By | Alpha Community
An AI startup founded at the end of 2025, which hasn't even publicly launched a product yet, has secured $700 million in Series A funding, propelling its valuation to $6 billion. This round was led by Parkway Venture Capital, with participation from NVIDIA, AMD Ventures, Intel Capital, Qualcomm Ventures, and Salesforce Ventures.
It's evident that this company has raised a massive amount of capital in a short time and has received backing from the industry's top software and hardware tech giants.
This company, named Hark, has a clear focus: they aim to create the next generation of universal human-computer interface using a combination of "self-developed foundational models + custom hardware."
In essence, this is a new type of artificial intelligence interface. Its form is AI-native hardware, which can be broken down into a series of customized native hardware devices and computing devices equipped with agent capabilities, featuring end-to-end voice models and highly personalized memory. All these AI systems are multimodal, capable of understanding and interacting in natural ways.
When we saw Hark complete its funding round at a $6 billion valuation with simultaneous investment from NVIDIA and Qualcomm, we weren't surprised. Since 2024, Alpha Community has been strategically investing in the "proactive AI" direction—our early investment in Looki has already sold multimodal AI wearable devices to users worldwide, making it the world's highest-volume seller of multimodal wearable general intelligent devices; LightSail Technology has independently developed a native AI operating system for smart hardware and pioneered the category of AI headphones with visual perception capabilities.
Hark's massive funding round once again confirms an increasingly obvious trend: the next decade of AI is not just on screens, but even more so in the real world.
AI Has Gotten Smarter, But It's Still Using Old Shells and Interaction Methods
Hark was founded by Brett Adcock at the end of 2025, initially with a $100 million personal investment from him. Brett Adcock has previously founded companies like Archer, Figure, and Vettery.
Among them, Archer entered the electric vertical takeoff and landing (eVTOL) aircraft market and successfully went public. Figure is a humanoid robotics company. In 2024, Figure raised $675 million, and in September 2025, Figure completed over $1 billion in Series C funding, reaching a valuation of $39 billion. Its investors include Jeff Bezos, NVIDIA, Microsoft, OpenAI, etc.
Why would Brett Adcock want to start a company in the direction of "proactive" AI-native hardware? Because Figure's own path is essentially a systems engineering project of "AI + hardware + real-world interaction." Fundamentally, this shares a similar underlying technology stack with AI-native hardware; he knows where the pitfalls are. And recently, Figure's live demonstration of robots performing long-duration package sorting tasks indicates they have already solved some of these problems.
In addition to Brett Adcock himself, Abidur Chowdhury has joined Hark as Head of Design. He was formerly a product design executive at Apple, involved in the design of products like the iPhone and Air. Hark has also attracted engineers from Apple, Meta, Google, Tesla, and leading AI labs to join, covering AI research, hardware engineering, and design.
Looking back at the history of personal hardware terminal development, it's actually a history of alternating advances in hardware form factors, interaction methods, and applications: hardware form factors and interaction methods evolve, giving rise to new applications, unlocking new capabilities, and spreading to a broader user base.
For example, when the PC form factor was established and its size became sufficiently small, coupled with the maturity of interaction interfaces like the mouse and GUI, it became easier for ordinary people to use. And with the proliferation of the internet, it moved from business and creative professional circles into the mass market.
The next breakthrough came with the iPhone. This breakthrough was not only about integrating computer and phone capabilities into a very small form factor but also about the multi-touch interaction method, which further lowered the interaction barrier, elevating the user base of smartphones (including tablets) by an order of magnitude compared to PCs.
Moreover, its App Store ecosystem directly became the software standard of the mobile internet era. In 2024, the global App Store ecosystem facilitated approximately $1.3 trillion in developer billings and sales.
Now, the problem with AI is that it has intelligence, strong software capabilities, but currently primarily operates and interacts through chat interfaces and non-AI-native devices like computers/phones. It lacks persistent memory of user identity and hardware specifically designed for intelligent interaction.
A preliminary industry consensus is that the next stage requires intelligent agent systems that can naturally interact with people and the real world. These systems need to anticipate needs, reduce cognitive load, and operate like collaborative partners rather than waiting for commands like traditional software.
Currently, at the software level, AI has already spawned super-startups like OpenAI and Anthropic with valuations approaching a trillion dollars. Once AI-native hardware further develops, its impact on the tech industry is likely to be on the scale of the iPhone.
However, for "proactive" AI-native hardware to mature is a complex engineering task. For instance, Hark needs to build this entire system across layers like models, AI hardware, interaction, and memory.
First, their models will possess agent capabilities, multimodal capabilities, and memory capabilities, able to remember who the user is, what they've said, and work across products and services the user already uses.
They will design AI-native hardware, integrated with Hark's foundational models. And judging from their recruitment for real-time voice infrastructure positions, their interaction interface is likely to start with voice.
Developing "Proactive" AI: Chinese Startups Hold an Advantage
Current AI, whether ChatBots or Agents, are temporarily just tools because they are trapped on screens. Only when people need them do they issue commands and get results.
Compared to these "passive" AIs, why is "proactive" AI important? Because it transforms AI from a tool into a collaborator. AI can, to a certain extent, independently think for people, act for people, and accomplish tasks.
Building a "proactive" AI system requires an AI-native hardware that combines software and hardware. It needs to possess perception, memory, intelligence, new and lower-barrier interaction methods, and needs to be always by the user's side (always on).
During the previous exploration period for AI hardware (e.g., smart speakers, etc.), it had perception but could only store, intelligence was insufficient, and interaction was rigid.
In the current new AI boom period, perception capabilities have further improved, AI memory and intelligence have made huge leaps, interaction is still being explored, but the exploration path has become initially clear.
True "proactive" AI has already taken a big step forward.
And for "proactive" AI to mature further, it's not a race for single-point breakthroughs. It requires the joint development and progress of foundational models, agent operating systems, personalized memory, and hardware terminals. The competition in AI-native hardware is a comprehensive one.
In innovating and exploring this field, Chinese startups have a greater chance of success. They possess three unique advantages: first, the manufacturing ecosystem advantage, with places like Shenzhen having the world's most complete supply chain foundation; second, the market scale advantage, as China is both the largest manufacturing country and the largest application market; third, the policy support advantage, with the country listing AI as a strategic priority, providing certainty for long-term investment.
