Enterprise AI & Security
Dreamteam Raises $53M as Anthropic Updates Claude and Amazon Builds
San Francisco startup Dreamteam raised $53 million led by Together Fund and Accel for its AI-native CRM, Anthropic released Claude Mythos 5.2 and Fable 5.2 with updated cybersecurity guardrails, and Amazon is developing a World Product Graph with Coca-Cola and Bayer.

Your listening Queue & catch up
Add a story or build a catch-up queue to start listening.
Enterprise software architecture, frontier cybersecurity model evaluations, and e-commerce product discovery are undergoing structural shifts as former Freshworks executives launch an AI-native customer relationship management platform, Anthropic releases updated security benchmarks for Claude Mythos 5.2 and Fable 5.2, and Amazon develops an expansive product research database.
Dreamteam Raises $53 Million to Challenge Salesforce and HubSpot
San Francisco-based startup Dreamteam has raised $53 million across seed and Series A funding rounds to compete with established customer relationship management providers Salesforce and HubSpot, according to TechCrunch. The seed and Series A rounds were led by Together Fund and Accel, respectively, with participation from Premji Invest. Co-founded in 2025 by chief executive Anand Venkatraman—who spent more than 25 years in enterprise software at companies including Salesforce and Freshworks—alongside Senthil Kanthaswamy, who spent more than 9 years at Freshworks, and Ramesh Parthasarathy, who led AI and platform engineering at Freshworks, the company employs more than 35 people across engineering operations in Chennai and go-to-market teams in the United States.
Dreamteam centers its platform on a virtual memory system called the Company Brain, which automatically captures context from emails, meetings, and customer conversations to maintain a shared record of customer relationships. Four built-in AI agents use that context to research prospects, draft follow-up emails, flag deal risks, and configure CRM workflows, while a feature called Magic Apps allows teams to build custom applications and dashboards using natural-language instructions without writing code. Within 3 months of active sales, Dreamteam has signed up more than 100 customers across 13 countries, replacing Salesforce at healthcare company DMS Health and HubSpot at other businesses. The advertising sales team at media streaming company Plex also selected Dreamteam, with Plex co-founder and chief product officer Scott Olechowski citing its integration with existing AI tools. Kanthaswamy said startups have a window of 2 to 5 years to gain ground while incumbents rebuild their architectures, and Dreamteam targets eight figures in annual revenue by the end of 2027.
Anthropic Rolls Out Claude Mythos 5.2 and Fable 5.2 Cybersecurity Updates
In frontier model engineering, Anthropic has rolled out Claude Mythos 5.2 to selected partners and Claude Fable 5.2 to regular users, introducing improvements in vulnerability discovery and reduced false-positive safety blocks, BleepingComputer reported. Anthropic describes Mythos 5.2 as its most capable model for cybersecurity engineering, outperforming Fable 5.2 at finding vulnerabilities, reproducing bugs, and developing fixes. In Anthropic's CVE Rediscovery evaluation, Mythos 5.2 detected 10.2% more vulnerabilities than Claude Opus 5.5 and 14.4% more vulnerabilities than Mythos 5.1, while a partner security firm reported that one in six confirmed bugs identified by Mythos 5.2 had been missed by every other frontier model tested.
Although Fable 5.2 and Mythos 5.2 share the same parent model, Anthropic applies additional cybersecurity and biology safeguards to Fable 5.2 for general users, blocking penetration testing, exploit generation, and certain binary analysis tasks and automatically switching high-risk cybersecurity queries to Opus 4.8. After Fable 5.1 triggered frequent blocks on routine administration tasks such as configuring AWS accounts, Anthropic reported that Fable 5.2 has reduced unnecessary cybersecurity blocks by approximately 60% across Claude.ai, Claude Code, and Cowork. Anthropic also evaluated Mythos 5.2 against sandbox escape attempts while noting that pre-release evaluations cannot catch every potential failure as models become better at recognizing test environments.
Amazon Develops World Product Graph to Compete With Google and AI Bots
Meanwhile, Amazon is developing an internal project called World Product Graph designed to help shoppers research items across the web, including products Amazon does not stock, according to Hoodline. Citing Reuters and internal documents organized by Selection & Catalog Systems division vice president Robert Tekiela, the report notes an aspirational release date of October 1, 2027, alongside companion tools called Buy for Me and Shop Direct that would allow customers to purchase items sold elsewhere through Amazon. According to Economic Times reporting cited in the coverage, Amazon has signed Coca-Cola and Bayer to contribute product data on an ongoing basis and held a chief-executive summit with Procter & Gamble regarding a potential trial.
The initiative addresses a research gap identified in a 2025 Morgan Stanley report showing that only 16% of online shoppers start product research on Amazon compared with 57% on Google, whereas 41% of shoppers who already know what they want start on Amazon versus 23% on Google and 18% on Walmart. While the internal document noted that Amazon's Rufus assistant previously handled about 83 million searches a week, Amazon stated that after rebranding Rufus as Alexa for Shopping in May, the assistant reached roughly 97% accuracy globally, and CEO Andy Jassy reported on Amazon's third-quarter earnings call that more than 400 million people used Alexa for Shopping over the past year. In September 2026, Amazon blocked Meta's Muse AI shopping agent—which had gained more than 10 million users in the U.S. and Canada—following its ongoing 2025 lawsuit against Perplexity, while a Billions Network and AmanTrue audit of 51 large U.S. retailers found that 5 explicitly restrict shopping agents and 37 block them through existing bot rules, even as a July 17, 2025 Useberry and adMarketplace survey of 1,018 U.S. consumers found 60% expect AI to become standard for online shopping.
Sources
Follow the coverage
Published reporting connected by Cybersecurity.
- Poland Halts ePUAP as Steam Queue Swells and Store Coolers Face Hacks
- OpenAI Releases ChatGPT Work Data as ARC Evals Tests Autonomous Agents
- Dreamteam Raises $53M as Anthropic Updates Claude and Amazon BuildsYou are reading this story
- Corma Hits $1B Value as Holiday Extras Uses AI & U.S. Caps Chips
Comments (0)
No comments yet. Be the first to share a thought on this story.
Request a correctionContact the newsroom