Sequoia Capital is making two major infrastructure bets that reveal how venture capital sees the future of AI scaling. The firm is leading a $1 billion funding round for Valar Atomics, a nuclear reactor startup seeking a $6 billion valuation, while simultaneously backing Sable, an AI employee platform that just raised $45 million. These investments, announced within days of each other, show Sequoia betting on both the physical infrastructure and software layers needed to power the next generation of AI systems.

Why Is Nuclear Power Suddenly a Venture Capital Priority?

Valar Atomics builds small modular nuclear reactors (SMRs), essentially miniaturized power plants designed to be cheaper and faster to deploy than traditional reactors. The company is three years old and based in El Segundo, California. Earlier this month, Valar demonstrated that its nuclear reactor could power an Nvidia AI chip, and the company announced a partnership with Nvidia to explore nuclear energy for future AI data centers.

The timing reflects a real infrastructure crisis. Data center electricity demands are projected to grow sharply over the next several years, but utilities in many regions are years away from adding enough new capacity. That gap has turned nuclear power, long plagued by cost overruns and regulatory delays, into one of the most closely watched corners of the AI infrastructure boom. Valar is not alone in chasing this opportunity. Other startups building next-generation reactors aimed at tech and industrial customers include Kairos Power and TerraPower, which is backed by Bill Gates, as well as NuScale Power, the only SMR developer with U.S. regulatory design approval.

Valar's technology is based on a helium-cooled, high-temperature gas reactor. The company says it eventually plans to build hundreds of SMRs to power data centers. However, while SMRs are theoretically cheaper to manufacture than traditional reactors, the technology is still nascent and it remains unclear how long it will take to be deployed at industrial scale.

The company was founded by Isaiah Taylor, who dropped out of high school at age 16. Now 27 years old, Taylor has said he launched two startups before Valar and noted that his great-grandfather worked as a nuclear physicist on the Manhattan Project. Valar counts Palmer Luckey, the Anduril founder, and Shyam Sankar, Palantir's chief technology officer, among its backers.

How Are Venture Capitalists Structuring These Mega-Rounds?

The Valar funding round illustrates a shift in how venture capital structures deals in the AI era. The startup is raising a $1 billion equity round at a $6 billion valuation, but part of that capital was raised previously at a lower valuation. Specifically, Valar raised $450 million, including $340 million in equity and $110 million in debt, at a $2 billion valuation in March 2026. Deals structured in multiple installments at varying valuations, executed at different times, are becoming increasingly common in today's AI-fueled fundraising environment.

• Multi-Tranche Structures: Investors in the same round can end up paying different prices for the same company, creating the perception of a single uniform valuation when prices actually vary across tranches.
• Valuation Progression: Valar's valuation jumped from $2 billion in March 2026 to $6 billion in the current round, reflecting rapid investor confidence in nuclear-powered AI infrastructure.
• Mixed Funding Sources: Companies are combining equity raises with debt financing, as Valar did with its $110 million debt component, to diversify capital sources and reduce dilution.

What Is Sable's AI Employee, and Why Does It Matter?

While Valar addresses the physical infrastructure powering AI, Sable is tackling the software layer. The startup announced a $45 million funding round led by Sequoia Capital and 8VC on July 16, 2026. Sable was founded in October 2025 by four Harvard University graduates whose research backgrounds span reinforcement learning, multimodal AI, and post-training. Several team members previously worked at SpaceX, Google, Meta, and Together AI. The startup has fewer than 20 employees and is already being used in production by firms like Notion and Decagon.

Sable's core product is Aidan, an AI system designed to manage customer interactions end to end. The phrase "AI employee" hides significant technical nuance. Traditional chatbots have long struggled with real-time collaboration because they do not actually interact with software interfaces. Sable's approach leans on what it calls Interactive Intelligence. The system can see the user's screen, click through interfaces, and verbally guide prospects through workflows inside a live environment. Aidan is positioned to compress qualification, demo, and onboarding stages into a single continuous interaction.

At the core of the system is LiveBox, a virtual workspace that allows Aidan to present software directly to a potential customer. The AI can answer detailed questions as it navigates, pull contextual information from internal materials, and evaluate engagement signals. If an AI assistant can detect which parts of a product resonate with prospects, sales teams can prioritize high-value follow-up conversations.

The architecture also relies on the Brain, a continuously updated internal knowledge layer. Organizations upload product documentation, sales call transcripts, marketing materials, and interviews with top performers. Over time, the AI uses this corpus to refine how it communicates. This type of iterative knowledge updating mirrors trends across other enterprise AI tools, though this implementation features a purpose-built orientation toward presales and support flows.

Why Are Enterprises Demanding AI Agents for Customer Engagement?

Industry data explains why Sable's model is gaining traction. According to IDC, spending on AI-centric systems is expected to reach $423.6 billion in 2025, with customer experience as a major driver. Meanwhile, Gartner reported that by 2026, roughly 80 percent of customer service organizations plan to use generative AI to enhance agent productivity and self-service channels.

From the buyer perspective, Forrester has noted that more than 60 percent of B2B technology buyers now prefer digital-first, self-serve engagement. Enterprise buyers frequently encounter friction during software evaluations, bouncing across qualification calls, bespoke walkthroughs, follow-up emails, and onboarding sessions handled by different employees. Each handoff increases the risk of miscommunication. Vendors such as Drift and Intercom are moving deeper into AI-based interaction, while ZoomInfo is automating prospect research with its own models. Aidan enters this market behaving less like a chatbot and more like a virtual presales engineer.

McKinsey has indicated that companies investing deeply in AI for sales and customer operations report up to 10 to 20 percent increases in sales growth and customer satisfaction. Given these shifts, Sable's push into interactive demos and real-time product guidance aligns with broader enterprise priorities.

What Governance Challenges Do AI Employees Face?

Deploying an AI system to represent a product in real time without a human in the loop requires strict governance. NIST's AI Risk Management Framework and ISO's AI lifecycle standards are increasingly referenced in enterprise deployments, and enterprise buyers expect alignment with those guidelines to manage operational risks. Some organizations will likely deploy Aidan in limited contexts initially, such as internal training or controlled demos, while fast-growing software companies may adopt it aggressively.

The involvement of partners from Sequoia Capital and 8VC gives the project additional visibility. Their presence on the board indicates that institutional investors view real-time interactive AI as a sustained market opportunity rather than a short-term automation trend.

How Do These Two Bets Reflect Sequoia's AI Strategy?

Sequoia's simultaneous backing of Valar and Sable reveals a comprehensive view of AI infrastructure. The firm is betting that the next wave of AI scaling requires not just raw computing power, but also the energy to run it and the software to make it productive. Nuclear power solves the energy bottleneck that threatens to constrain AI data center expansion. AI employees like Aidan solve the productivity bottleneck in enterprise software, allowing companies to automate complex customer interactions that currently require human intervention.

Both companies are at different stages of maturity and technical readiness. Valar's SMR technology is nascent and faces regulatory uncertainty, though the company has taken an aggressive legal stance toward its regulator. Last year, it joined several states and rival startups in suing the Nuclear Regulatory Commission, arguing the agency wrongly applies the same lengthy licensing process to small test reactors that it uses for full-size commercial plants. The case remains unresolved, with both sides repeatedly pausing litigation, which suggests some kind of settlement is in the works.

Sable, by contrast, is already in production with real customers and has a clear go-to-market path. The startup is recruiting engineers in Israel, where one of the founders noted the talent pool aligns well with the product roadmap. The $45 million raise provides the capital to build out engineering capabilities in both the United States and Israel.

As AI agents evolve from static text boxes into active participants in the customer journey, and as data center electricity demands strain existing infrastructure, Sequoia's dual bet on nuclear power and interactive AI suggests the firm sees both as essential to the next phase of AI scaling. The investments signal that venture capital is moving beyond funding AI models themselves and toward funding the infrastructure, energy, and software systems that will make those models economically viable at scale.

" }