

Second, the team struggled to find a steady, repeatable business model that could support rapid growth and hire plans.

Leaders tried product tweaks and new directions, but the pace of change stayed slow and revenue did not climb as hoped. In some cases the product was kept alive by small wins; in others it felt stalled.

BTCUSD now trading at $87,700. Chart: TradingView

Investors will get back most of the money they put in. That makes this shutdown cleaner than some collapses where user funds were at risk.

Reports say refunds will be handled through formal steps and planners are working out the details.

The company’s founder has suggested they may shift their career focus away from crypto, possibly into fields like medical research, though that path is not certain.

Featured image from Pexels, chart from TradingView

相关问答

QWhat is the main reason for Entropy's decision to shut down?

AThe business could not reach the size investors wanted, and the team struggled to find a steady, repeatable business model that could support rapid growth and hiring plans.

QHow much capital is Entropy planning to return to its investors?

AThe team will return roughly $25–$27 million that had been put into the project.

QWhich well-known venture capital firms invested in Entropy?

AEntropy raised capital from Andreessen Horowitz and Coinbase Ventures.

QWhat was the final product direction Entropy was working on before shutting down?

AThe company was working on a crypto automations platform, similar to n8n or Zapier, aimed at making crypto workflows easier.

QHow does this shutdown differ from other crypto company collapses?

AThis shutdown is cleaner because investors will get back most of their money, unlike some collapses where user funds were at risk.

你可能也喜欢

BTC“数字黄金”的叙事是不是失败了?

这篇文章从三个核心问题探讨了比特币的现状与未来,强调提供的是思考框架而非投资建议。 **如何看待比特币资产?** 作者认为比特币是一种全新的、更优秀的“黄金”资产。其优势在于总量恒定、转移便捷、交易可审计。尽管早期与灰色地带关联,但合规化是趋势。目前全球数字货币渗透率仅3%-4%,类比互联网和电商的早期阶段,意味着比特币仍处于发展初期,潜力巨大但波动性也极高。 **如何理解本轮下跌?** 比特币自2025年10月高点(近12.6万美元)持续下跌,2026年2月一度跌破6.1万美元,单日跌幅达15%,随后又快速反弹。这被解读为遵循四年减半周期的共识性获利了结。特别之处在于,美国比特币ETF的批准引入了机构资金,也促使早期低成本持有者(如矿工和信仰者)进行大规模“换手”,这是资产迈向主流化的必经过程。历史数据显示,比特币历次大跌的幅度在收窄(从93%到当前的约50%),表明资产正在成熟,波动率逐步下降,但高波动仍是其获取超额回报的固有特征。 **长期如何看待发展?** 长期价值可对标黄金。当前比特币市值仅为黄金市值的约7%,若“数字黄金”叙事实现一半,上行空间依然显著。但作者提醒,短期市场脆弱,换手可能未完,底部无法预测。真正的风险并非资产归零(概率较低),而在于错误的仓位管理(如All-in或加杠杆)以及对资产缺乏深刻理解。投资者必须计算并承受潜在的最大回撤(例如从已跌50%的位置再跌50%),才能存活至长期价值兑现。 文章最后以亚马逊在互联网泡沫后暴涨为例,指出关键不在于比特币未来是否上涨,而在于投资者能否通过理性的仓位管理和深度认知,扛过剧烈波动存活到那一天。文末提问引导读者反思:当前黄金涨、比特币跌的局面,究竟意味着“数字黄金”叙事失败,还是资产进化过程中的换手阵痛?这取决于每个人对比特币最底层的信仰。

marsbit9小时前

BTC“数字黄金”的叙事是不是失败了?

marsbit9小时前

BTC“数字黄金”的叙事是不是失败了?

标题:BTC“数字黄金”的叙事是不是失败了? 作者:@wuk_Bitcoin 本文从三个核心问题出发,探讨比特币的现状与未来。 **如何看待比特币?** 作者认为比特币是一种全新的、更优秀的“黄金”类资产。其优势在于:总量恒定(2100万枚);资产可转移性极强,在全球不确定性时代具备溢价;所有交易链上可审计,透明度高。反驳了比特币主要用于灰色地带的过时观点,指出其正走向合规。目前全球数字货币渗透率仅约3%-4%,类比互联网和电商早期,意味着该资产类别仍处早期,潜力与巨大波动并存。 **如何理解本轮下跌?** 比特币自2025年10月高点(近12.6万美元)持续下跌,2026年2月初曾单日暴跌15%,跌破6.1万美元。这被视为遵循其四年减半周期的规律性回调,是长期持有者在周期高点锁定利润的结果。本轮下跌的特殊性在于:美国比特币ETF的批准引入了大量机构新资金,但也促使成本极低的早期持有者(矿工、OG)进行历史性抛售,即从“早期信仰者”向“长期配置机构”的换手过程。历史数据显示,比特币历次大回撤的跌幅在逐步收窄(从93%到目前的约50%),表明资产在成熟,波动率在下降,但高波动仍是获取超额回报的代价。 **长期怎么看?** 若将比特币视为“数字黄金”,其当前总市值(约1.4万亿美元)仅为黄金总市值(约20万亿美元)的7%。即使该叙事仅部分实现,上行空间依然可观。但作者强调短期风险:换手可能未结束,市场脆弱,不排除进一步下跌。真正的风险不在于资产归零(概率极低),而在于错误的仓位管理(如All-in、加杠杆)和对资产缺乏深度理解,这可能导致投资者无法承受巨大波动而提前被迫出局。 **最后对比** 作者以亚马逊在互联网泡沫破裂后股价跌95%又最终上涨42倍为例,指出关键在于“活着等到那一天”。对于比特币,核心同样是能否通过理性仓位管理活到其价值兑现之时。文末提问:当黄金大涨而比特币大跌,这究竟是“数字黄金”叙事的失败,还是资产进化过程中的阵痛?答案取决于每个人对比特币最底层的信仰。

链捕手10小时前

BTC“数字黄金”的叙事是不是失败了?

链捕手10小时前

从代码到认知:机器人大脑进化的万字指南

本文概述了机器人大脑从传统代码控制到现代人工智能模型驱动的演进历程。文章首先回顾了前大型语言模型(LLM)时代,机器人依赖手工编码的模块化技术栈(感知、状态估计、规划、控制)和行为树,虽稳定但泛化能力差。随后,深度学习改进了感知,强化学习和模仿学习进入了控制层,但策略仍较为狭窄。 ChatGPT的出现带来了转折。LLM最初被用作自然语言编译器,将指令转化为机器人可执行的原子技能序列(如谷歌的SayCan)。但更重要的突破是视觉-语言-动作模型(VLA),例如谷歌的RT-2和开源的OpenVLA,它能将视觉、语言信息融合,直接输出动作指令,实现了推理与行动的耦合。 目前最先进的系统采用“双脑”架构(如Figure AI的Helix、NVIDIA GR00T):一个慢速、参数多的“系统2”负责高层次推理和规划;一个快速、小巧的“系统1”负责高频动作生成。其下还可能有一个“系统0”反射层处理平衡等底层控制。出于延迟和可靠性考虑,安全关键的控制回路通常在机器人本地(如NVIDIA Jetson模块)运行,而对话界面和集群学习等任务可交由云端。 开源模型(如OpenVLA、GR00T、π0)降低了行业门槛,让初创公司能在其基础上用自有数据微调。然而,当前VLA机器人仍存在任务中途恢复能力弱、样本效率低、缺乏物理常识和长期规划能力等局限。 这催生了下一代方向:世界模型。这类模型(如NVIDIA Cosmos、Meta V-JEPA)能根据当前状态和动作预测未来结果,让机器人在行动前进行模拟和评估,从而改善恢复能力、泛化能力和长期规划。架构上主要分为像素级视频扩散、联合嵌入预测架构(JEPA)和潜在动作世界模型等流派。 文章最后指出,数据采集(特别是远程操作数据)是核心竞争力,仿真训练至关重要,机器人成本正在迅速下降。当前物理AI的发展阶段大约相当于“GPT-2时代”,虽未完全自主,但正通过架构的持续演进(从代码到感知、规划、策略,最终到世界模型),朝着更通用、更强大的方向稳步前进。

marsbit10小时前

从代码到认知:机器人大脑进化的万字指南

marsbit10小时前

交易

现货
合约
活动图片