德国初创企业SpiNNcloud创始人的融资诀窍：尽早实现盈利、自力更生

（原标题：德国初创企业SpiNNcloud创始人的融资诀窍：尽早实现盈利、自力更生）

尽管融资条件困难，德国初创企业数量仍在逆势增加。2025年初，德国初创企业协会统计数据显示，当地2024年有2766家初创企业成立，比2023一年（2498家）多出约11%。

德国联邦外贸与投资署（德意志联邦共和国经济发展机构，下称“GTAI”）趋势与创新探索主管Philipp K?vener曾对经济观察报表示，初创企业是德国经济恢复增长的最为关键的因素之一。GTAI等不同部门为不同行业、不同发展阶段和不同地区的初创企业提供不同的支持方式，有时甚至是财政支持。

Hector A. Gonzalez所在的SpiNNcloud（一家专注于开发AI人工智能芯片和高性能计算系统的德国初创企业），便是上述企业之一。

身为SpiNNcloud联合创始人兼联合首席执行官，Hector A. Gonzalez给经济观察报提供的第一个关键词，便是“国际化”。

从2021年企业创立之初，他和团队就立志推进公司的国际化。因为在AI产业领域，国际上的“大玩家”都非常国际化，且具有强大的高科技产业背景。如果一家创业企业想参与其中，就必须实现国际化，才能与这些“大玩家”同台竞争。

想要实现这个目标并不容易。

在德国境内，企业的融资难度正在增加，2024年比2022年、2023年更难获取风投资金。在这种环境下想要崭露头角，初创企业需要直面多重挑战。

Hector A. Gonzalez曾为知识产权商业化而头痛，也为融资资金、专业人才建设而奔走。

在危机之中，他找到了新路。

SpiNNcloud联合创始人兼联合首席执行官，Hector A. Gonzalez（受访者供图）

经济观察报：你来自哪里，为什么想创业，是什么想法或市场需求促使你创办这家企业？

Hector A. Gonzalez：SpiNNcloud是一家位于德国德累斯顿的神经形态计算领域的初创企业，更确切地说，是大脑启发计算领域的初创企业。

公司的技术由 ARM 架构（目前全球几乎所有手机都采用该架构）的发明者史蒂夫-福伯（Steve Furber）共同开发，是欧盟最大的研究计划“人脑计划 ”的一部分。这项技术的第二代产品主要在德累斯顿工业大学（TU Dresden）开发，团队看到了这项技术在解决人工智能超级计算目前面临的一些最大挑战方面的巨大潜力，例如人工智能计算中能效的缺乏，以及稳健可靠的人工智能的开发。有鉴于此，团队创办了SpiNNcloud公司，作为德累斯顿工业大学的衍生公司。

SpiNNcloud的创始团队汇集了50多年的行业经验和学术经验。作为从德累斯顿工业大学分拆出来的公司，团队一开始就拥有一个巨大的优势——完全成熟的超级计算技术，而不是从零开始。

团队在硬件和软件系统设计方面的综合背景，加上大家的商业头脑，迅速加快企业的成长和发展。

就我个人而言，我拥有高能效人工智能芯片和软件方面的业内背景。 

经济观察报：创业初期，如何确定企业的核心业务和发展方向？中途有没有改变过这个方向？

Hector A. Gonzalez：作为超级计算领域一家年轻的初创公司，团队想将尖端技术推向市场，因此团队早期采取了以开源软件为中心的战略，以达到用户所期望的成熟度。

这种预判促使团队扩大软件团队和社区，重点开发软件堆栈和特定用例应用程序。在这个过程中，团队从一家主要以硬件为中心的公司发展成为一家同时连接硬件和软件领域的公司。

团队的创始团队非常特别，因为它是自发组建起来的。没有哪一个人决定去招募不同背景的人才，而是围绕着强大的芯片设计团队Christian Mayr这个中心焦点而聚在一起。凭借团队各自拥有的不同的背景和专业技能，在决定创立SpiNNcloud的时候，团队很自然也很清楚每个人能为团队带来什么价值。 

经济观察报：现在是创业的好时机吗？为什么？德国当前的经济环境和社会趋势为你的初创企业提供了哪些机遇？

Hector A. Gonzalez：在SpiNNcloud，团队认为现在永远是创办初创企业的好时机——尤其是像团队这样已经拥有功能齐全的突破性技术的企业。团队坚信，创业心态能以多种方式造福社会，这也是团队积极鼓励他人踏上创业之路的原因。对于（年轻的）创始人来说，德国政府和欧盟都实施了许多创业和财政支持计划。这些计划不仅惠及初创企业个人，也惠及德国和欧盟的整个初创企业生态系统。团队坚信，通过利用这些机会和杰出的创意、技术或解决方案，团队都能从由此带来的创新和增长中受益。这种进步不仅会推动德国的进步，也会推动整个现代社会的进步。

尽管目前欧洲风险投资有所下降，但团队坚信：强大的产品和令人信服的商业案例相结合，总会吸引那些认同你的愿景的投资者。根据团队的经验，有几个因素可以显著提高成功的机会：坚持不懈、尽早实现盈利、研发具有良好应用前景的创新技术，以及尽可能长时间自力更生。

此外，地方、国家和欧洲各级政府都有各种补助金和财政支持机制。这些机制可以有效地充当引导替代方案，使公司能够在不稀释股权的情况下启动和发展。 

经济观察报：在融资过程中，你们更倾向于哪种融资方式，如风险资本、银行贷款或政府基金等，为什么？

Hector A. Gonzalez：与许多商业决策一样，融资的选择完全取决于你的具体目标。风险资本可以提供加速发展和扩大市场规模所需的资源。另一方面，银行贷款有助于为生产运行提供前期资金或满足短期资金需求。最后，政府资金，尤其是用于研究项目的资金，可以帮助你保持在创新领域的领先地位。

在SpiNNcloud，团队不相信一刀切的融资方法。相反，团队会对每种情况进行单独评估，然后选择最符合团队预期结果的金融工具。团队在战略性地利用各种融资工具方面拥有丰富的经验，这些融资工具支持团队不断进行市场渗透和技术创新。 

经济观察报：现在这一行业的竞争有多激烈？SpiNNcloud如何在市场中定位自己，与竞争对手相比有哪些独特优势？

Hector A. Gonzalez：在人工智能、云计算和专用计算硬件进步的推动下，超级计算/高性能计算行业竞争激烈，发展迅速。目前，主流硬件架构包括用于通用计算的CPU（中央处理器）和用于人工智能工作负载（包括训练和推理）的GPU（图形处理器）。然而，神经形态计算等新兴半导体范式有望彻底改变整个市场。大脑启发方法使团队能够克服主流处理器架构所面临的一些固有挑战，尤其是在人工智能生态系统正在经历算法突破，而这些算法并不适合现有解决方案的情况下。后者使团队在能效方面具有关键优势。 

经济观察报：你们如何了解和分析竞争对手，是否制定了具体的竞争战略，使自己与竞争对手区别开来？

Hector A. Gonzalez：团队非常了解竞争对手的生态系统。与其他竞争者相比，SpiNNcloud的解决方案的关键差异在于，团队的架构并非源自主流人工智能趋势。相反，团队的架构源自欧盟人脑旗舰项目数十年的研究。这使团队能够支持主流人工智能软件社区在近期算法突破中发现的动态稀疏原则，而不是受限于它们的局限性。 

经济观察报：你认为扩大市场份额的最大挑战是什么？

Hector A. Gonzalez：SpiNNcloud正在整个技术应用生命周期内战略性地扩大市场份额。团队目前的客户群包括欧盟和美国的知名机构和数据中心这些早期采用者。同时团队也在积极拓展其他行业、高性能计算中心和超级分频器等市场。

作为脑启发超级计算领域的先驱，团队还开始引导行业对话，并为下一代人工智能计算制定新标准。通过SpiNNcloud围绕大脑启发计算的思想领导力计划，团队正在展示突破性解决方案如何从根本上改变人工智能的格局，改变整个计算模式，并为社会创造前所未有的价值，同时定义节能、灵活的高性能计算的未来。

作为一家深度技术分拆公司，团队将德累斯顿工业大学和曼彻斯特大学开发的SpiNNcloud平台商业化。这一超级计算解决方案基于 SpiNNaker2芯片，该芯片使用大量低功耗处理器，可高效计算人工智能和其他工作负载。芯片架构和系统设计的灵感均来自人脑原理——强调并行性、能效和实时处理能力。该平台可扩展至超级计算机级别，并可作为云解决方案提供给数量有限的用户，其定位是首个由大脑启发的商用高性能计算系统，彻底改变了传统的计算模式。

团队扎根于该领域一些最负盛名的大学院系。在可预见的未来，团队仍将与学术界保持紧密联系。另一方面，团队拥有早期客户，并获得了地区、国家和欧洲机构的资助，这为团队积极推行商业化战略提供了支持。当然，这也包括定期评估必要的内部研发步骤，因为团队还在硬件方面大力发展下一代技术的研发团队，并开发易于使用的软件栈，使未来的客户能够创建在团队的硬件上运行的节能、强大的人工智能模型。 

经济观察报：目前国际化进展如何？遇到了哪些问题？

Hector A. Gonzalez：公司已经通过位于旧金山的子公司进入了美国市场，并且正在积极探索其他国际市场，目前包括阿拉伯联合酋长国。在应对不断变化的全球贸易格局时，团队正在实施战略方法来解决市场准入要求、认证和进口法规等问题。团队的国际扩张战略考虑了这些市场动态，同时确保团队继续向全球关键市场提供尖端解决方案。团队认为，促进国际合作和市场准入对推动超级计算领域的技术创新至关重要。

在每年年初，团队的首要目标始终是继续团队的增长道路。对团队来说，无论规模大小，渐进式增长都很重要。基于这一点，自SpiNNcloud成立以来，在过去几年中表现良好，持续扩张。这仍然是团队2025年的目标。增长定义了目标，团队将继续追求增长，同时努力成为节能AI基础设施领域的世界领导者。

The Economic Observer:Tell us where you are from, why you wanted to start your own business and what idea or market demand drove you to start this start-up?

Hector A. Gonzalez:SpiNNcloud is a Dresden-based startup in the area of Neuromorphic Computing, or, more specifically, brain-inspired computing. Our technology was co-developed by Steve Furber, the inventor of the ARM-architecture (which powers virtually all of today’s mobile phones world-wide) as part of the European Union’s largest research initiative, The Human Brain Project. With the second generation of this technology developed predominantly at the Dresden University of Technology (TU Dresden), we saw the big potential of this technology to solve some of the biggest challenges that AI supercomputing faces today, like for instance a lack of energy efficiency in AI computing and the development of robust and trustworthy AI. With this in mind, we started SpiNNcloud as a Spin-off of the TU Dresden.

Our founding team brings together extensive experience of over 50 years from both industry and academia to the table. As a spin-off from TU Dresden, we started with a significant advantage –a fully developed supercomputing technology – rather than starting from scratch. Our team's comprehensive background in hardware and software system design, combined with our business acumen, has enabled us to significantly accelerate our company's growth and development. In my personal case, I have a background in energy-efficient AI-powered chips and software. 

The Economic Observer:How did you determine the core business and development direction of the company at the initial stage of starting the company, and have you ever changed this direction?

Hector A. Gonzalez:As a young startup in the supercomputing field with the mission to bring cutting-edge technology to the market, we adopted an early strategy that is centered around open-source software to achieve the maturity expected by our users. This insight led us to expand our software team and community with a focus on developing both software stack and specific use case applications. Through this process, we evolved from a primarily hardware-focused company into one that bridges both hardware and software domain.

Our founding team is quite special because it assembled itself. No single person decided to recruit different profiles; rather, we all gravitated around a central focus point—Christian Mayr’s strong chip design group. With our diverse backgrounds and expertise, it was natural and clear what each of us brought to the team when we decided to found SpiNNcloud.

The Economic Observer:Is this a good time to start a business? And why? What opportunities does the current economic environment and social trends in Germany offer for your start-up?

Hector A. Gonzalez:At SpiNNcloud, we believe it is always a good time to found a startup – particularly when, as in our case, you already start with a fully functional groundbreaking technology. We are strong believers that entrepreneurial mindsets can benefit society in numerous ways, which is why we actively encourage others to embark on this incredible journey. For (young) founders, both the German government and the European Union have implemented numerous entrepreneurial and financial support programs that benefit not only individual startups but the entire startup ecosystem in Germany and the EU. We strongly believe that by leveraging these opportunities alongside brilliant ideas, technologies, or solutions, we can all benefit from the resulting innovation and growth. This advancement will drive progress not only for Germany but for the modern society as a whole.

Despite the current decline in European VC investments, we firmly believe that a combination of a strong product and compelling business case will always attract investors who share your vision. From our experience, several factors can significantly improve the chances of success: persistence, generating early revenue, developing innovative technology with promising future applications, and bootstrapping for as long as possible.

Additionally, various grants and financial support mechanisms are available at local, national, and European government levels. These can effectively serve as bootstrapping alternatives, enabling companies to launch and grow without diluting equity. 

The Economic Observer:Which financing method do you prefer in the financing process, e.g. venture capital, bank loans or government funds etc. and why?

Hector A. Gonzalez:As with many business decisions, the choice of funding depends entirely on your specific goals. Venture capital can provide the resources needed to accelerate growth and market scaling. Bank loans on the other hand can be instrumental in pre-financing production runs or to bridge short-term financial needs. Lastly, government fundings, particularly for research projects, can help maintain your position at the forefront of innovation.

At SpiNNcloud, we do not believe in a one-size-fits-all funding approach. Instead, we evaluate each situation individually and then select the financial instrument that best aligns with our intended outcomes. We have had good experience in strategically leveraging a diverse mix of financing instruments that support our ongoing mission for market penetration and technological innovation. 

The Economic Observer:How competitive is the industry in which your company operates, how do you position yourself in the market and what are your unique advantages over your competitors?

Hector A. Gonzalez: The supercomputing/HPC industry is highly competitive and rapidly evolving, driven by advances in AI, cloud computing, and specialized compute hardware. Currently, the mainstream hardware architectures consist of CPUs for generic computing and GPUs for AI workloads, including both training and inference. However, emerging semiconductor paradigms like neuromorphic computing are poised to revolutionize the entire market. Our brain-inspired approach enables us to overcome several inherent challenges faced by mainstream processor architectures, especially as the AI ecosystem is experiencing algorithmic breakthroughs that don’t fit well in existing solutions. The latter gives us a key advantage when it comes to energy efficiency. 

The Economic Observer:How do you understand and analyze your competitors and have you developed specific competitive strategies to differentiate yourself from them?

Hector A. Gonzalez:Absolutely, we are well acquainted with the competing ecosystem. The key differentiation of our solution compared to other players is that our architecture is not derived from mainstream AI trends. Instead, it stems from decades of research conducted by the EU Flagship Human Brain Project. This allows us to support the dynamic sparse principles found in recent algorithmic breakthroughs by the mainstream AI software community, rather than being constrained by their limitations.

The Economic Observer:What do you think is the biggest challenge in increasing your market share and how do you deal with it?

Hector A. Gonzalez:We are strategically expanding our market presence across the Technology Adoption Life Cycle. While our current customer base includes prestigious institutions and data centers in the EU and US as early adopters, we are actively expanding into other industries, HPC centers, and hyperscaler markets. To serve these diverse market segments, we are continuously enhancing our software stack and solution suite. Our growing team is focused on developing comprehensive value propositions that resonate with both academic and industrial users.

As pioneers in brain-inspired supercomputing, we have also started shaping the industry dialogue and setting new standards for next-generation AI computing. Through our thought leadership initiatives around brain inspired computing, we are demonstrating how our groundbreaking solutions are fundamentally transforming the AI landscape, changing entire computing paradigms and creating unprecedented value for society while defining the future of energy efficient and flexible high-performance computing.

As a deep tech spin-off, we commercialize the SpiNNcloud platform developed by TU Dresden and the University of Manchester. This supercomputing solution is based on the SpiNNaker2 chip, which uses an extremely large number of low-power processors for efficiently computing AI and other workloads. Both the chip architecture and system design are inspired by the human brain's principles – emphasizing parallelism, energy efficiency, and real-time processing capability. The platform is scalable to supercomputer level and is available as a cloud solution for a limited number of users, positioning itself as the first commercial brain-inspired HPC system that revolutionizes traditional computing paradigms.

Being rooted in some of the most prestigious university departments of this field, we will remain having a strong connection to academia for the foreseeable future. On the other hand, having early customers and receiving grants from both regional, national and European institutions supports us in actively pursuing our commercialization strategy, which of course also contains a regular assessment of the necessary R&D steps internally as we also strongly grow our team in the areas of researching and developing the next generation of the technology on the hardware side and on developing an easy to use software stack that enables future customers to create energy efficient and robust AI models running on our hardware. 

The Economic Observer:Has your company ever considered internationalization? If so, how far have you gotten so far and what problems have you encountered?

Hector A. Gonzalez: We are already present in the United States through our San Francisco-based subsidiary, and we are actively exploring additional international markets, including currently the United Arab Emirates. In navigating the evolving global trade landscape, we are implementing strategic approaches to address market entry requirements, certifications and import regulations. Our international expansion strategy accounts for these market dynamics while ensuring that we continue to deliver cutting-edge solutions to key markets worldwide. We believe that fostering international collaboration and market access is crucial for advancing technological innovation in the supercomputing sector.

At the beginning of each year, our primary goal is always to continue our growth path. For us, incremental growth is important, regardless of its scale. Based on this, we have performed well in the past years since founding SpiNNcloud, consistently expanding. This remains our goal for 2025. Growth defines our objectives, and we will continue pursuing it while striving to become world leaders in energy-efficient AI infrastructure.