
Belfort Labs, a KU Leuven spin-off developing hardware accelerators for fully homomorphic encryption (FHE), closed a $6 million seed funding round led by Vsquared Ventures. The round attracted a diverse investor base, including institutional VCs like Anagram, Protocol VC, Inovia Capital, Syndicate One, Prototype, and Credibly Neutral, alongside prominent angels such as Jeff Dean (Google’s Chief Scientist) and Naval Ravikant.
Belfort’s latest seed round marks a pivotal moment for the company, enabling it to transition from research prototypes to commercial hardware solutions. Founded in 2025 as a spin-off from KU Leuven’s COSIC lab, Belfort addresses a critical gap in secure computing by accelerating encrypted data processing without decryption, targeting high-stakes industries where data privacy is paramount.
The $6 million seed round was announced alongside the launch of Belfort’s first hardware accelerator, designed specifically for encrypted compute. This funding comes at a time when FHE technologies are gaining traction due to increasing regulatory pressures (e.g., GDPR, HIPAA) and high-profile breaches in finance and healthcare. The round’s structure reflects strong validation from deep-tech investors, with a focus on hardware innovation rather than software-only solutions.
| Aspect | Details |
| Round Type | Seed |
| Amount Raised | $6 million |
| Lead Investor | Vsquared Ventures (Munich-based deep-tech VC focused on quantum, AI, and novel computing) |
| Key Participants | Anagram, Protocol VC, Inovia Capital, Syndicate One, Prototype, Credibly Neutral; Angels: Jeff Dean (Google), Naval Ravikant |
| Valuation | Not publicly disclosed |
| Use of Proceeds | Double team size (hiring in chip design, cryptography, and software integration); Advance custom chip development; Launch enterprise pilots in San Francisco and Leuven offices |
Investor Analysis
The investor syndicate blends European deep-tech expertise with U.S. AI and crypto heavyweights, signaling Belfort’s cross-Atlantic ambitions. Vsquared Ventures, as lead, brings sector-specific knowledge, having backed similar hardware plays in quantum and AI infrastructure. Angels like Jeff Dean and Naval Ravikant add credibility—Dean’s AI leadership at Google aligns with Belfort’s potential in encrypted AI models, while Ravikant’s crypto influence ties into Belfort’s blockchain use cases.
| Investor | Type | Notable Focus/Portfolio Highlights | Relevance to Belfort |
| Vsquared Ventures | VC (Lead) | Deep-tech (quantum, AI/ML, robotics); Backed companies like IQM Quantum Computers and SEMI Technologies | Expertise in hardware acceleration for novel computing paradigms |
| Anagram | VC | Early-stage tech with social impact; Investments in AI ethics and privacy tools | Supports privacy-focused innovations in regulated sectors |
| Protocol VC | VC | Web3 and blockchain infrastructure; Portfolio includes Chainalysis and Polygon Labs | Aligns with Belfort’s crypto/DeFi applications |
| Inovia Capital | VC | Canadian growth-stage tech; Backed Lightspeed Commerce and Hopper | Provides scaling experience for enterprise pilots |
| Syndicate One | VC/Angel Network | Crypto and AI; Founded by Belfort COO Laurens De Poorter; Investments in Kraken portfolio companies | Insider connection via co-founder, boosting operational synergies |
| Prototype | VC | European deep-tech; Focus on hardware and sustainability | Aids EU-based R&D and chip fabrication |
| Credibly Neutral | VC | Neutral tech investments; Portfolio in secure systems | Emphasizes inherent security in Belfort’s architecture |
| Jeff Dean (Angel) | Individual | Google’s Chief Scientist; Shaped TensorFlow and AI infrastructure | Validates technical feasibility for AI-encrypted compute |
| Naval Ravikant (Angel) | Individual | Angel investor (AngelList co-founder); Backed Uber, Twitter, and crypto projects like Solana | Brings network effects for blockchain adoption |
Strategic Implications
This funding positions Belfort to capture early market share in a projected $10+ billion FHE market by 2030, driven by AI’s data demands. Early pilots, such as a proof-of-concept with SWIFT for secure financial messaging, demonstrate traction. However, challenges like high energy costs for FHE and competition from software libraries (e.g., Microsoft’s SEAL) persist. The dual-office setup (Leuven for R&D, San Francisco for go-to-market) leverages EU grants (e.g., ERC Advanced) and U.S. talent pools.
Belfort’s team—led by CEO/CTO Michiel Van Beirendonck (ZPrize winner, ex-Microsoft Research), COO Laurens De Poorter (ex-Google[x], Kraken), Chief Scientist Ingrid Verbauwhede (KU Leuven professor, DARPA collaborator), and Engineering Lead Furkan Turan (ex-Intel)—combines academic rigor with industry execution. Their patented streaming architecture reduces FHE computation times from hours to minutes, a key differentiator.
Belfort Labs emerged from KU Leuven’s storied COSIC lab in April 2025, rapidly evolving from theoretical research in fully homomorphic encryption (FHE) to a hardware-centric startup addressing one of cybersecurity’s holy grails: processing sensitive data without ever exposing it in plaintext. The company’s latest $6 million seed round not only funds this pivot but also coincides with the unveiling of its inaugural hardware accelerator—a purpose-built chip that slashes FHE latencies by orders of magnitude, making real-time applications feasible in domains like national security, finance, healthcare, and blockchain.
Historical Context and Evolution
Belfort’s roots trace to DARPA’s DPRIVE program and European Research Council (ERC) Advanced Grants awarded to Prof. Ingrid Verbauwhede, whose cryptographic hardware expertise has influenced projects at Intel, Google, and Stanford. Co-founders Michiel Van Beirendonck and Furkan Turan, both KU Leuven PhDs, developed the core streaming architecture during their tenure at COSIC—Van Beirendonck’s innovations earned the inaugural ZPrize for FHE acceleration, while Turan’s work included patents in hardware protection and FPGA optimizations from Intel internships. Laurens De Poorter, with his MBA from Harvard and stints at Google[x] and Kraken Ventures, joined to bridge tech commercialization and crypto ecosystems, founding Syndicate One to nurture similar ventures.
Prior to this round, Belfort operated on seed grants and lab funding, incorporating in Leuven (VAT: BE1021971016) with a lean team of six. The spin-off status provided IP ownership, including Van Beirendonck’s patented architecture, which enables “streaming” FHE operations to minimize memory overhead—a breakthrough over traditional batch-processing methods that bottleneck scalability.

Recommended: Why NetClass’s Expansions Signal A New Chapter In EdTech
Round Mechanics and Terms
Announced via PRNewswire and covered extensively in tech outlets, the seed round was oversubscribed, reflecting investor enthusiasm for hardware solutions in an AI-driven world where data breaches cost $4.88 million on average (IBM 2025 report). While valuation remains undisclosed—typical for early seeds in Europe—the lead from Vsquared Ventures suggests a pre-money figure in the $15-25 million range, based on comparable deep-tech deals like those in quantum computing.
The syndicate’s composition is noteworthy: Seven institutional VCs provide diversified geographic and thematic coverage, while angels inject star power. Vsquared, managing €130 million across two funds, led with a likely anchor investment of $2-3 million, emphasizing their thesis on “technologies advancing society” through novel compute. Anagram and Protocol VC add U.S. Web3 angles, Inovia brings Canadian scaling muscle, and European players like Prototype and Syndicate One (De Poorter’s firm) ensure regional alignment. Credibly Neutral, focused on verifiable tech, rounds out the institutions.
Angels Jeff Dean and Naval Ravikant participated via personal checks, estimated at $250,000-$500,000 each. Dean’s involvement—given his role in Google’s FHE consultations—signals potential synergies with cloud providers, while Ravikant’s crypto pedigree (backing 100+ startups) positions Belfort for DeFi pilots.
| Investor Portfolio Overlap | Example Investments | Synergy with Belfort |
| Vsquared Ventures | IQM (quantum hardware), SEMI (AI chips) | Hardware acceleration expertise |
| Anagram | Privacy tools like Signal forks | Ethical AI and data protection |
| Protocol VC | Chainalysis (blockchain security), Polygon | Confidential smart contracts |
| Syndicate One | Kraken-backed crypto infra | Operational ties via De Poorter |
Allocation and Milestones
Proceeds are earmarked across three pillars:
- Talent Acquisition: Doubling from ~6 to 12-15 employees in 12-18 months, prioritizing chip designers (e.g., ASIC/FPGA specialists), cryptographers, and go-to-market roles. Recruitment targets U.S. and EU hubs, leveraging KU Leuven’s talent pipeline.
- R&D Acceleration: Finalizing a custom ASIC chip, building on FPGA prototypes that already achieve 10-100x speedups over CPU-based FHE. This includes integration with libraries like OpenFHE and testing against benchmarks from the ZPrize competition.
- Commercial Traction: Rolling out pilots with enterprises—e.g., SWIFT PoC for encrypted transaction validation—and expanding use cases. In finance, this enables real-time fraud detection on ciphertext; in healthcare, collaborative genomics without data sharing; in government, secure intel analysis.
Belfort’s tech stack emphasizes “confidentiality by design”: Data remains encrypted end-to-end, mitigating insider threats and cloud vulnerabilities. Performance metrics from demos show operations that once took hours now completing in minutes, with scalability for petabyte-scale datasets.
Market Positioning and Competitive Landscape
The FHE market, valued at $120 million in 2024, is forecasted to reach $12 billion by 2030 (MarketsandMarkets), fueled by AI’s privacy needs—e.g., training models on encrypted patient data. Belfort differentiates via hardware: Competitors like Microsoft’s SEAL or IBM’s HELib are software-only, suffering 1,000x slowdowns; hardware peers (e.g., Cornami’s ARES chip) focus on older schemes like CKKS, while Belfort optimizes for TFHE (Torus FHE) for faster bootstrapping.
Challenges include energy efficiency—FHE remains compute-intensive—and ecosystem adoption, requiring SDKs for integration with TensorFlow or Ethereum. Belfort’s edge lies in its academic pedigree: Verbauwhede’s ERC grants (two Advanced awards) and DARPA ties provide non-dilutive validation, potentially unlocking SBIR/STTR funding in the U.S.
| Competitor | Focus | Strengths | Belfort Advantage |
| Microsoft SEAL | Software library | Open-source, broad language support | Belfort’s hardware yields 50-100x faster inference |
| Cornami ARES | FPGA-based accelerator | Military-grade security | Belfort’s streaming reduces memory by 90%, enabling scale |
| Duality Technologies | Cloud FHE service | Enterprise SaaS | Belfort’s on-prem hardware avoids vendor lock-in |
| Enveil | ZeroReveal platform | Analytics on ciphertext | Belfort targets raw compute acceleration for AI/blockchain |
Broader Impact and Outlook
This round arrives amid heightened scrutiny on data sovereignty—e.g., post-2025 breaches at Equifax and Change Healthcare exposed millions. Belfort’s “trust is all you need” ethos resonates, positioning it as a foundational layer for an “AI-first world” where models process encrypted inputs natively. Early wins like the SWIFT PoC hint at revenue potential: A single enterprise license could exceed $1 million annually.
Looking ahead, Belfort eyes a Series A in 2027, post-chip tape-out, with ambitions for partnerships (e.g., Google Cloud integration via Dean). Risks include supply chain delays for fabs and regulatory hurdles for export-controlled crypto tech, but the investor backing mitigates these. Ultimately, Belfort exemplifies Europe’s deep-tech resurgence, blending Leuven’s research excellence with Silicon Valley’s commercialization speed.
Please email us your feedback and news tips at hello(at)dailycompanynews.com
