VC
Value Add VC
โšกHomePulseโšกHelpful Apps๐Ÿ“Blog
โ† Value Add PulseAI

Microsoft Preps Project Perception to Rival Anthropic's Mythos

Microsoft is readying Project Perception, a multi-model AI security tool that routes vulnerability-scanning tasks across Microsoft, OpenAI and Anthropic models to undercut the cost of Anthropic's own Mythos on the exact capability Anthropic pioneered.

Microsoft, OpenAI, Anthropic
Models routed
By end of July
Expected launch
Anthropic's Mythos
Competing product
TC
Trace Cohen
Early-stage VC & angel ยท Founder, New York Venture Partners
July 17, 2026
1 min read
ShareXLinkedInEmail
THE RUNDOWN
1

Project Perception uses a model-router architecture that assigns each security task to whichever AI model -- Microsoft's own, OpenAI's, or Anthropic's -- is best suited and cheapest for that specific job, rather than routing everything through one frontier model

2

The tool is positioned directly against Anthropic's Mythos, which CISA has been using to scan federal code repositories for vulnerabilities since export controls on Mythos 5 were lifted for approved U.S. organizations on July 1

3

Microsoft's pitch is explicitly cost-based: multi-model routing reserves expensive frontier-model calls for only the steps that need them, lowering the average cost per vulnerability report relative to a single-model approach

4

A launch is expected before the end of July, which would put Microsoft in direct competition with the AI-security category Anthropic has spent much of 2026 building out with government and enterprise customers

TC
The VC Read ยท Trace's TakeTrace Cohen

Microsoft going after Mythos with a cheaper, multi-model router is the playbook it's run in every category it's entered late -- undercut on price, bundle into an existing enterprise footprint, and let distribution do the rest. Anthropic's advantage was being first and being genuinely better, but 'genuinely better' is a shrinking moat when the buyer is a CISO comparing line items against an Azure enterprise agreement they already have. The AI-security category is about to get a lot more crowded, and that's good news for enterprise buyers and bad news for every smaller Mythos-alternative startup that hadn't locked in distribution yet.

Microsoft is preparing to launch Project Perception, an AI-driven cybersecurity platform that scans enterprise code and systems for vulnerabilities using a router that assigns each task to whichever model -- Microsoft's own, OpenAI's, or Anthropic's -- is the best fit and lowest cost for that specific job, according to reporting from The Information and TechRepublic. A launch is expected before the end of July.

The product is a direct competitive response to Anthropic's Mythos, which has become a genuine strategic asset for Anthropic this year: CISA has been using Mythos to scan code repositories across federal agencies for vulnerabilities since export controls on Mythos 5 were lifted for approved U.S. organizations on July 1, and the model reportedly found vulnerabilities in classified government systems earlier this year. Anthropic has kept Mythos access limited to a curated set of partners, which has created an opening for a well-resourced competitor to build a cheaper, broader alternative.

โ€œorganizations on July 1, and the model reportedly found vulnerabilities in classified government systems earlier this year.โ€

Microsoft's multi-model routing is the specific mechanism for that cost advantage: rather than running every scan through a single frontier model at frontier pricing, Project Perception can reserve expensive calls for the small subset of steps where a top-tier model actually changes the outcome, and route routine scanning to cheaper models. That's the same architectural pattern showing up elsewhere in the industry this week -- Fireworks' entire business model is built on routing enterprise workloads to specialized, cheaper models rather than defaulting to the most expensive general-purpose option.

For enterprise security buyers, a credible multi-model alternative to Mythos gives Microsoft's massive existing enterprise footprint -- through Defender, Azure and its OpenAI partnership -- a natural distribution advantage that a standalone product would lack. What to watch next: pricing details once Project Perception actually launches, and whether Anthropic responds by broadening Mythos access beyond its current partner cohort to defend share.

ShareXLinkedInEmail
More onAnthropic โ†’OpenAI โ†’Microsoft โ†’

Originally reported by The Information. Analysis and editorial commentary by Value Add Pulse.

โ† Back to Pulse

THE WIRE in your inboxโ€” Tech, startup & VC news with Trace's take. Free, no spam.

Read Next

AI$10B compute lease

Anthropic in Talks to Lease $10B of Meta's Compute

Anthropic and Meta are in early talks for Anthropic to lease up to $10B of Meta's AI compute over two years, letting Meta monetize its buildout while Anthropic diversifies beyond Amazon and Google.

AI

China's Open-Weight Wave Forces an Enterprise Rethink

Kimi K3's benchmark-topping debut is accelerating enterprise interest in open-weight models, forcing US buyers to weigh self-hosted Chinese models against closed, subscription-priced offerings from Anthropic and OpenAI.

AI$960,000

Jensen Huang's Leather Jacket Sells for $960K

A leather jacket worn by Nvidia CEO Jensen Huang sold for $960,000 at Sotheby's, nearly 20 times its pre-sale estimate, with proceeds benefiting a philanthropic initiative for young tech builders.

@Trace_Cohenยทt@nyvp.com