Venture Capital & Startup Funding Roundup, August 3, 2026: Atreides, Battery Ventures, NEA, Sequoia, Point72 & More

Over the past day, investors have plowed into the underlying hardware and infrastructure needed to power the AI boom. The largest rounds are pushing capital toward real-world enablers: for example, El Segundo–based Valar Atomics raised $1 billion (Series B) at a $6 billion valuation to mass-produce modular nuclear reactors that can power data centers and AI systems. That same AI-driven theme shows up in chip funding: UK startup OLIX secured $312 million (Series B) for its photonic AI inference chips, signaling a push for specialized compute beyond today’s GPUs. Even in enterprise software, money is chasing AI agents and automation. San Francisco’s Freehand drew $75 million to expand its AI-powered supply-chain management platform, and New York’s Ellis AI raised $10 million to automate private credit workflows.

Cybersecurity is another surge point: Horizon3.ai netted $250 million (Series E) to build autonomous penetration-testing tools, reflecting rising demand for AI-driven security. Finally, niche science and frontier sectors also attracted funding – e.g., Shanghai’s Active Technology pulled about $45 million in a record angel round for its brain-computer interface platform. In sum, today’s biggest rounds underscore a clear shift: VCs are funneling capital into AI’s physical infrastructure and domain-specific applications (energy, chips, biotech, cybersecurity, etc.) rather than generic apps.

The Macro Environment: AI’s Physical Frontier

Venture funding in 2026 is increasingly concentrated on “hard” infrastructure and deep technology that support AI’s growth. Investors are signaling that the bottlenecks for the next wave of innovation lie in compute, energy, and specialized hardware. The Valar Atomics round is a prime example: “AI is building very quickly. We need a lot of power in every direction,” CEO Isaiah Taylor remarked. By backing mass-produced nuclear microreactors for data centers, Sequoia and other top investors are betting nuclear can solve AI’s energy crunch. Likewise, the OLIX funding reflects a push to augment Moore’s Law with photonic chips, boosting inference throughput using light-based interconnects. These are not small, speculative bets – the sheer size of the rounds (hundreds of millions to a billion dollars) and the presence of marquee backers show a “move-big-or-die” mindset on AI infrastructure.

Investors also see opportunity in systems that apply AI in traditional industries. The sizable check into Freehand’s supply-chain AI and Ellis AI’s credit-management platform indicate that enterprise and fintech operators with domain expertise are again drawing VC dollars. Venture firms appear to be weary of hype without results; they want technology that solves real operational pain points. Cybersecurity is a clear beneficiary: Horizon3.ai’s triple-digit million raise at a $2 billion valuation demonstrates that companies are willing to invest in autonomous security tools as cyber threats escalate. Overall, founders are being reminded that capital efficiency will be judged in the context of scale – the largest rounds are going to companies with tangible production goals (building reactors or chips) or with platforms built for critical enterprise functions.

The public markets and late-stage investments have also influenced this climate. After years of AI fever, LPs are pushing GPs to diversify beyond one or two foundational models (OpenAI, etc.), so VCs are scouting for adjacent layers like energy and robotics. Likewise, macroeconomic stability (low inflation, steady growth in 2026) is leaving some dry powder for speculative tech. But at the same time, excessive valuations earlier this year have cooled a bit, so VCs are focusing on sectors where they see clear ROI. In short, capital is flowing into the physical frontier of AI – factories, chips, power plants and specialized platforms – where the next bottlenecks lie. Founders and investors alike are taking a more pragmatic view: is your startup solving one of these concrete infrastructure gaps or mission-critical enterprise workflows? The funding patterns of the day suggest that question is on everyone’s mind.

Valar Atomics raises $1B in funding to build factory-made nuclear reactors for AI data centers

Valar Atomics develops small modular nuclear reactors designed to be built in factories and deployed like equipment for data centers or industrial sites. This $1 billion Series B funding (led by Sequoia’s Shaun Maguire) values Valar at $6 billion. Investors care because Valar claims to have achieved a major nuclear milestone – in July it powered an Nvidia AI chip for an entire web page – making it a front-runner in meeting AI’s escalating power needs. The round comes as hyperscalers grapple with the extreme energy demands of AI models. CEO Isaiah Taylor emphasizes that AI growth requires “a lot of power in every direction,” and Valar’s microreactors could deliver data-center-scale kilowatts without overloading local grids.

This raise signals that energy infrastructure is now considered an AI bottleneck worthy of tech investment. It also underlines investors’ willingness to back capital-intensive plays with long time horizons – Valar’s approach is analogous to “the nuclear answer to AI,” and the $1B round (plus a $200M credit line) gives it firepower to ramp production. Competitive analogs include Commonwealth Fusion (now well-funded) and TerraPower. But Valar’s hybrid strategy (casting reactors in factories, then shipping them) could slash costs and timelines versus traditional nuclear. With Sequoia, Point72 and other funds on board, and a planned 30 MW facility with Nvidia, Valar is positioned to be a strategic supplier to Big Tech.

Funding Details:

• Startup: Valar Atomics
• Investors: Sequoia (Shaun Maguire), Atreides Management, Point72, Snowpoint Ventures, etc.
• Amount Raised: $1,000,000,000
• Total Raised: ~$1,000,000,000 (Series B)
• Funding Stage: Series B
• Funding Date: Aug 3, 2026
• Headquarters: El Segundo, CA
• Sector: Energy Infrastructure (nuclear reactors for data centers)

OLIX raises $312M in funding to make photonic AI chips

OLIX Computing, a London-based AI hardware startup, has raised €270.5 million (≈$312 million) in a Series B at a €2.8 billion ($3.3B) valuation. The round was led by growth fund Fundomo, with new and existing backers including Arm, hedge fund Hudson River Trading, and even Reed Hastings. OLIX is building a new class of inference chip called the “Optical Tensor Processing Unit” that integrates on-chip photonic interconnects. In other words, instead of sending data over copper, these chips use light to move data between cores – potentially much faster and more power-efficient for massive AI models.

Investors are betting OLIX’s approach can leapfrog existing GPUs by tackling the “memory wall” bottleneck. Its prototype stores all data in high-speed SRAM rather than off-chip DRAM, and uses optical channels to overcome bandwidth limits. The new funds will go toward tapeout of OLIX’s chips and developing software/toolchains. By shipping a hardware-software stack (they’re hiring for compilers), OLIX is pushing to deliver chips to customers as early as late 2027. This positions OLIX as a pioneer in frontier compute hardware, analogous to Cerebras or Lightmatter. For investors, OLIX’s success would signal that photonics can become a mainstream path for AI infrastructure.

Funding Details:

• Startup: OLIX Computing
• Investors: Fundomo (lead), Arm, Hudson River Trading, Reed Hastings, Hummingbird Ventures, Crane, Plural, Creandum, Phoenix Court, Transition, etc.
• Amount Raised: $312,000,000
• Total Raised: $312,000,000 (Series B)
• Funding Stage: Series B
• Funding Date: Aug 3, 2026
• Headquarters: London, UK
• Sector: Semiconductor / AI Hardware (photonic inference chips)

Horizon3.ai raises $250M in funding to automate security testing

Horizon3.ai, a San Francisco-based cybersecurity startup, has closed a $250 million Series E led by NEA and NightDragon. This new funding more than triples the company’s valuation (now above $2 billion) since its $650 million Series D last year. Horizon3 offers “autonomous penetration testing” – essentially AI-driven network scanning and exploit simulation – aiming to fully automate pentesting instead of relying on occasional human red teams. With 7,200 customers (roughly half of US federal agencies use Horizon3) and ~$100 million ARR, investors see a huge market.

Companies care because Horizon3’s NodeZero platform can continuously probe live systems safely. With threats and attack surfaces multiplying, clients want more than annual manual tests – they want on-demand scans and constant coverage. The rise of AI has also generated fresh security concerns (security labs breached open models recently), so tools that proactively find vulnerabilities are in higher demand. The round underscores the trend that cybersecurity itself is tech-hungry and enterprise buyers will fund automated solutions.

Funding Details:

• Startup: Horizon3.ai
• Investors: NEA (lead), NightDragon, Qualcomm Ventures, SAIC, EDBI (Singapore), and others.
• Amount Raised: $250,000,000
• Total Raised: $250,000,000 (Series E)
• Funding Stage: Series E
• Funding Date: Aug 3, 2026
• Headquarters: San Francisco, CA
• Sector: Cybersecurity (AI-driven penetration testing)

Freehand raises $75M in funding for AI supply-chain automation

Freehand (formerly “ScopeAI”) in San Francisco announced a $75 million round co-led by Battery Ventures and NewRoad Capital. The startup provides AI agents and orchestration for enterprise procurement and supplier management – essentially automating complex, rules-driven spend processes across departments. Freehand’s platform reads invoices, contracts, and workflows to guide purchase decisions and enforce compliance.

Investors backed this round because supply-chain spend remains a huge pain point. Enterprises spend billions annually, yet legacy systems are fragmented. Freehand’s customers reportedly include Fortune 500 companies, suggesting the product resonates. Battery’s Chamath Palihapitiya remarked that AI agents like Freehand’s are “the future of work”, hinting at growth ambitions. The round gives Freehand a war chest to expand beyond procurement into adjacent areas (e.g., outbound payments, inventory) and to scale its account executive team.

Funding Details:

• Startup: Freehand
• Investors: Battery Ventures, NewRoad Capital (co-leads), plus PSP Growth, Nexus, and 1kx.
• Amount Raised: $75,000,000
• Total Raised: $75,000,000 (undisclosed round)
• Funding Stage: Growth Equity (post-seed/Series?)
• Funding Date: Aug 3, 2026
• Headquarters: San Francisco, CA
• Sector: Supply Chain / Enterprise Software (AI-driven procurement)

Active Technology raises nearly $46M in funding for brain-computer interfaces

Shanghai’s Active Technology announced a record RMB 330 million (≈$46 million) angel funding round, the largest seed/angel ever in China’s BCI sector. Active is developing non-invasive brain-computer interface (BCI) devices for direct human-AI interaction. The capital came from VC 360Xi, VCs including 54 Capital, plus corporate investors like Lenovo Ventures and 360’s co-founder Zhou Hongyi.

Investors are excited by the convergence of AI and biotech. Active’s devices aim to let users control digital systems or augment cognition via thought, and it already has product trials with sports and gaming groups. This round will fund human trials and supply chain ramp-up. With companies like Neuralink and Numenta making headlines, Active’s focus on consumer/enterprise BCI gets attention as a possible future computing platform. Strategically, this shows money chasing next-gen interfaces; if consumer AR/VR and AI meet physical reality, BCI could be a frontier.

Funding Details:

• Startup: Active Technology
• Investors: 360Xi Ventures, 54 Capital, Lenovo Industry Fund, Yunqi Partners, IDG Capital, among others.
• Amount Raised: ~$45,700,000 (RMB 330 million)
• Total Raised: ≈$46 million (angel)
• Funding Stage: Angel
• Funding Date: Aug 3, 2026
• Headquarters: Shanghai, China
• Sector: Frontier Tech (Brain-Computer Interfaces)

P-1 AI raises $50M in funding to build an “AI mechanical engineer”

P-1 AI, a San Mateo startup founded by ex-DeepMind scientists, announced the initial close of a $50 million Series A led by NEA. Its product is called Archie, an AI agent that can autonomously perform mechanical and electrical design tasks for industrial engineering. Essentially, P-1 is training AI to replace or assist human engineers by understanding CAD, schematics, and industrial requirements.

Investors are excited by the unique angle: “P-1 is building an AI mechanical and electrical engineer,” NEA partner Reed Sturtevant said. With shortages of skilled engineers in manufacturing, a viable AI assistant has huge promise. The $50M will accelerate R&D and pilot deployments (P-1 had closed a $23M seed last year). The raise also underscores that even specialized deep tech (here combining robotics, AI, and industrial design) can attract venture dollars if execution looks strong. Strategic implications include disrupting industrial design workflow and possibly integrating with robotics suppliers.

Funding Details:

• Startup: P-1 AI
• Investors: NEA (lead), with participation from Atlantic Bridge, Founders Fund, Radical Ventures, Gula Tech Adventures, and others.
• Amount Raised: $50,000,000 (initial Series A close)
• Total Raised: ~$73,000,000 (including prior $23M seed)
• Funding Stage: Series A
• Funding Date: Aug 3, 2026
• Headquarters: San Mateo, CA
• Sector: AI / Industrial Automation (AI agent for mechanical engineering)

Balance Theory raises $19M in funding for AI-driven security budgeting

Balance Theory, a Columbia, MD–area startup, announced a $19 million Series A funding led by SYN Ventures. Its platform helps CISOs and security teams optimize their budgets and vendor portfolios using AI. For example, it ingests all an enterprise’s security products, spend, and risks into a unified model, then uses machine intelligence to recommend where to reallocate funds for maximum ROI. Existing investors DataTribe and TEDCO also participated.

The raise highlights growing investor attention on cybersecurity spend optimization. Balance Theory claims its platform manages over $1 billion in security spend today, with customers seeing massive cost savings. With security toolsets proliferating, CISOs often have hundreds of products – this startup promises a “north star” to navigate those choices. Investors see defensive value: better budget decisions could avoid costly breaches. While $19M is modest, having a dedicated funding round and experienced backers (SYN Ventures’ Alex Tosheff joined the board) shows confidence. This round will fund product expansion (e.g., adding automated procurement tools) and sales growth.

Funding Details:

• Startup: Balance Theory
• Investors: SYN Ventures (lead), DataTribe, TEDCO.
• Amount Raised: $19,000,000
• Total Raised: $19,000,000 (Series A)
• Funding Stage: Series A
• Funding Date: Aug 3, 2026
• Headquarters: Columbia, MD (suburb of Baltimore)
• Sector: Cybersecurity (AI-driven investment/portfolio management)

Ellis AI raises $10M in funding to automate private credit deals

Ellis AI, a New York fintech startup, emerged from stealth with a $10 million seed led by First Round Capital, 645 Ventures and others. Its goal is to streamline the cumbersome back-office of private credit firms (secondaries, direct lending). The Ellis platform uses generative AI agents to aggregate spreadsheets, documents and market data – for example, it can help close a fund’s books at month-end by automatically extracting and reconciling records.

The funding – backed by names like Khosla Ventures and Thrive – underscores venture interest in vertical AI tools for finance. Private credit has ballooned over the decade, but deal processing still relies on manual workflows and Excel. Ellis attacks this pain point, promising clients faster reporting and due diligence. Its founder Ryan Williams, known for fintech success with real-estate platform Cadre, already had credibility with investors. Strategic implications: if Ellis’s agents work well, they could expand to other opaque asset classes. The $10M will grow the engineering and sales teams, as the startup pilots with credit managers.

Funding Details:

• Startup: Ellis AI
• Investors: First Round Capital, 645 Ventures, Harlem Capital, Khosla Ventures, Thrive Capital, Slow Ventures, Mellody Hobson (Ariel Investors), etc..
• Amount Raised: $10,000,000
• Total Raised: $10,000,000 (Seed)
• Funding Stage: Seed
• Funding Date: Aug 3, 2026
• Headquarters: New York, NY
• Sector: Fintech / Enterprise AI (private credit automation)

FAST Metals raises $4.3M in funding to recycle mining waste

FAST Metals (Stamford, CT) closed a $4.3 million pre-seed round led by New Climate Ventures. The startup has developed a process to extract valuable minerals (like rare earths and aluminum) from “red mud” – the toxic iron-rich waste left from bauxite ore processing. As part of the raise, FAST Metals announced its first commercial deal: it will process red mud at Metalox’s plant in Florida, proving the technology at scale.

This funding round is smaller but strategically important: it points to renewed VC interest in clean mining and circular economy solutions. Investors include mining-industry insiders (Azolla Ventures, Glencore’s Astorg family office) and even Rio Tinto’s Founders Factory, reflecting belief that industrial AI/technology can reduce waste. For FAST Metals, the capital will scale its pilot plant and accelerate partnerships. In broader terms, this shows that beyond AI software, VCs are also banking on startups that make existing industries greener and more efficient.

Funding Details:

• Startup: FAST Metals
• Investors: New Climate Ventures (lead), Azolla Ventures, Humba Ventures, Astor Swiss (with Glencore ties), Rio Tinto/Founders Factory, etc..
• Amount Raised: $4,300,000
• Total Raised: $4,300,000 (Pre-seed)
• Funding Stage: Pre-Seed
• Funding Date: Aug 3, 2026
• Headquarters: Stamford, CT
• Sector: Cleantech / Mining (mineral recovery from industrial waste)

What Today’s Funding Activity Reveals

The clustering of today’s rounds reveals two clear patterns. First, infrastructure-oriented AI plays are commanding attention. Valar Atomics and OLIX alone account for over $1.3 billion of today’s financing – highlighting that power and compute underpinnings of AI are now premium bets. Even smaller rounds like FAST Metals tie into this theme (it aims to secure materials for technology in a sustainable way). In short, investors are treating AI’s physical needs as core sectors, akin to funding data centers or power grids.

Second, domain-specific AI applications remain a priority. Freehand and Ellis, for example, raised meaningful rounds by automating established enterprise workflows (supply chain, finance) using AI agents. Horizon3 and Balance Theory demonstrate that even cybersecurity is being reframed around AI-driven efficiency. These rounds suggest VCs are focused on startups that apply AI to concrete industry problems with clear ROI, rather than chasing broad platform concepts.

Geographically, the funding split is also telling: a big chunk is U.S.-centric (Valar, Horizon3, Freehand, Ellis, P-1, Balance, FAST), but we also see the globalization of these trends. OLIX shows Europe’s investors championing AI chip innovation, while Active Technology highlights China’s interest in next-gen AI interfaces. Even domain-crossing – a nuclear reactor startup in the U.S. solving data-center power – signals that we’re seeing cross-pollination between tech and other industries (energy, heavy industry, biotech).

Finally, the investor base provides context: top-tier venture firms and even strategic corporates (Arm, Lenovo, Qualcomm, Rio Tinto, Nvidia partnerships) are prominent. This suggests a convergence of startup finance with traditional industrial players, especially where technology and hard assets meet. In sum, today’s fundraising hints at an era where venture capital is flowing into the nuts and bolts of the digital economy – the chips, the electrons, and the AI-powered processes – rather than solely into consumer or social apps.

Venture Funding Table

*Humigent’s amount was not disclosed; investors participated in a Series A close.

Strategic Takeaways for Founders and Investors

For founders: Focus your startup on solving deep, high-value constraints in AI and industry. Today’s biggest deals suggest the market rewards companies addressing things like compute bottlenecks (new chips, power sources), automation of critical workflows (finance, procurement), and essential safety/defense needs (security, autonomous testing). Demonstrate real production or customer traction (Valar’s reactor demo, Horizon3’s revenue growth, etc.) – investors are skeptical of vague AI plays without physical results. Also note the premium on domain expertise: Freehand’s supply-chain domain knowledge and Balance Theory’s cybersecurity focus helped them raise serious capital. If you’re founding an AI startup in 2026, the bar is high: you need to show how your tech fills a tangible infrastructure or enterprise gap, not just how it uses AI.

For investors: The dominant theme is “AI infrastructure.” Allocate capital not only to apps but also to the underlying layers – chips, data-center hardware, power generation, industrial AI agents. We’re seeing a capital concentration around a few large opportunities (Valar and OLIX each raised hundreds of millions), so competition for those deals is intense. Smaller rounds (Ellis, Balance, FAST) indicate niche spaces where disciplined investors can find value, but large VCs are clearly focused where every extra percentage of efficiency or innovation compounds at scale. Also watch geography: European and Asian investors are also pushing these themes, so don’t neglect high-potential startups outside the US. Evaluate investments by looking at end-market size (e.g., the multi-trillion-dollar data-center market) and strategic synergies (Nvidia partnering with Valar, for example). Finally, be mindful of valuation signals: when a startup jumps to unicorn status on a new round (Horizon3’s 3x re-valuation), consider whether that premium is justified by unique tech or just hype. In general, the infrastructure “arms race” suggests a two-tier strategy: anchor positions in big bets (AI energy, hardware) and selectively back efficient niche plays (vertical AI software) where returns can be sharp.

Conclusion

Today’s funding news paints a picture of venture capital that is no longer content with virtual hype alone. Instead, the flow of capital is tracing where AI’s rubber meets the road – in energy plants, chip fabs, and enterprise operations. Founders building physical enablers for the AI revolution are commanding the largest checks, while those applying AI to core industries (from mining to cybersecurity to finance) are also attracting interest. The upshot is clear: investors see the next wave of value in solving real-world constraints on AI growth. As a result, the ecosystem is tilting toward deep tech and industrial AI, suggesting the startup world is entering a phase where getting the underlying machinery right matters as much as the software. This shift could redefine the startup landscape over the coming months. Founders and VCs who recognize this trend – and orient their strategies accordingly – will likely have the edge in a market where capital chases the foundations of tomorrow’s technology.