This confidence, which is a giant with an artificial intelligence -backed technology, does not come slightly when the rod is already too high. Behind this movement has a pile of AI that already works hot and pushes the current supply limits.
No need to say, bets rose and Wall Street’s expectations have increased.
Morgan Stanley calls Nvidia the best chip choice for the increase in earnings in the second half of 2025. Justin Sullivan & Sol; Getty Images
As business and cloud players are locked in advanced capacity, as they pass through the old plugs, Nvidia’s Blackwell GPUs are not a secret that they kill him.
Even with the fact that AMD has recently increased the MI350 prices to stretch Trust, Nvidia still dominates the GPU debate.
Trendorce data projects, where Blackwell can potentially represent 80% of NVIDIA’s senior GPU shipments this year.
This points to a large ramp that started in the second quarter and continued at the closing of the year.
A big swing in terms of product mixture, but emphasizes how fast the adoption moves.
Nvidia -backed stocks such as Coreweave and Dell have recently confirmed this change with the first large -scale commercial distribution of Blackwell Ultra.
Beyond that, it is difficult to miss the performance of performance.
Nvidia claims that Blackwell offers a 40 -fold leap in AI training and inference performance compared to Hopper.
The ultra version takes one or two notches, increases the yield of “AI factory” in a close to 50x, and gives cloud formats a strong new base line in raw calculation and execution.
Surprisingly, even the middle -layer Blackwell systems, including the B200, have shown a faster training rate of close to 57% than the H100. In addition, for self -survivor, they reduced the cost of workload, which is a major incentive in the midst of tightening AI budgets.
This demand is already real, concrete income.
Related: NVIDIA AI Outlook is changing on meta platforms, Microsoft plans
Nvidia, Blackwell Chip sales, underlining the enormous power of the cycle, leaving the summit of Hopper’s summit, he said.
The combined effect is that Nvidia continues to squeeze her grip in the AI.
By offering unique performance and superior price-exit rates, it allows cloud providers to upset competitors in the process, while allowing cloud providers to access more margins, eliminating NVIDIA’s ecosystem dominance.
However, perhaps the biggest change for Nvidia recently is its progress to continue its H20 chip exports to China.
For the context, the H20 is a special -made version of Nvidia’s hopper architecture, which is effectively compatible with US export controls.
At the same time, although not as developed as Blackwell, it is still punching in high -performance AI acceleration.
Related: Morgan Stanley re -calibrates his opinion on Apple Stock before earnings
However, it fills a critical performance layer for China’s AI market, where local alternatives are still delayed.
The US allowed NVIDIA to license H20 sales over and over again as a part of rare land -access negotiations.
Analysts marked this as a meaningful step to alleviate technology tensions.
Taking advantage of the development, Nvidia ordered a 300,000 large unit H20 with TSMC, a sign of directness in demand.
However, friction is also difficult to ignore.
While China’s cyber security investigation enters the H20, the US national security voices already forced the White House to rethink.
Nevertheless, no official restriction has been restructured, and the early active position of Nvidia gives him the first crack of resurrection.
Yes, supply restrictions remain, but this is often a signal of rise.
Even the limited output is absorbed efficiently. The cyber security investigation in China’s H20 enters the noise, while being new.
More news:
More importantly, no ban was issued. Nvidia remains on the market with the opportunity to deepen its basis before domestic competitors are scaled.
Return from Washington is expected, but so far the policy continues well and Nvidia is the only AI chip supplier who benefits immediately.
Nvidia received a great vote of confidence from Morgan Stanley and has to do with Blackwell.
Experienced analyst Joseph Moore, confirming the overweight note in the AI giant, increased its 12 -month price target from $ 170 to $ 200.
The call, published on July 30, is based on what Moore defines Nvidia as “extraordinary” demand for the new generation of Blackwell AI chips.
Related: Teceli analyst Tesla Post-Winnings reduces a 3-word decision
This is actually the demand that continues to exceed the supply.
Moore’s team measured this movement by multiplying the 2026 valuation floors.
Nvidia’s mid -year EPS estimation from $ 6.02 to $ 28x’ten 33x earned.
This mathematics emphasizes Nvidia’s position as the best chip stock game of Morgan Stanley, while this mathematics is effectively locked on an excellent price target.
The note points to healthy and clear calculation needs in both the inference and the large model AI workloads.
Customer controls show the investment of infrastructure developed from hyper -scale and large business buyers, which forces it to adopt AI faster and more broader than expected.
Although the current supply problems are a head wind, Moore quickly changes it.
Production and logistics are expected to heal in the back half of this year and unlock a major step in yield and gain acceleration.
This amazing dynamic is what Moore calls “convex gain leverage ..
In short, the executive validity will quickly compound the opposite in the Nvidia stock.
