Recent developments highlight a significant milestone in the world of artificial intelligence as a rising unicorn in the AI large model sector, known as StepStar, successfully completed a Series B funding roundThis round amassed several hundred million dollars, attracting a consortium of investors, including state-owned capital from Shanghai, Tencent Investments, Wuyuan Capital, and Qiming Venture Partners, among others.

The capital influx will primarily be channeled towards further enhancing the capabilities of foundational AI models, particularly emphasizing multi-modal functionalities and enhancing complex reasoning capabilitiesThese upgrades aim to broaden the reach of consumer applications and ultimately to elevate user experiences across various platforms.

StepStar was founded relatively recently, in April 2023, and is headquartered in Xuhui District, Shanghai, under the helm of DrJiang Daxin, a former global vice president at Microsoft.

Although many may not be familiar with StepStar, its impact on the AI landscape is growing rapidlyThe company has successfully carved out a place among the so-called “Five Little Tigers” in the industry, a term that denotes emerging successful AI startups.

Traditionally, the AI market has been dominated by what is known as the "Four Dragons and Five Tigers" framework, a descriptor that divides the players into two tiers

Advertisements

On one hand, the "Four Dragons" — SenseTime, Megvii, CloudWalk, and Yitu — enjoy immense notoriety due to their advanced image recognition technologies and have attracted significant investmentThe "Five Little Tigers" refers to a newer group of unicorns, including Zhihu AI, The Dark Side of the Moon, Baichuan Intelligence, MiniMax, and Zero One Universe, all of which have emerged as formidable competitors.

In this heightened competitive environment, considerable funding is essential, often concentrating on leading-name companiesHowever, with no clear killer application prevailing in artificial intelligence, the differences in strength among various firms remain minimal, leading to a situation unique to the Chinese market that diverges from the dominance seen with firms like OpenAI abroad.

The rapid pace of innovation within the tech industry is both breathtaking and unpredictable; thus, companies must remain flexible and adaptiveThis year, the established order of the "Four Dragons and Five Tigers" has been shaken as StepStar entered the fray as a new contenderThis shift underscores the dynamism of the landscape, which can change almost overnight.

In June, the whispers about StepStar's newest funding round began circulating, with estimates valuing the company at $2 billion (approximately 140 billion RMB). This marked its official entry into the ranks of unicorns, well before this recent funding announcement brought it further recognition.

The Low-Profile Unicorn

StepStar has maintained a notably low profile compared to its counterparts in the "Five Little Tigers." To date, publicly available estimates regarding the company have been scarce

Advertisements

Even on prominent organizations like Tianyancha and Qichacha, information on its funding rounds is limited to sporadic reports.

From piecing together various snippets of information, it appears that top-tier venture capital firms like Qiming and Wuyuan participated in the initial funding rounds after the company's inceptionThere were also rumors regarding Alibaba's potential investment during the reported $2 billion evaluation round, although it remains unclear whether that materializedNonetheless, the buzz surrounding StepStar indicates a healthy interest from the capital markets.

The recent funding announcement marks the most comprehensive disclosure of information about StepStar thus far, owing much to major players like the Shanghai state-owned capital investment firm and other strategic investors like Tencent, Wuyuan Capital, and Qiming Venture Partners.

Jiang Daxin's reserved demeanor and leadership style contribute to the company's low-key approachIn a rare media dialogue, he openly identified himself as "extremely introverted," preferring to operate outside of the public spotlight.

StepStar's roots are deeply intertwined with MicrosoftJiang joined Microsoft Research Asia in 2007, kicking off a career spanning 16 years with the technology giantAs a leading figure in areas such as data mining and natural language processing (NLP), he published nearly 200 papers in top-tier journalsOver his tenure, he occupied significant roles, including vice president and chief scientist at Microsoft Research, where he spearheaded various projects related to Bing, Cortana, Microsoft Azure, and Microsoft 365.

The advent of models like ChatGPT sparked a transformative realization for Jiang, prompting him to depart from Microsoft and embark on his entrepreneurial journey.

In April 2023, StepStar was formally founded

Advertisements

Jiang's extensive experience in AI and his ability to galvanize talent facilitated the swift assembly of a team comprising experts in the fieldKey figures in the organization include Zhu Yibo, responsible for systems, and Jiao Bixing, overseeing data operationsZhu boasts a robust history with institutions like Microsoft, ByteDance, and Google, codifying his expertise in large-scale system management, while Jiao previously led Bing's core search team.

While other prominent AI firms have attracted widespread attention, StepStar's commitment to steady product development during the past year has been strategic, placing them on the periphery of the public eye despite their timely entry into the competitive field.

Reports indicate that StepStar had already garnered backing from several leading venture capital firms by early 2023, positioning itself as a strong contender in terms of funding, talent, and data acquisition compared to its peers.

In what has emerged as one of the most active years for AI large models, StepStar focused keenly on product developmentBy July 2023, the team commenced training for their modelsLess than two months later, they unveiled Step-1, a large model boasting superior performance compared to ChatGPT-3.5 with 100 billion parametersBy November, Step-1V, another 100 billion parameter multi-modal model, was also completedMarch 23, 2024, will see StepStar take the stage at the Global Developer Conference to showcase a preview of Step-2, marking a notable first for a domestic startup to present a trillion-parameter model.

When questioned by the media regarding potential anxiety stemming from the rapid advancements and hype around competing firms, Jiang succinctly conveyed his focus: "Close the door and run with all your might, and only then can you truly be free."

The recent burst of funding news in June elevated StepStar to a new level of visibility, marking its transition from obscurity to recognition.

Prior to this point, industry insiders often referred to the primary startup leaders in the domestic large model domain as the "Five Tigers." This nomenclature included Zhizhu, The Dark Side of the Moon, Zero One Universe, Baichuan Intelligence, and MiniMax — a collective of five standout firms.

With StepStar's ascent, commentators suggest that this title may soon need updating

The landscape of domestic foundational AI model startups may evolve to recognize not just five but now six significant contenders, reshaping the competitive framework.

The latest funding announcement also signals a shift in StepStar's trajectory, illuminating its potential to claim a larger share of the market.

Funding Dynamics in AI

In this funding round, StepStar has garnered interest from diverse segments of the investment landscape, showcasing the intermingling of state capital and large internet corporations.

Shanghai's state-owned capital and Tencent are among the leading investors in this round, signifying a notable trend where major players from local and national governance, alongside tech giants, are stepping into the AI sphereRecently, Zhizhu AI, one of the "Little Tigris," announced that it had also secured a substantial 3 billion RMB funding round, marking the involvement of state capital once again in key tech venturesThis reinforces the notion of a strong relationship between state funds and burgeoning AI projects.

According to reports, from January to August, there have been over 400 public announcements regarding projects won by domestic large model enterprises, with over 60% of these contracts attributed to state-owned enterprises, highlighting a robust appetite for innovation in public sector frameworks.

Additionally, the Zhongguancun Science City recently led a new fundraising round for Zhizhu AI, further illustrating the growing financing activities accessibly available to AI startups

This representation of local government investment channels emphasizes a broader strategy being employed at municipal levels to support the AI industryNotably, the Beijing municipal government's AI Industry Investment Fund has invested across a range of companies, including those mentioned earlier, revealing a focused effort to sustain a competitive AI environment.

In a closely related vein, shortly before this, AI solutions provider Wuwen Chip received significant backing from over a dozen new investors, highlighting the significance of state-owned funds in the adoption of AI technology across the region.

With major tech conglomerates such as Tencent becoming increasingly engaged in AI investment efforts, the atmosphere has poised for a dynamic competition between AI startups seeking capital and attention in a progressively crowded marketplace.

The industry landscape may evolve further, prompting a consolidation of power, but whether this will be favorable or detrimental remains uncertain.

Expert analysis suggests that the trend of state investment is likely geared towards early-stage and promising startups, making it imperative for these firms to demonstrate commercial viability in the near futureThe uncertainties associated with such investments can potentially lead to a shift where the focus may navigate away from emergent entities should they fail to sustain their trajectories of innovation and market capture.

Meanwhile, tech giants are not just passive investors; with their established AI endeavors, they may prioritize inward investment due to competition

They could choose to focus investments on bolstering their proprietary AI capabilities, especially in a time when external rivalries are intensifying.

In the rapidly changing narrative of large models, an expected pivot towards commercialization and its associated challenges will define the next stage in this battle.

Reflecting on the early part of 2023, the domestic market saw an explosive surge in large AI model launches, totaling over 100 firms going liveThis situation fostered a competitive environment where innovation flourished.

However, the challenge remains: how to navigate through such an extensive roster of models? As this landscape thins out, only a few foundational AI models will persist, moving forward to claim their positions in the market.

In conclusion, the evolving dynamics suggest that merely a couple of entities are likely to emerge successfully from this fieldThe standout players in this competitive space include tech giants like Baidu, Alibaba, Tencent, and ByteDance, alongside promising entities like the "Four Dragons" and the newly recognized "Six Tigers."

Henry Chesbrough, a prominent figure in open innovation, illustrated the concept of a "funnel" in his worksThis analogy describes how, despite an initial abundance of competition, only a select few will progress to the commercial and industrial stages of development while others gradually fade into obscurity

Advertisements

Advertisements