Datavault AI (NASDAQ: DVLT) has signed a definitive agreement to acquire 100% of CyberCatch Holdings (TSXV: CYBE, OTCQB: CYBHF) for USD $94.5 million in cash, or $3.53 per share across roughly 26.8 million shares outstanding. The deal is structured as a court-approved plan of arrangement under the Business Corporations Act of British Columbia. All outstanding dilutive securities get exchanged on a cashless-exercise basis.
CyberCatch would operate as a San Diego-based subsidiary. Founder, chairman and CEO Sai Huda becomes president of the subsidiary, reporting to Datavault AI CEO Nathaniel T. Bradley.
What Datavault AI Is Actually Buying
CyberCatch sells an AI-enabled SaaS platform for continuous compliance and cyber risk mitigation. The pitch is that annual penetration testing is a broken model: most organizations can’t afford the human expertise a comprehensive test requires, and the ones that can typically run it once a year, leaving eleven and a half months of unmeasured exposure.
The platform replaces that cadence with agentic AI. Specialized agents handle reconnaissance, vulnerability detection, TTP selection, exploitation, evidence gathering and reporting, running continuously against three attack surfaces: outside-in, inside-out and social engineering. Generative AI confirms required controls are in place and produces a Cyber Hygiene Score; the agentic side produces a Cyber Breach Score. Dashboard reporting maps failures to NIST CSF 2.0, NIST 800-171, CMMC 2.0, ISO 27001, HIPAA, PCI DSS, CAN/DGC 104 and EU NIS.
The second asset is encryption. CyberCatch picked up multi-authority attribute-based encryption with revocation through a February 2026 acquisition, now branded MARS-MABE, and is working to make it quantum-resistant. The claimed advantages over RSA and AES-256 are fine-grained attribute-gated access, restriction to data subsets rather than whole sets, and instant revocation without re-encrypting everything. That last point is the commercially interesting one, since re-encryption cost is what makes revocation painful at scale in real deployments.
Third is the roster. CyberCatch’s board and advisory board include Tom Ridge, the first Secretary of Homeland Security; Dr. Marv Langston, former DARPA information systems director and Navy cybersecurity chief; and Scott Tait, former Navy commander and national security advisor to the Joint Chiefs. Huda holds USPTO patent 11,297,094 on automated continuous cybersecurity assessment and co-authored Canada’s national cybersecurity standard. For a company selling into the defense supply chain, that roster is part of the product.
The Price Sits Well Above the Business
Here is where the transaction gets hard to read on operating fundamentals alone. CyberCatch’s trailing twelve-month revenue is somewhere in the region of $230,000. Fiscal 2025 revenue came in around $404,000, down roughly 38% year over year, against losses of about $5.6 million. Market capitalization sat near $44 million as of August 10, before the announcement.
So Datavault AI is paying roughly double the pre-announcement market value, and several hundred times revenue, in cash rather than stock.
That matters because of where the cash comes from. On its Q1 2026 call in May, Datavault AI put working capital at approximately $140 million following a $60 million registered direct offering, plus a binding term sheet for $120 million in non-dilutive financing earmarked for the SanQtum edge rollout. Q1 revenue was $3.4 million against a $53.1 million net loss, driven by bitcoin fair value losses and a NYIAX impairment. A $94.5 million cash outlay against that balance sheet is not a bolt-on. It is a substantial share of stated working capital, spent on an asset whose value is entirely prospective.
Datavault AI reports Q2 results before market open on Wednesday, August 19. That call is where the funding question gets answered, and it lands four days after this announcement.
The Post-Quantum Framing
The strategic case leans on two timelines. Google has set 2029 as its internal deadline for migrating authentication systems to quantum-resistant cryptography, and Google Quantum AI research has put the qubit threshold for breaking elliptic curve cryptography an order of magnitude below earlier estimates. Separately, CrowdStrike’s 2026 Global Threat Report puts AI-enabled adversary attacks up 89% year over year in 2025, with average eCrime breakout time down to 29 minutes.
Both trends are real and both cut in favor of continuous testing over annual audits. Whether they justify this price for this asset is a different question. The release also cites Gartner’s $213 billion 2025 information security spending projection, which is the total addressable market for the entire industry, not a figure CyberCatch’s platform is positioned to capture any meaningful slice of at current scale.
What to Watch
Closing requires CyberCatch shareholder approval, British Columbia court approval of the arrangement, TSX Venture Exchange sign-off and customary conditions. Shareholder approval looks close to assured at roughly double the market price.
The integration story is that CyberCatch becomes the compliance layer across DataValue, DataScore and the Information Data Exchange, all running on Available Infrastructure’s SanQtum edge platform. Continuous compliance attestation against NIST, CMMC, ISO 27001, SOC 2, HIPAA and PCI DSS is increasingly a procurement precondition in federal and regulated markets, so the logic holds on paper. Execution is the open item, and Datavault AI has assembled a wide portfolio of acquisitions and licensing arrangements over the past eighteen months without yet demonstrating that the pieces compound.