Brutal Market Warning: Fund Manager Exposes Weakest Sector, #1 AI Trade | Matthew Tuttle

This is not March 2000

Tuttle distinguishes the current AI rally from the dot-com bubble, citing real infrastructure demands and revenue generation, though acknowledges that calling the top is impossible and the trend could continue until hyperscalers cut spending.

The Broadcom warning signal

Broadcom's 14% drop despite strong earnings illustrates that expectations have moved from 'strong' to 'perfect,' creating an entry problem where even solid execution fails to support inflated multiples.

Position sizing over timing

Rather than predicting bubble peaks, investors should limit individual positions to 1-2% allocations to survive potential 50% corrections without catastrophic portfolio damage.

Avoid broad software ETFs like IGV

Tuttle criticizes the IGV software index for holding companies that AI will render obsolete, arguing investors must pick individual winners rather than owning 'everything' through passive funds.

Winners and losers in the AI transition

Microsoft, Palantir, ServiceNow, Salesforce, and cybersecurity names will thrive or remain additive to AI, while legacy software companies face obsolescence and crushed margins.

Bitcoin lost leading indicator status

Bitcoin's 50% decline from highs reflects rotation from prior 'bottleneck trades' into newer AI infrastructure plays rather than serving as a predictive signal for broader equity markets.

Memory is the critical constraint

DRAM contract prices jumped 60% in a single quarter and Micron rallied 275% from March lows because AI demand continues to outpace supply despite efficiency improvements like Google's Turbo Quant.

Peeling the onion on infrastructure

Beyond memory, investors should monitor sequential bottlenecks including photonics (data movement), space (data center locations), substrates, and inference optimization as each becomes the next constraint.

Halo assets provide portfolio balance

Complement high-growth tech with 'heavy asset, low obsolescence' infrastructure plays including utilities, copper, and oil companies that AI cannot replace and that benefit from data center energy demands.

SpaceX lessons for new listings

Investors should avoid IPO day one to prevent becoming 'exit liquidity' for pre-IPO insiders at firms like Fidelity, instead waiting for post-hype entry points on must-own names.

Anduril over OpenAI

Tuttle favors defense-tech unicorn Anduril for its position at the convergence of AI and military hardware over crowded software IPOs like OpenAI or Anthropic, citing the 'convergence trade' where drones and lasers replace traditional tanks and ships.

Critical materials watch

While not currently concerned about Strait of Hormuz disruptions affecting helium or semiconductor materials, Tuttle monitors geopolitical supply risks that could emerge as secondary bottlenecks.