Lede: A Carnegie Mellon study uncovered 6 million fake stars created by 301,000 accounts across 18,617 repositories, with AI and LLM projects receiving the highest volume among non-malicious categories. Paragraph 1: The ICSE 2026 paper by CMU, NCSU and Socket researchers analyzed 20 terabytes of GitHub data spanning 2019-2024 and noted a surge in 2024, when 16.66% of repositories with over 50 stars were affected (ICSE 2026). This builds upon Dagster's 2023 hands-on investigation that identified vendors through Google searches and tested retention rates from budget and premium services including GitHub24 at EUR 0.85 per star (Dagster, March 2023). What much of the original coverage missed was the direct tie to venture capital practices, as Redpoint and other firms run scrapers using star counts with a median of 2,850 at seed stage as sourcing signals (AwesomeAgents.ai, 2025). Paragraph 2: Marketplaces price stars between $0.03 and $0.85 depending on account quality, with services like Fiverr gigs offering packages from $5 and Telegram channels enabling easy access; FTC's 2024 rule on fake social influence carries $53,088 penalties per violation while SEC has charged founders for inflated traction (FTC, 2024; AwesomeAgents.ai, 2025). Independent analysis of 150 profiles across 20 projects showed 36-76% of stargazers exhibiting zero followers and fork-to-star ratios 10x below organic baselines (AwesomeAgents.ai, 2025). Mainstream reporting has under-emphasized these patterns in favor of anecdotal blockchain coverage. Paragraph 3: Synthesizing these sources reveals AI repositories accumulated 177,000 fake stars and 78 manipulated projects reached GitHub Trending; GitHub deleted 90.42% of flagged repositories and 57.07% of flagged accounts by January 2025 (ICSE 2026). This acceleration coincides with the AI boom and converts $0.06 stars into $1-10 million seed rounds, distorting open-source discovery and funding pipelines in patterns previous coverage failed to connect across studies.