To The Moon

OpenAI and Anthropic Just Reignited the AI Hardware Trade

The AI trade is rotating back toward hardware.

OpenAI's GPT-6 Astra launch and Anthropic's latest research breakthroughs have helped revive a simple investment thesis: frontier AI progress still requires enormous physical infrastructure.

Better AI still needs more compute

GPT-6 Astra delivered another major jump in reasoning and scientific capabilities. More important for hardware investors, $NVIDIA(NVDA)$   CEO Jensen Huang said Astra was trained on more than 100,000 Nvidia Grace Blackwell NVLink72 GPUs and that another 400,000 GPUs are "coming online next."

Anthropic is sending a similar capability signal, with Claude recently advancing a problem related to the Riemann hypothesis and completing the first full computer-checked formalization of Fermat's Last Theorem.

The takeaway is simple: model progress has not broken the link between intelligence and compute.

Memory and optics wake up

The market is starting to trade that message.

Nvidia moved back to within roughly 1.5% of its record high, while memory stocks rebounded sharply, led by $SK hynix(SKHY)$,$Micron Technology(MU)$ and $SanDisk Corp.(SNDK)$.

Optical names such as $Lumentum (LITE.US)$ and $Coherent (COHR.US)$ also rallied. Larger AI clusters need more high-speed connectivity, making optics another direct beneficiary as GPU deployments scale.

Neoclouds join the rebound

$CoreWeave (CRWV.US)$ and $IREN Ltd (IREN.US)$ also posted strong gains.

The logic is straightforward: if frontier labs need compute faster than hyperscalers can build it internally, neoclouds remain an important source of incremental GPU capacity.

That makes them one of the highest-beta ways to trade another acceleration in AI infrastructure demand.

Power is part of the same trade

$Bloom Energy (BE.US)$ surged 7.35% during Friday's regular session and gained another roughly 5% after S&P Dow Jones Indices announced it would join the $S&P 500 Index (.SPX.US)$ .

The index catalyst added fuel, but the broader AI power thesis remains intact. Hundreds of thousands of additional GPUs also mean more data-center electricity demand, making power availability a critical constraint on the next stage of AI deployment.

What investors should watch

The key question is whether this becomes a sustained rotation back into AI infrastructure.

If frontier models keep scaling, the beneficiaries extend well beyond Nvidia: memory feeds the compute, optics connects it, neoclouds add capacity and power keeps everything running.

@TigerStars  @CaptainTiger  @TigerWire@Daily_Discussion  @Tiger_chat  @Tiger_comments  @MillionaireTiger