Every week brings another bold claim about where the next big winners in AI will emerge. But sometimes the most important signals come not from the headlines — but from the way the industry reallocates its resources. That's what's happening right now. | Nvidia's dominance isn't over. But for the first time in years, it's no longer the only story. A different part of the AI stack is gaining traction — and the companies positioned there are beginning to move faster than investors expected. | And that makes Eric Fry's warning — and his new "Nvidia alternative" pick — far more relevant than most realize: |
|
| A MESSAGE FROM OUR PARTNER | When Eric Fry picks a stock, it often starts moving quickly. | One month ago, Eric went live with a controversial call. | "Sell Nvidia," he said. | It's been a great stock, but its biggest customers are turning into its biggest competitors. And that could spell doom. | Instead of Nvidia, Eric said to buy a more off-the-wall hardware pick that's garnering almost zero media attention. | Which is shocking when you realize that in a single AI data center, there is enough of this company's hardware to circle the globe up to 8 times! | Since Eric went live with this recommendation, this little-known "Nvidia alternative" is already beginning to soar. | In fact, it's returned 18% over just the past 30 days while NVDA has gone negative – just as Eric predicted! | | | | But, Eric says, it is not at all too late to jump into this exciting – and totally misunderstood – stock. | He believes it could soon overtake the growth trajectories of all the Mag 7 stocks. | To get details on this exciting company while it's still cheap – plus 7 more timely trade ideas straight from Eric – click here now. |
|
| | What's Really Changing in AI Hardware | The clearest sign that the market is changing came through in November's earnings cycle. Nvidia is still growing — fast — but the tone has shifted. Growth remains massive, but the easy trajectory is gone. Demand for Blackwell chips is overwhelming, yet the bottlenecks around packaging, memory, and power density are starting to dictate the pace. | That's why hyperscalers are no longer waiting. | Amazon launched a $110 million program to pull researchers onto Trainium 2. Microsoft-backed d-Matrix released a commercial inference product that skips the CUDA ecosystem entirely. Cerebras continued securing funding for wafer-scale systems designed to bypass GPU clusters altogether. | None of these players aim to replace Nvidia outright. They're carving out territory around it — and in doing so, they're changing the structure of the entire hardware market. |
|
|  | | | Presented by Investor Place | |
|
|
|
|
|
|
|
| Why Big Tech Is Pulling Away From Nvidia | While most attention stays on the chips, the more urgent developments are happening one step underneath. | At the SC24 supercomputing conference, the dominant topic wasn't model training — it was cooling. Rack densities above 100 kW are pushing air systems to their limit, and data centers are scrambling to adopt liquid-to-chip setups. Vertiv and nVent rolled out new systems specifically built for Blackwell-class loads. On the communications side, the shift to optical interconnects is accelerating. As clusters expand to tens of thousands of GPUs, copper simply can't keep up. Companies like Broadcom, Marvell, Coherent, and Lumentum are supplying the components that allow these machines to talk to each other without choking bandwidth. | These are the segments absorbing most of the new demand — long before it shows up in quarterly results. |
|
| The Companies Powering the AI Factory | The next stage of AI growth depends less on GPUs and more on the systems that keep them running. Cooling leaders like Vertiv and nVent are becoming essential as data centers hit power and heat levels older infrastructure can't support. Their liquid-to-chip designs, introduced at SC24, give operators a workable path to scale. In networking, Broadcom and Marvell are enabling the optical links required for clusters that now span tens of thousands of accelerators. As copper hits physical limits, these suppliers are stepping into a critical role. | Meanwhile, Cerebras and d-Matrix are positioning themselves for specialized workloads — wafer-scale training on one side, high-volume inference on the other — filling the gaps the GPU ecosystem can't reach alone. | These companies are shaping the foundation that allows the entire AI stack to expand. |
|
| | | | | | | | | Final View | AI's hardware landscape isn't splintering — it's restructuring. Nvidia will remain central, but the fastest-moving opportunities are forming around the suppliers handling power, cooling, connectivity, and inference efficiency. | Investors who only watch the chip race will miss the broader shift.
The build-out of the "AI factory" is already underway — and the companies enabling that build-out often move first. |
|
| | | Written by Deniss Slinkins
Global Financial Journal |
|
|
|
Tidak ada komentar:
Posting Komentar