Berlin has long been known as the birthplace of early internet startups focused on consumer, marketplaces and software, from e-commerce giants like Zalando to quick commerce companies like Delivery Hero.
But with companies increasingly embedding AI in their solutions, a whole new generation of startups are being built in the German capital.
In the first half of 2025, Berlin startups raised €1.9bn across 91 deals — a dip from the first half of 2024, which saw over €2.5bn raised across 92 deals, according to Sifted data.
Recent big fundraises including CarOnSale,a B2B marketplace for used car trading which closed a €70m Series C, and Talon.One, which helps retailers offer promotions and loyalty schemes
So which companies in Berlin are on investor’s radars? Sifted asked VCs from Cherry Ventures, Project A, 3VC and Atlantic Labs to share their top picks that are not from their portfolios.
Filip Dames, founding partner at Cherry Ventures
n8n
Founded in 2019, n8n is a workflow automation platform that enables developers and non-developers to automate workflows and integrate apps using an open, extensible and self-hostable low-code platform. As more companies build internal automations and AI-enabled workflows, platforms like n8n become essential infrastructure.
Langdock
Langdock, founded in 2013, is an all-in-one platform helping companies deploy AI in their organisations. The company is one to watch given the strong team and current traction. It’s compelling as it helps companies deploy secure, customisable AI copilots for internal knowledge and workflows, streamlining enterprise adoption of LLMs, and providing a flexible and effective setup offering exactly what CIOs are looking for.
Langfuse
Founded in 2022, Langfuse offers observability and analytics for LLM applications, making it easier for teams to monitor, evaluate, and iterate on AI product performance. I like it because it’s more of a picks-and-shovels play for the way tech teams are using AI, with an interesting OpenSource approach.
Eva Arh, partner at 3VC
Interloom
Founded in 2024, Interloom offers AI-powered automation software for enterprises. As AI becomes further adopted, we will use not just copilots, but also AI Agents that take over certain tasks. To fully leverage agents, it is important that humans and agents can co-exist and collaborate in an effective, complementary way. Interloom is building a platform for human-agent orchestration in the enterprise, helping companies decide which tasks should be done by humans and which by agents.
TipTap
TipTap, founded in 2022, is an open-source solution that enables companies to implement collaborative editing functionality quickly, securely, and in a scalable manner. Their OS traction is remarkable and includes companies like LinkedIn and Anthropic. With content creation reaching new heights thanks to AI, I believe the need for collaboration is also increasing.
Jack Wang and Philipp Werner, partners at Project A
Reflex Aerospace
Founded in May 2021, Reflex Aerospace is building customised, high-performance satellites with modular, software-defined designs which it can deliver within 12 months. This is special because traditional aerospace firms typically take years with limited customisation, while NewSpace startups often lack reliability and scalability.
Juna.ai
Juna, founded in 2023, optimises industrial manufacturing processes through autonomous AI agents — essentially, turning factory controls into ‘self-driving’ systems. Its AI agents plug into existing production lines and continually fine-tune thousands of settings to boost output and cut waste. Where traditional industrial software stops at static dashboards, Juna’s self-learning agents take the wheel, delivering real-time optimisation for heavy industries like steel, cement and chemicals.
Daniel Niemi, partner at Atlantic Labs
Dunia
Founded in 2022, Dunia automates the discovery of materials through robotics-powered laboratories that accelerate R&D cycles in the chemicals industry. By combining physical automation with AI-driven research, the company aims to reduce dependency on petrochemicals while cutting discovery timelines from years to months. Its integrated approach to both software and hardware positions it to capture value across the materials innovation stack.
Lyceum
Lyceum — a company founded early this year — is building a GPU cloud platform to help companies train machine learning models faster and more cheaply. As European teams seek alternatives to hyperscaler dependency, Lyceum’s approach to scheduling and scaling removes operational overhead, allowing researchers to focus on model development rather than infrastructure management.
The company recently announced it has raised a total of €10.3m, led by Redalpine with participation from 10X founders.
Peec AI
Peec AI, aimed at marketing teams, helps measure, analyse and improve a company’s brand performance on AI search platforms like Google Search. It addresses the emerging challenge of brand visibility in AI-generated responses. As search behavior shifts toward conversational AI, companies need tools to monitor and optimise their presence across language models. Peec’s real-time tracking of prompts, sources, and sentiment provides the visibility layer that enterprises will increasingly require as AI becomes the primary discovery interface.
The company, founded in January 2025, recently raised a €5.2m seed round led by 20VC, with participation from Antler, Foreword VC, Identity VC, Combination VC and S20.
Read the orginal article: https://sifted.eu/articles/berlin-startups-to-watch/
