Anthropic’s AI Deals Shake Up Cyber, Software

Anthropic has introduced Claude Mythos, described as its most advanced artificial intelligence model, but is restricting access due to its ability to identify and exploit software vulnerabilities. Rather than a broad public release, Anthropic is making the large language model available to about 50 leading technology companies through an initiative called Project Glasswing, aimed at strengthening cybersecurity defenses.

According to Anthropic, Mythos has already uncovered thousands of vulnerabilities across every major web browser and operating system. The company has warned it expects malicious actors to attempt to use AI to exploit such weaknesses, underscoring why access is being tightly controlled and focused on trusted partners.

Participants in Project Glasswing will receive a total of $100 million in usage credits from Anthropic. After these credits, access to Claude Mythos is priced at $25 per million input tokens and $125 per million output tokens, setting clear commercial terms for ongoing enterprise use of the model.

Two cybersecurity vendors, CrowdStrike (CRWD) and Palo Alto Networks, have been named among 12 leading members within Project Glasswing. Their inclusion signals that Anthropic is looking to established cybersecurity providers to play central roles in AI-era protection rather than building competing security solutions itself.

Wedbush analysts characterized the announcement as significant for CrowdStrike (CRWD) and Palo Alto Networks, arguing that access to Claude Mythos could help them become an “AI enforcement layer” instead of being sidelined by AI-driven disruption. The model’s advanced vulnerability-finding capabilities reinforce expectations that cyber defenses will need to incorporate AI tools to counter AI-enabled attacks.

Both companies are pursuing platform-oriented strategies, which the reports suggest may benefit from Glasswing participation as organizations prioritize integrated security over disparate point tools. Commentary from The Motley Fool and Yahoo Finance notes that growing AI-related cyber risks are likely to keep cybersecurity spending high and that CrowdStrike and Palo Alto were already viewed as leaders before the Anthropic announcement.

Separately, Bloomberg reporting cited by PYMNTS says investors have approached Anthropic about a new funding round that could value the company at as much as $800 billion. The firm has so far resisted these offers, according to unnamed sources, following a February fundraise that reportedly implied a $350 billion pre-money valuation.

The same report states that Anthropic’s AI tools have gained strong traction among business customers, driving revenue that helps it compete with OpenAI. Anthropic is said to have reached $30 billion in annual run-rate revenue, up from $19 billion only months earlier, and has discussed the possibility of going public, potentially later this year, though no decision or terms were disclosed.

PYMNTS also highlights that large enterprises adopting AI often face organizational hurdles rather than purely technical barriers. More than 70% of surveyed executives at companies with at least $1 billion in annual revenue cited organizational readiness as the main limitation on AI performance, versus 11% who pointed to the technology itself.

Palantir Technologies sits at the center of a separate debate about AI software valuations. Articles from Yahoo Finance and The Motley Fool describe the stock trading at a trailing price-to-earnings multiple of around 215 and roughly 100 times this year’s projected earnings of $1.30 per share, despite a 35% pullback from recent highs and a 36% decline from its October peak.

On the bullish side, Palantir has continued to monetize its AI platform and recently drew public praise from President Donald Trump for what he called the company’s “great warfighting capabilities” supporting the U.S. military. Cathie Wood’s Ark Invest has also been buying shares on weakness after previously trimming its position at higher levels.

In contrast, investor Michael Burry has disclosed a bearish stance. In a Substack post, he said he holds Palantir put options with strike prices of $100 expiring in December 2026 and $50 expiring in June 2027, stating he believes the company’s “fundamental value” is well under $50 per share. Burry highlighted the stock’s rapid post-2022 appreciation and premium valuation relative to broader market benchmarks and large AI peers such as Microsoft (MSFT) and Alphabet (GOOGL).

Analysts cited by The Motley Fool place a consensus 12‑month price target for Palantir at $191.29, nearly 50% above its current level of about $135.70, while acknowledging heightened volatility and uncertainty as competition from other AI software providers, including Anthropic, intensifies.