Alice has reportedly raised $140 million at a valuation between $700 million and $800 million, according to coverage from Calcalist and Bloomberg. The financing places the startup among the companies seeking to define how security should work as businesses deploy AI systems, including tools and models from major providers.

Bloomberg described Alice as an AI security startup working with Anthropic and Google. Calcalist framed the financing around the company’s view that AI will require a new security layer rather than simply an extension of conventional software protection. The available reporting does not identify the investors, the round’s structure, or the precise valuation, so those details remain unconfirmed from the source material.

Funding targets a new AI security category

The central news is not only the size of Alice’s financing, but the category it is attempting to build around. Traditional enterprise security products were designed largely for software, networks, endpoints, identities, and data stores. AI systems introduce additional control points: models can interpret instructions, call tools, retrieve information, generate actions, and interact with employees or customers.

Alice’s reported thesis is that these systems need a dedicated security layer. That could mean controls that observe or govern how AI applications access data and services, although the supplied coverage does not specify Alice’s product architecture or the precise risks it addresses. It would be premature to assume that the company is focused on one particular area, such as model monitoring, prompt injection, identity, data leakage, or agent permissions.

The valuation range reported by Calcalist—$700 million to $800 million—signals that investors see a potentially large market in the security infrastructure surrounding AI adoption. It is not, by itself, evidence that Alice has established a market standard or that its technology has been independently validated.

Anthropic and Google links raise the stakes

Bloomberg’s description of Alice as working with Anthropic and Google provides the strongest indication in the available evidence that the startup is engaging with major AI platforms. However, the report does not explain the nature of those relationships. “Working with” could refer to a technical integration, a customer relationship, a research effort, or another form of collaboration; the supplied source does not establish which.

That distinction matters to buyers and builders. A formal integration could make Alice easier to deploy alongside widely used AI services. A research relationship could suggest that the company is addressing problems recognized by model developers. A customer engagement would indicate a different kind of validation. Without further details, the relationships should be treated as reported connections rather than proof of product adoption or endorsement.

The same caution applies to the financing. Bloomberg and Calcalist both report the $140 million raise, but the provided extracts contain no investor list, executive comments, use-of-funds plan, revenue figures, customer count, or performance benchmarks. Those omissions limit what can be concluded about Alice’s current scale.

Evidence is strong on the financing, thin on the product

The financing and valuation are the clearest reported facts in this story. Both source headlines identify a $140 million raise, while Calcalist supplies the $700 million-$800 million valuation range. Bloomberg identifies the company’s positioning as AI security and mentions work with Anthropic and Google.

The evidence is much thinner on the technical and commercial case. There is no supplied information about which AI environments Alice protects, whether it serves model developers or enterprise application teams, how it measures security outcomes, or how its system differs from established cloud and cybersecurity products. There are also no independently reported benchmarks or verified adoption figures.

That gap is important because AI security claims can cover very different products. A tool that blocks malicious prompts has a different deployment model from one that manages agent permissions. A platform that monitors model outputs may compete with application observability vendors, while a system that governs access to enterprise data may compete with identity and data-security providers. The phrase “new security layer” describes a strategic position, not a sufficiently detailed product category.

What the raise means for AI builders and buyers

For AI product teams, Alice’s financing is a sign that security is moving closer to the core architecture of AI applications. Teams building AI agents or model-powered workflows will need to decide which actions systems may take, which data they may retrieve, and how human approval fits into automated processes. Security controls added after deployment can be difficult to retrofit, especially when an application connects models to business systems.

For enterprise buyers, the practical question is whether a new security product can reduce operational risk without adding another disconnected console. Buyers will likely want clear answers about deployment, integration with existing identity and access systems, audit logs, policy management, incident response, and support for multiple model providers. They will also need to test whether controls work in real workflows rather than only in controlled demonstrations.

Alice’s reported valuation may intensify competition among startups, cloud providers, model companies, and established cybersecurity vendors. But funding alone does not establish which layer of the stack will capture value. The winning products may be those that combine strong controls with low-friction deployment and explainable decisions for security teams.

What to watch next

The next meaningful signals will be concrete product and market disclosures from Alice. Key questions include which AI applications and model providers it supports, whether its relationships with Anthropic and Google involve integrations or customers, and what security problems its platform is designed to prevent.

Investors and prospective customers should also watch for the identity of the financing participants, the company’s use of the new capital, independent customer references, and evidence of recurring commercial adoption. Technical documentation would help clarify whether Alice is securing models, AI agents, data access, tool calls, or the broader application lifecycle.

Further reporting may also establish whether the valuation is based on a completed financing or a range associated with the transaction. Until those details emerge, the raise is best understood as a strong investor bet on the importance of AI security, not as proof that Alice has solved the category’s hardest problems.

Creati.ai perspective

Alice’s financing reflects a real architectural pressure: AI applications can turn model outputs into access requests, tool calls, and business actions, creating security requirements that conventional application controls may not fully address. The opportunity is substantial, but the category remains difficult to define.

For builders and enterprises, the important test will be operational rather than rhetorical. Alice will need to show exactly where its security layer sits, how it works with existing systems, and whether it can reduce risk without slowing useful AI workflows. The reported relationships with Anthropic and Google are notable, but product transparency and verifiable customer outcomes will matter more than the financing headline.