Meet the New Generation of AI Disruptors

The next wave of the AI era is being built not by incumbents but by a new generation of companies. These disruptors are developing physical AI and software AI applications that support the advances in AI-based perception, manipulation, and real-time decision making that are being adopted in construction, agriculture, health care, heavy industry, and for other uses. These emerging AI leaders are also attracting talent and capital at a pace that reflects a broader recognition—that the window for defining the AI era’s foundational platforms, tools, and business models is open now, and closing fast.

A new BCG analysis of 1,000 US private companies in software and physical AI, computing hardware, and quantum computing finds that these emerging disruptors share at least one of these attributes: They make autonomous systems capable of performing end-to-end work. They redesign hardware and software to overcome existing AI constraints. Or they turn previously unmeasurable data into intelligence that can be embedded directly into how companies operate.

We delved into what each factor looks like in practice, and compiled examples of disruptor companies that are successfully executing on them.

What’s Driving the New Generation of AI Disruptors

AI innovations are advancing so fast they’ve outpaced the capabilities of existing computing power, memory, and energy—limitations that are impeding AI models’ ability to scale. From 2018 to 2025, growth of AI model capability outpaced growth of computing throughput by 500,000 times. (See Exhibit 1.)

Addressing that constraint has moved hardware back into the spotlight as a structural necessity and renewed source of value in the tech stack.

The convergence of physical and software AI is still new. But there is no doubt it is where tech industry capital and talent are concentrating. Among the top 100 US private companies, physical AI and robotics, computing hardware, and quantum computing now rank alongside software as representing some of the largest value pools. (See Exhibit 2.)

Factors That Set Emerging AI Disruptors Apart

To understand what has made the emerging disruptors so attractive to investors and customers, we analyzed US companies in physical AI, computing hardware, and related sectors across four variables: vision, business model, novel technology, and leadership. (See “How We Identified Emerging AI Disruptors.”)

We found that disruptors embody at least one of three factors that set them apart from competitors.

Factor 1: Shifting from Assistance to Autonomous Execution

Disruptors productize end-to-end delivery, focus business models on outcomes, and cultivate domain-specific expertise.

In earlier eras that ushered in new forms of automation, fixed rules replaced routine tasks. ATMs, for example, processed standard banking transactions. But people still handled anything that required a judgment call. Autonomous AI crosses that threshold, with systems that can make decisions, handle exceptions, and complete end-to-end work, as shown in the example of how Sierra’s autonomous AI agents respond to customer inquiries. (See Exhibit 3.)

This change is affecting tech business models. With SaaS, value scaled as customers added seats and logins. AI execution platform business models flip that. If fewer users need to interface with autonomous systems, value scales as the volume and quality of outcomes improves. In recognition of this, some disruptors offer outcome-based pricing. They charge for resolved service cases, successful placements, or other results. In some cases, the companies effectively assume responsibility from their customers for delivering results.

Disruptors don’t rely on general AI capabilities alone to build a durable “moat,” a long-term, sustainable competitive advantage. They convert that capability into domain-specific execution systems that combine industry context, operating logic, and real-world judgment so customers can delegate tasks to the systems with confidence. Then they use that domain specificity and expertise to differentiate themselves. Some of the disruptor companies that illustrate this are Sierra, Nimble, and Basis.

Factor 2: Re-architecting Around System Bottlenecks

Disruptors identify critical limitations and redesign entirely new systems that circumvent them.

Previous eras of computing innovation scaled when governing restrictions were removed. The cheaper, more durable transistors that replaced vacuum tubes miniaturized electronics, paving the way for the digital age. Cloud computing removed the need to own hardware in order to build and deploy software.

Modern disruptors remove bottlenecks created by insufficient computing power, memory, or energy, or by fragmented information trapped inside corporate software and systems that can’t communicate with each other, which keeps companies from acting on data they already own. Disruptors aren’t building cars that go faster. They’re building the equivalent of an entirely new transportation system. By implementing changes that lead to significant reductions in cost, latency, and decision-making speed, they aren’t just winning market share in existing categories, they are creating new ones, as exemplified by Lightmatter in the computer chip sector. (See Exhibit 4.)

These companies address areas where constraints are architectural and shared across industries. In addition to Lightmatter, some of these disruptors include d-Matrix, Glean, and QuEra.

Factor 3: Embedding Unstructured Data into Customer Operations

Disruptors transform underused data into intelligence for day-to-day workflows.

In earlier tech eras, data was structured in ways that were relatively simple to aggregate and monetize. Nielsen turned TV viewership into commercial commodity. Bloomberg converted fragmented financial data into market intelligence.

In the past quarter century, people have generated massive amounts of data that is unstructured, from video, audio, wearables, internet-of-things devices, and other digital sources. Because unstructured data is so complex, it hasn’t been as easy to interpret at scale, which has limited how financially feasible it’s been to aggregate.

AI changes that. AI can interpret previously unmeasurable signals, extracting meaning from video, speech, physical interactions, and tacit expertise—and doing it at scale. Merely providing access to historically unattainable data is not enough of a competitive advantage, though. Disruptors turn unstructured data into intelligence and embed it into customer workflows in ways that, combined with usage, become difficult to replace, as the example of ElevenLab’s voice AI illustrates. (See Exhibit 5.)

Emerging AI technology leaders build interpretation layers that sit across industries and embed them directly into workflows. In addition to ElevenLabs, other disruptors capitalizing on this ability include Twelve Labs, Carbon Robotics, and World Labs.

How to Capitalize on AI Disruptors

As the market for physical and software AI applications expands, it’s creating different opportunities for stakeholders in and outside of the tech industry.

For emerging disruptors.

Disruptors’ successes to date are proof that AI-based technologies and workflows are taking hold across a variety of industries. To compete, determine if your offerings go deep enough in a specific domain or are embedded far enough into specific workflows that they couldn’t easily be replaced by general-purpose alternatives. Know what small wins you need to accumulate that would produce momentum over time, including aligning product engineering, go-to-market, customer success, and value realization. Remove roadblocks that could keep enterprise customers from buying, including their expectations and requirements for reliability, quality, and compliance. If you’re considering partnering with existing companies that sell to enterprise customers, understand the ecosystems they operate in and take steps to build relationships with them.

For incumbent tech companies.

The threat that disruptors pose isn’t just from the novelty or superiority of their products or business models. Disruptors are designing new ways of operating and getting things done, which could render existing systems and workflows outdated or irrelevant. To remain competitive in an era of hypergrowth, update your usual product development cadence. Launch fast-moving internal “skunkworks” initiatives with the goal of accelerating innovation in specific target areas. It’s better to disrupt yourself than be disrupted by emerging competitors.

For enterprises.

For companies adopting AI, disruptors represent ways of working that were not possible before. If you’re considering using a disruptor, engage with them initially to understand what their solutions are and the value those apps, tools, or services could create for your organization. Do a small, proof-of-concept project to understand what you’d need to adopt and integrate it into your workflow. If all goes well, move quickly to full-scale deployment so you start to realize measurable value as soon as possible and are not spending time on POCs unnecessarily. If after deployment your operational performance metrics are not leading the industry, it’s a sign that you haven’t taken AI-enabled disruptive solutions far enough.

Disruptors signal where value creation is heading. Players defining the next phase of the AI era are building it now, before the window of opportunity closes, and others would be wise to follow their lead.