Following SoftBank’s full divestment and Bridgewater’s significant reduction of its stake in NVIDIA, Peter Thiel, a billionaire and one of the most influential investors in Silicon Valley, has also exited his position entirely. The consecutive stake reductions by multiple investment giants undoubtedly send a clear signal of caution to the fervent AI market.

According to the latest 13F holding report submitted to the U.S. Securities and Exchange Commission (SEC), Thiel Macro LLC completely withdrew from its investment in NVIDIA in the third quarter. This move is particularly eye-catching as it occurred when NVIDIA’s market capitalization surpassed $5 trillion and Wall Street analysts generally regarded it as an “unshakable” leader. Moreover, NVIDIA is set to release its latest earnings report on November 19.

In tandem with the complete divestment, there was a thorough reshaping of the investment portfolio. The fund significantly slashed its total U.S. stock holdings, reducing them from approximately 212millioninthesecondquartertojust74.4 million. At the same time, the fund established substantial new positions in Microsoft and Apple. Major Portfolio Overhaul: Clearing Out the AI Leader and Reducing Total Holdings by Two-Thirds
Thiel Macro LLC’s operations in the third quarter were quite aggressive. The documents show that the fund not only completely sold off its NVIDIA stake, which accounted for a hefty 40% of its investment portfolio weight in the previous quarter, selling all 537,700 shares, but also divested another significant holding, Vistra Energy, which made up 19% of its portfolio weight.

This move stands in stark contrast to NVIDIA’s own strong fundamentals. According to reports, NVIDIA’s quarterly sales soared from 39.3billionto46.7 billion, with its data center business revenue growing by 56%. However, Thiel’s choice was clearly based on his judgment about future valuation rather than current performance.

Instrument the end-to-end path

Measure queueing, execution, retries and final user outcome rather than only component health. Traces should make it possible to explain why a request was slow or wrong from the user’s perspective.

Version important configuration

Prompt templates, model routes and data dependencies can change behavior without code changes. Versioning these inputs makes regressions reproducible and rollback safer.

Prefer bounded failure

When dependencies degrade, reduce optional work or return a clear fallback instead of allowing latency and error rates to cascade through the whole system.

Define the production contract

Agree on expected quality, latency, cost and failure behavior before optimizing components. A service can look efficient in isolation while violating the user-facing objective.

Fallback and rejection behavior should be explicit so edge cases do not improvise at runtime.

Make state ownership explicit

Models, caches, queues and databases often hold overlapping representations. Each durable state should have one authoritative owner and a version so stale copies cannot silently win.

Test degradation and recovery

Simulate unavailable dependencies, stale caches and overload. A production system needs predictable fallbacks and a clear path back to normal operation rather than manual repair for every incident.

Instrument the end-to-end path

Measure queueing, execution, retries and final user outcome rather than only component health. Traces should make it possible to explain why a request was slow or wrong from the user’s perspective.

Version important configuration

Prompt templates, model routes and data dependencies can change behavior without code changes. Versioning these inputs makes regressions reproducible and rollback safer.

Prefer bounded failure

When dependencies degrade, reduce optional work or return a clear fallback instead of allowing latency and error rates to cascade through the whole system.