Earn up to 50% APR
Boost earnings with NFTs
Play, HODL & earn more
Grok's AI Stock Portfolio Is Up 59% in Nine Months. Claude's Is Up 2.6% in Two Months
-
Two AI-managed stock portfolios running live on the Autopilot mirror-trading platform are producing starkly different results, offering one of the most concrete public comparisons of how different AI models translate market data into investment decisions. Grok's portfolio, managed autonomously by the xAI-branded agent under the @grkportfolio account, has returned 59% over its first nine months with $17 million currently invested — compared to the S&P 500's 36% gain over the same window. Over the past three months alone, Grok added 12.6% against the SPY's 9.75%. The outperformance has been driven by concentrated positions in semiconductor, memory, defense, and power stocks — a direct bet on the hyperscaler capital expenditure cycle that has driven AI infrastructure spending to record levels. Grok's approach is described as a multi-stage pipeline: macro forecasting, fundamental screening of approximately 1,000 stocks, deep research using real-time web and X search, scenario modeling with price targets, and portfolio construction — the full analyst workflow automated end-to-end.Claude's portfolio launched in April 2026 with $50,000 in seed capital and has underperformed significantly in its first two months, returning approximately 2.6% while the S&P 500 gained 8.3% over the same period.
The Claude agent rotated into enterprise software names like ServiceNow and Zeta Global while trimming Microsoft, leaning into what operators describe as second-order AI plays in enterprise software, fintech, and power. That tilt missed the mega-cap semiconductor rally that produced Grok's headline returns. The performance gap reflects different model philosophies as much as different stock picks — Grok's concentrated infrastructure bet captured the most aggressive part of the AI-driven rally, while Claude's more hedged, probability-weighted approach prioritized scenario discipline over momentum. The comparison covers very different time periods — nine months for Grok versus two months for Claude — making direct conclusions premature, though the directional difference is real.
