Datavault AI’s CyberCatch Deal Turns a Stock Offer Into a $94.5 Million Cash Test

Datavault AI signed a definitive agreement to buy CyberCatch for $94.5 million in cash, replacing an earlier all-stock proposal announced in May. The deal reached Google News as a completed negotiation, but it is not a completed acquisition. Financing details, shareholder approval, court approval, and integration work still stand between the agreement and closing.

That change in consideration is the central story. Datavault AI first proposed issuing approximately 49.9 million shares for CyberCatch. It now plans to pay cash for approximately 26.8 million outstanding shares, valuing each CyberCatch share at $3.53.

The new structure removes the direct share issuance contemplated by the original proposal. It also transfers attention toward Datavault AI’s liquidity, financing capacity, and competing capital commitments. The company has not explained the cash source in its public acquisition announcement.

This is why the transaction deserves more scrutiny than a routine cybersecurity acquisition. CyberCatch offers a plausible security layer for Datavault AI’s data, tokenization, and edge-computing plans. However, the buyer must first show that its financing and integration plans match the scale of its ambition.

The definitive agreement changes the transaction, not the closing risk

Datavault AI has moved from an acquisition proposal to a signed agreement, but several material conditions remain unresolved.

The company announced the definitive acquisition agreement on August 14, 2026. Datavault AI agreed to acquire all issued and outstanding CyberCatch shares through a court-approved plan of arrangement under British Columbia law.

CyberCatch is listed on the TSX Venture Exchange under CYBE and trades over the counter as CYBHF. Datavault AI trades on Nasdaq under DVLT. The cross-border structure introduces approvals beyond those involved in a simple private-company purchase.

CyberCatch shareholders must approve the transaction. A British Columbia court must approve the arrangement, and the parties must obtain necessary exchange and regulatory clearances. Other customary closing conditions also apply.

Until those conditions are satisfied, CyberCatch remains independent. Datavault AI’s announcement consistently describes the operational combination as an expected outcome following closing.

That distinction matters because some Google News headlines compress “signed an acquisition agreement” into “acquired.” The latter wording implies that ownership has already transferred. Datavault AI has signed a binding deal, but it has not reported a completed closing.

If the transaction closes, CyberCatch will become a wholly owned Datavault AI subsidiary. The business is expected to continue operating from San Diego, with founder and CEO Sai Huda becoming the subsidiary’s president.

Huda would report to Datavault AI CEO Nathaniel Bradley. That arrangement provides leadership continuity, which can reduce disruption among employees, customers, and partners during an integration.

The consideration has changed significantly since the parties announced their binding letter of intent on May 1. That earlier proposal contemplated an all-stock transaction.

Under the May terms, CyberCatch shareholders would have received approximately 49.9 million newly issued Datavault AI shares. The proposal valued approximately 26.8 million CyberCatch shares at C$5.11 each.

CyberCatch shareholders were expected to own approximately 7.52% of Datavault AI after that transaction. Existing Datavault AI shareholders would have retained approximately 92.48%, before accounting for full dilution.

The definitive agreement discards that ownership exchange. Datavault AI now plans to pay $94.5 million in cash, or $3.53 for each CyberCatch share. Outstanding dilutive CyberCatch securities will be handled through cashless exercises.

A cash agreement can offer CyberCatch shareholders greater value certainty. They no longer need to evaluate the future trading price of Datavault AI shares when estimating their consideration.

For Datavault AI shareholders, the immediate benefit is the absence of the 49.9 million acquisition shares described in May. The tradeoff is that the buyer must provide substantial cash instead.

That creates a different risk profile. Share dilution attached directly to the acquisition has disappeared, but debt, financing costs, or a separate capital raise can still affect shareholders.

The definitive announcement does not identify a committed acquisition facility, lender, escrow balance, or other funding source. It also does not describe a financing condition.

Those omissions do not prove that financing is unavailable. They mean outside readers cannot yet evaluate the transaction’s funding structure from the announcement alone.

The deal’s legal status has therefore advanced while its financial mechanism has become less transparent. That is the reversal readers should remember as the story moves beyond Google News summaries.

Why CyberCatch fits Datavault AI’s wider platform plan

CyberCatch gives Datavault AI a continuous security and compliance layer that connects several otherwise separate product ambitions.

Datavault AI describes its business around data valuation, digital credentials, tokenization, high-performance computing, and acoustic technologies. These activities can involve sensitive records, access permissions, digital assets, and regulated customers.

CyberCatch sells software designed to monitor cybersecurity controls continuously. Continuous compliance means testing whether required safeguards remain active, instead of relying only on a periodic audit.

The platform maps results against frameworks used by regulated organizations. Datavault AI identifies NIST CSF 2.0, NIST 800-171, CMMC, ISO 27001, HIPAA, and PCI DSS among the supported frameworks.

CyberCatch also uses automated security testing from three directions. These include outside-in testing, internal testing, and social-engineering exercises.

According to Datavault AI, specialized AI agents can perform tasks associated with a penetration test. These tasks include reconnaissance, vulnerability detection, exploitation attempts, evidence collection, reporting, and remediation recommendations.

A penetration test is a controlled attempt to find exploitable weaknesses in a system. CyberCatch’s proposed distinction is frequency, since its software is designed to run tests continuously.

CyberCatch also generates a Cyber Hygiene Score and a Cyber Breach Score. The company says these measures help customers identify control failures and prioritize corrections.

These descriptions are company claims rather than results from an independent technical evaluation. Buyers should still examine testing scope, false positives, remediation workflows, and performance against human-led assessments.

The strategic fit is nevertheless understandable. Datavault AI wants to place CyberCatch across DataValue, DataScore, and its Information Data Exchange.

The same security layer is expected to support workloads running through the SanQtum edge platform. Edge computing processes data closer to its source, reducing dependence on a distant centralized data center.

Distributing workloads also expands the number of systems requiring consistent security controls. A compliance dashboard is more useful when it can assess every node under one operational model.

Datavault AI CEO Nathaniel Bradley summarized the rationale directly: “Cybersecurity is no longer a separate stack from data and AI.” His statement frames security as a prerequisite for the company’s other services.

That argument becomes especially relevant in regulated markets. Federal contractors, healthcare providers, financial companies, and manufacturers often need evidence that required controls remain operational.

For those customers, a failed control can delay procurement or create audit problems. Datavault AI wants CyberCatch to produce a real-time compliance signal alongside its data and computing services.

The planned combination also includes MARS-MABE, CyberCatch’s encryption technology. The name refers to multi-authority attribute-based encryption with revocation.

Attribute-based encryption controls access according to defined user characteristics. A system might authorize data only when a user has the required role, organization, clearance, and location.

Multi-authority systems distribute responsibility for issuing those attributes. Revocation allows administrators to remove access when a user’s circumstances change.

CyberCatch says its design can revoke access to specific data subsets without re-encrypting an entire collection. That capability would be useful in environments where permissions change frequently.

The technology entered CyberCatch through its February 2026 acquisition of Atriarch. CyberCatch’s interim financial filing confirms the share exchange used to acquire that encryption intellectual property.

Datavault AI says CyberCatch is converting MARS-MABE toward quantum resistance. That work remains a development objective, not an independently verified finished capability.

Post-quantum cryptography uses algorithms designed to resist attacks from future quantum computers. It does not require a quantum network or a functioning quantum computer.

The distinction matters because the acquisition announcement combines several labels, including quantum-secured infrastructure, quantum resistance, and a Quantum Private Network. Those terms describe related ambitions, but they are not interchangeable.

Datavault AI’s own annual filing states that SanQtum uses algorithms based on NIST’s finalized post-quantum standards. The filing identifies ML-KEM, ML-DSA, and SLH-DSA as relevant standards.

CyberCatch’s conversion work must eventually show how its attribute-based encryption interacts with those standardized components. Buyers will need evidence covering security, performance, key management, and interoperability.

The acquisition gives Datavault AI additional software and expertise for that work. It does not make every part of the combined security stack quantum-resistant upon signing.

The cash structure puts Datavault AI’s balance sheet under pressure

The main opponent in this transaction is not another cybersecurity vendor. It is Datavault AI’s strategic ambition versus its financing capacity.

The cash obligation is material because Datavault AI is pursuing several projects at once. Its plans span tokenization, acquisitions, distributed GPU infrastructure, licensing, and data services.

The company previously announced a definitive agreement to acquire NYIAX. Its 2025 annual report estimated the value of that stock-based transaction at $59.2 million.

Datavault AI has also discussed deploying an edge network across more than 100 U.S. cities. In May, management described a planned fleet of approximately 48,000 GPUs.

Those targets require capital even before CyberCatch enters the picture. Equipment purchases, site preparation, network operations, integration, security testing, and customer support all consume resources.

Datavault AI announced a proposed $120 million cash contribution and revenue-participation arrangement with Scilex in April. It also completed a $60 million registered direct offering in May.

Those announcements show that the company has actively pursued outside capital. They do not, by themselves, establish which funds are available for the CyberCatch purchase.

The CyberCatch announcement does not connect the acquisition payment to either transaction. It also does not explain whether any earlier proceeds remain restricted for infrastructure or working capital.

This creates the most important unanswered question following the Google News cycle: where will the acquisition cash come from, and on what terms?

Several financing routes are possible. Datavault AI could use existing cash, borrow money, obtain financing from a strategic partner, sell assets, or raise additional equity separately.

Each option shifts risk differently. Existing cash can reduce operational flexibility, while debt adds interest and repayment obligations.

A strategic financing agreement can include revenue sharing, collateral, or restrictive covenants. A separate equity issuance can dilute shareholders even though the acquisition agreement itself uses cash.

The absence of acquisition shares therefore should not be confused with the absence of dilution risk. Investors need to evaluate the full financing chain.

The timing puts additional weight on Datavault AI’s scheduled second-quarter results. The company plans to report before the market opens on August 19, followed by an 8:30 a.m. Eastern conference call.

That earnings call offers management an immediate opportunity to clarify liquidity and acquisition funding. CEO Nathaniel Bradley and CFO Brett Moyer are scheduled to present.

The most useful disclosure would separate unrestricted cash from capital committed elsewhere. It would also identify any debt facility, acquisition financing, or closing-related funding requirement.

Investors should watch whether management describes financing as committed, expected, or still under negotiation. Those words indicate very different levels of certainty.

Datavault AI’s annual filing provides important context for evaluating management’s plans. The company’s 2025 Form 10-K discusses dependence on partners and the risks surrounding technology development, commercialization, and financing.

The document also describes an expanding collection of businesses and planned platforms. That breadth creates opportunities for cross-selling, but it raises execution demands.

Every acquisition brings accounting systems, contracts, employees, security policies, and customer obligations. Each additional product also requires technical integration and a coherent sales story.

CyberCatch itself reported minimal revenue and high liquidity risk in its January 2026 interim statements. Its operations depended on financing from related parties and outside sources.

That does not erase the value of its intellectual property or customer relationships. It does suggest the acquisition is primarily a capability investment rather than the purchase of a large established revenue stream.

Datavault AI has not published a CyberCatch revenue contribution target in the acquisition release. It has not disclosed expected cost savings, integration expenses, or a timetable for financial accretion.

Without those figures, readers cannot build a reliable return model. The strategic rationale can be assessed, but the economic payoff remains difficult to measure.

The all-cash structure therefore sharpens the burden of proof. Datavault AI must show not only that CyberCatch fits, but that the purchase preserves enough capital for everything else.

The cybersecurity promise still needs technical validation

Buying CyberCatch creates a broader security story, but integration and verification will determine whether that story becomes a working product.

Datavault AI wants CyberCatch to sit across several platforms. That scope includes data valuation, tokenization, exchange infrastructure, edge computing, acoustic technologies, and regulated workloads.

A common security layer can reduce fragmentation. It can also become difficult to implement when the underlying products use different architectures, identities, data models, and deployment environments.

Continuous compliance depends on accurate asset discovery. A platform cannot reliably test controls on systems it does not know exist.

It also needs current mappings between technical evidence and regulatory requirements. Frameworks evolve, customer environments vary, and one control can require different evidence across industries.

Agentic penetration testing adds another challenge. Automated agents must operate inside carefully defined boundaries to avoid service disruption, data exposure, or unauthorized actions.

Customers will need controls over which systems can be tested. They will also need audit records, approval workflows, rate limits, and ways to reproduce important findings.

False positives can waste security teams’ time. False negatives are more serious because they can create misplaced confidence about systems that remain vulnerable.

Datavault AI’s announcement explains the intended functions but provides no independent benchmark. It does not compare CyberCatch’s detection quality with established scanning platforms or professional penetration-testing teams.

The announcement also does not disclose customer retention, deployment duration, alert volumes, or remediation rates. Those operating metrics would help enterprise buyers assess maturity.

CyberCatch competes within a crowded security market. Large vendors already offer combinations of vulnerability management, cloud security posture management, compliance reporting, identity controls, and attack simulation.

Specialists also provide continuous penetration testing and automated breach simulation. CyberCatch must differentiate through accuracy, regulatory mapping, ease of deployment, or integration with Datavault AI’s infrastructure.

Its strongest potential distinction is not simply adding AI to security testing. Many vendors already use machine learning or automated workflows.

The more specific proposition combines continuous compliance, security testing, attribute-based access control, and Datavault AI’s data infrastructure. That package could appeal to organizations seeking fewer disconnected tools.

However, the package also creates a wider attack surface. Integration code, shared identities, centralized dashboards, and automated actions can introduce new failure points.

Datavault AI and CyberCatch will need to show how they separate customer environments. They must also explain permissions, credential storage, logging, incident response, and software-update controls.

MARS-MABE carries a separate validation burden. Attribute-based encryption can provide detailed access policies, but complexity can make implementation and recovery harder.

The announced conversion to quantum resistance will require more than replacing one cryptographic primitive. The combined design must preserve revocation, performance, and policy enforcement.

Independent cryptographic review would materially strengthen the claim. So would clear documentation showing which standardized algorithms protect each operation.

Datavault AI cites Google’s 2029 migration target for authentication systems as evidence that post-quantum preparation is becoming urgent. Google’s migration guidance reflects a broader concern about long transition periods.

That concern is legitimate. Large organizations cannot replace identity and encryption systems overnight.

However, an urgent migration timeline does not validate a particular vendor’s implementation. Buyers must still test the design, software, key lifecycle, and operational controls.

The combined company will also need to avoid vague language around quantum security. “Quantum-ready” can describe preparation, while “quantum-resistant” should refer to specific cryptographic protections.

Clear terminology will help technical buyers distinguish deployed safeguards from planned upgrades. It will also prevent marketing claims from moving ahead of engineering evidence.

Datavault AI acknowledges many of these uncertainties in its forward-looking statement. The company identifies integration, regulatory approvals, technical development, market demand, and post-quantum conversion as risks.

Those warnings are not boilerplate details to ignore. They describe the precise areas where the acquisition thesis can succeed or fail.

What to watch after the Google News headline

Three signals will determine whether this agreement becomes a credible platform expansion or another unfinished Datavault AI commitment.

The first signal is acquisition financing. Datavault AI’s August 19 results should reveal its cash position and recent capital movements.

Management should explain whether the full CyberCatch payment is funded. It should also identify any lender, financing partner, material conditions, collateral, or expected closing costs.

A committed financing package with manageable obligations would strengthen the acquisition thesis. A vague reference to future capital would leave the largest execution question open.

Readers should also compare the financing plan with Datavault AI’s other obligations. CyberCatch cannot be evaluated in isolation from the company’s GPU network, NYIAX transaction, and operating requirements.

The second signal is the approval and closing process. CyberCatch must distribute transaction materials and obtain shareholder approval before the arrangement can proceed.

The British Columbia court process and TSX Venture Exchange requirements provide additional checkpoints. Delays, revised terms, or supplemental disclosures can reveal issues not visible in the initial announcement.

A completed closing on the announced terms would turn the story from an agreement into an acquisition. Until then, headlines should continue describing the transaction as pending.

The third signal is measurable technical integration. Datavault AI should identify the first products using CyberCatch and explain what customers can deploy.

Useful evidence would include a named integration release, a documented architecture, or a customer implementation. Independent security testing would carry more weight than another broad platform announcement.

The company should also specify which parts of MARS-MABE are currently deployed. It should distinguish existing encryption functions from planned post-quantum changes.

Customer metrics will matter after integration begins. Retention, deployment time, tested assets, confirmed findings, and remediation rates can show whether the system delivers operational value.

Revenue disclosures will provide another reality check. CyberCatch’s interim filing showed a small commercial base relative to the acquisition consideration.

Datavault AI does not need immediate large revenue for the transaction to work. It does need a credible path from acquired technology to customer adoption.

That path should explain which Datavault AI customers need continuous compliance, how the product will be sold, and who provides implementation support.

The acquisition also creates a governance test. Datavault AI is assembling several businesses while pursuing infrastructure, tokenization, and licensing programs.

Management must prioritize integrations that customers can understand and buy. Simply placing every product under one platform label will not produce technical or commercial unity.

For enterprise buyers, the near-term question is practical. Does the combined offering reduce security work, improve evidence quality, and fit existing controls?

Developers should watch the boundaries placed around automated penetration testing. Security agents require permissions, logging, reproducibility, and human oversight before they can enter sensitive production environments.

Knowledge workers and compliance teams should focus on evidence quality. Automated dashboards are useful only when their findings map accurately to real systems and current requirements.

Anyone following the story through Google News should separate four stages: announcement, approval, closing, and integration. Datavault AI has reached only the first stage under its definitive agreement.

The strategic logic is stronger than the available financial explanation. CyberCatch can fill a real gap between Datavault AI’s data ambitions and the controls regulated customers require.

The transaction still asks investors to accept several unproven links. These include acquisition funding, closing, product integration, customer adoption, and post-quantum validation.

Watch the August 19 financial report first. Then follow CyberCatch’s approval materials and Datavault AI’s initial integration release.

Those three checkpoints will provide more useful evidence than another round of acquisition headlines. Until they arrive, treat the deal as strategically coherent, financially unresolved, and technically unfinished.