The recent surge in interest surrounding the startup DeepSeek has sparked intense discussions within the European venture capital (VC) community, with many claiming that missing out on this opportunity is akin to neglecting the dawn of the internetMeanwhile, the Chinese investment scene is engaging in crucial introspection due to its conspicuous absence from this technological breakthroughAs European investors redirect their focus toward innovative Chinese enterprises for models worth emulating, venture capital firms across major cities from Beijing and Shanghai to Hangzhou and the shores of Shenzhen Bay are dusting off their project evaluations and preparing to dive back into the fray.

Typically, venture capital operates through equity funds to provide financial backing for startups, with the aim of capital appreciation through eventual IPOs or acquisitionsHowever, interviews conducted with various VCs across Europe and China reveal that this tech wave transcends mere technicalities, evolving into a critical test of the global innovation ecosystemThe seemingly vibrant venture capital sector is rife with undercurrents and ongoing power plays.

After the remarkable rise of DeepSeek—a Hangzhou company that achieved remarkable performance akin to GPT-4o with just a $557.6 million investment in computational power—the team behind it shattered the technocratic iron curtain often associated with AI giantsCentral to the narrative of DeepSeek is the migration of top talent back to ChinaIts lead member, Pan Zizheng, previously accepted a full-time position at Nvidia but opted instead to join DeepSeek's fledgling three-person team, spearheading the development of their R1 modelRegarded as the “AI Qian Xuesen”—an allusion to a celebrated Chinese scientist—his return to China has raised eyebrows, leading notable voices like Harvard's Graham Allison to ponder: “Who truly missed the boat with DeepSeek?”

This collective astonishment among investors prompted a seismic shift within the venture capital landscape

Advertisements

As Yu Wenhui, founding partner of Thunderstorm Capital, shared his experiences with rapid news platforms, he was traveling to AI conferences throughout Guangzhou, Shenzhen, and Hong KongIn his view, the eruption of DeepSeek represents a transformative event on a national scaleThis moment not only heralds a rebound in the valuation of Chinese companies but also reshapes the global AI innovation ecosystem, revitalizing a previously subdued venture capital domainDeutsche Bank expressed this sentiment aptly, likening China to Japan in the 1980s.

Central to this upheaval is the overthrow of existing technical paradigmsDeepSeek has utilized an open-source model to dismantle technological monopolies, challenging the prevailing belief in computational supremacy with its low-cost training systemsYu Wenhui stated bluntly, “Even with hardware disparities, China's advancements in software technology have left OpenAI blushingThe winds of change brought forth by DeepSeek are reaching across the ocean and stirring the tech battles in Europe and AmericaWhile venture capital aims for profitability, it also represents a significant financial innovation since World War II, propelling global developmentWithout Microsoft's $10 billion injection and the subsequent support of venture capital, ChatGPT would not exist.”

While Yu Wenhui was embroiled in this whirlwind of activity, his counterpart Cab in Berlin felt a contrasting sense of disappointmentRegularly seen at the Ben Rahim café in Berlin's Mitte district, she would scroll through LinkedIn to monitor her Chinese peers’ fervent discussions about DeepSeekThis café is situated in Germany's venture capital hub, and the grey waters of the Spree River reflect her perception of a dim European AI startup ecosystem.

Having previously ventured in entrepreneurship within China, Cab exhibits a sense of dismay over the current state of the European VC landscapeJokingly, she mused on social media about wanting to “run away from the European venture capital scene.” A mere three years prior, she had joined the investment world under the auspices of her German MBA, unaware that this land, once home to geniuses like Einstein and Leibniz, would fall by the wayside during the generative AI revolution.

The uproar from DeepSeek rippled through European VC circles, leading Cab—who is proud of her Chinese roots—to feel like an ancillary support staff member at a top investment bank, relegated to the periphery of the core AI "battle." Previously, the success of Mistral AI had sent European VCs into a frenzy, with investments cascading in from firms like LVP, Index, and a16z, crowning it as Europe’s “hope of the village.” However, its momentum proved insufficient, leading to its gradual decline last year, exposing various deficiencies within Europe’s tech ecosystem, such as insufficient capital, market fragmentation, talent drain, policy restrictions, and disconnects between academia and industry.

Nevertheless, the buzz surrounding DeepSeek has instilled a spark of hope in Cab

Advertisements

Many AI startups across Europe have started integrating DeepSeek's frameworks, with positive feedbackIn a recent conversation with the CTO of a protein engineering startup, the executive praised DeepSeek's deep reinforcement learning methods and plans to apply them to their model training.

As Chinese large models breach the walls of the technological bastions, European entrepreneurs are finally waking up to the reality that, rather than struggling in the saturated base model sea, it would be more beneficial to create their own waves at the application layerCab insists she can distinctly perceive the ripples from China's large model tech crossing continents, a phenomenon that she believes will influence investment decisions within European venture capital firms in the future.

Reflections on the global VC landscape highlight how the DeepSeek thunderstorm has sparked ripples across the arteries of the Chinese investment ecosystemHigh-profile executives once dismissive of startup projects are now actively seeking partnerships with Yu Wenhui, and personal investors are knocking on VC doors with millions in their coffersInstitutions are vying to secure entry into the space with angel investments reaching into the tens of millions.

“Investing is often painful, as it requires long-term commitment, and startups generally have low success rates amidst numerous hardshipsA project may take five or even ten years to mature, necessitating patience and an ability to delay gratificationAngel investing, in particular, hinges on faith in people, human nature, and luckStrong angels lead to strong startups; strong startups bolster national strength.” Amidst his journey, Yu has witnessed numerous missed opportunities, such as Baidu Ventures bypassing DeepSeek and his friends around Guangzhou eschewing investment in Yushu Technology.

In industry discussions, Yu emphasized the need to identify the next DeepSeekHis succinct assertion encapsulated the anxieties and hopes shared by many investors

Advertisements

He articulated that DeepSeek’s sudden rise encapsulates both inevitable and accidental factors. “Had it not been for the return of key personnel like Pan Zizheng to China, DeepSeek might not have emergedYet, the booming AI industry provides fertile soil for such innovative enterprises.” As the winds of change brought forth by DeepSeek reshuffle the foundational laws governing the VC ecosystem, an essential new paradigm emerges.

In his investment notes, Yu summarized succinctly: “Top-tier founders must choose top-tier directionsIf they opt for a mediocre path, they will only pursue mediocrityThus, top-tier founder plus top-tier direction equals top-tier enterprise.” As one gazes back from 2025's intersection, Yu’s investment landscape reflects a multifaceted spectrum of China’s technological ascentFrom grassroots charging station businesses to AI translators expanding abroad, from specialized models in niche sectors to the healthcare arena, every move by Thunderstorm Capital is a validation of this premise.

Firmly believing in the burgeoning fields of AI general intelligence, which encompass vertical industry models, robotics, low-altitude economies, and AI agents, Yu exchanged views on the notion that AI agents possess a potential market scale in the trillions of dollars—eventually replacing human roles across various sectors, becoming the next investment frontier.

As DeepSeek’s technological current washes away traditional barriers, only those investors willing to bet on the “next step” can hope to gain a foothold in the age of AI multipolarityYu underscores this sentiment, emphasizing, “This is not a competition between China and the United States; it’s a race between humanity and technological evolution.” On this track, every VC symbolizes a historical benchmark, where DeepSeek merely signifies the prologue.

Meanwhile, in contrast to the relative quiet of Guangzhou’s tech scene, Hangzhou commanded attention in this energetic competition

Following DeepSeek’s stunning debut, the so-called “Six Little Dragons” of Hangzhou—Game Science, Deep Exploration, Yushu Technology, Yunshen Technology, Qiangnao Technology, and Qunkong Technology—have emerged as a remarkable force within the global tech arenaTheir ascendance has prompted considerable reflection from multiple state media outlets questioning the underlying reasons behind “Why Hangzhou?”—a topic of heated discourse.

Wang Le, whose firm has operated in Hangzhou for nearly a decade, has witnessed technological upheavals from mobile internet to fintech and now to AI roboticsAs an alumnus of Zhejiang University, she possesses a profound insight into the transformative impacts wrought by DeepSeek. “The explosion of DeepSeek has set the entire VC community aflame,” she noted. “Currently, nearly all investment firms are keeping their fingers on the pulse of projects, and enthusiasm among investors surpasses anything we've seen before.”

Though DeepSeek originated in Hangzhou, its influence has breached the city’s confinesAs Zhu Xiaohu, managing partner at Jinsaijang Venture Capital, pointed out, Chinese institutions are eagerly positioning themselves, and Wang Le’s VC firm is no different, preparing for a more aggressive approach to AI investments.

Wang revealed that this year, Liang Wenfeng and his team are steadfastly focused on model development, maintaining a certain distance from external engagementEven in dealings with institutions led by Zhejiang University alumni like Wang, their interactions are limited, predominantly communicated through DeepSeek's internal development teamYet Wang feels the profound impact this project has cast over Hangzhou.

“This year, we intend to invest vigorously in applications based on DeepSeek or other foundational models,” she mentioned, exuding confidence about the future of AI in ChinaPreviously, analysts estimated that the technology gap between major Chinese enterprises and their American counterparts spanned more than two years

Today, each generational model’s disparities have shrunk to just two to three monthsThe impressive capacity for Chinese engineers to implement and explore systematically has invigorated the industry, boosting confidence all around.

DeepSeek’s model is not only open-source and controllable but also cost-effective, dramatically lowering inference expenses and laying the groundwork for extensive commercial applications reminiscent of the early prosperity of the Android ecosystemWang lamented that while many AI companies have launched from Beijing over the past few years, the quiet ambiance of Hangzhou belies the myriad diversifications and robust entrepreneurial support from the government, rendering it a cradle of innovation—a process of laying solid foundations paving the way for future breakthroughs.

Liu Xiaoying, a strategic partner at Yu Wenhui’s firm, previously invested in one of Hangzhou’s six little dragons, BrainCoHe revealed plans to establish a new branch in Hangzhou to capitalize on the innovative energy radiating from this vibrant cityReflecting on his past beliefs about “geographical advantage being secondary to cooperation,” he noted that witnessing the successes of projects like Black Myth: Wukong, Nezha 2, and DeepSeek alongside Yushu Technology’s breakthroughs illuminated the importance of timing and opportunityHe believes that great achievements stem from an intricate interplay of timing, location, and collaboration.

The excitement surrounding DeepSeek has created an apparent undercurrent within the venture capital community, marked by both celebration and unseen rivalries. “Bullets and guns are key,” a metaphor combining funding and exit strategies encapsulates the fundamental importance of financial backing and viable exit channels in the VC sectorThe current market excitement has led VCs to chase opportunities zealously, resulting in strategic diversifications among leading firms

Some have shifted from “single-star bets” to crafting “ecosystem strategies,” with some institutions jockeying for position through inflated valuations and weak terms, while the “wait-and-see faction” perceives looming valuation bubbles and opts to delay entry until the next technological or commercial validation emerges.

Feng Bo from Changlei Capital, with ten years of experience in venture capital and operating in Shanghai, intimately understands that, despite DeepSeek's capacity to evoke monumental waves within the VC realm, the absence of meaningful improvement in funding and exit mechanisms could cause this momentum to dissipate rapidlyRemembering the industry’s exceptionally prosperous period—when investment firms reached tens of thousands and project resources abounded—Feng reflected on the stark reduction in the number of IPOs scheduled for 2024 compared to peak levels, with only about one-fifth still on course.

The era of windfall profits in the industry has long since waned; presently, any fund that can yield three times returns is deemed exceptionally successfulHowever, the harsh reality is that over 90% of funds struggle to even reclaim their principal, and far fewer achieve twofold or even singlefold returnsPrior to DeepSeek’s emergence, numerous funds had invested heavily in opaque models, with several reaching valuations of tens or even close to billionsYet with the free, open-source versions of DeepSeek appearing, those startups that had poured substantial funds now find themselves maneuvering through treacherous waters, with investors anxious about their investments potentially evaporating, while other foundational models immediately lose competitive edge, leading to hesitance in funding further developments in this regardSome institutions that previously heavily invested in AI are now wary of “technological illusions,” questioning whether foundational models will maintain their first-mover advantages or if contributions from open-source communities can solidify into sustainable business barriers.

Within this context, Feng Bo’s Changlei Capital, operating as a private small fund, grapples with limited capital and a challenging exit landscape, confining them to invest in early-stage, low-valuation projects with potential exit avenues or collaborate with publicly-listed companies

Advertisements

Advertisements