#美股AI概念股普涨

---

1️⃣ AI-concept US stocks recently bought on Gate and the trading ideas behind them

In Gate’s US stock trading section, I mainly focused on three targets: Micron (MU), Intel (INTC), and a small amount of Nvidia (NVDA). The specific ideas are as follows:

**Micron (MU) — The hardest storage logic**

- **Buying point thinking:** Micron surged nearly 10% in a single day on June 8, but before that it had already been pulling back for three straight weeks. Seeing that HBM capacity is still in short supply and that Micron has fully booked its 2026 HBM capacity, on June 8, Feeiheng (Fed) rebounded nearly 6%. MU led the storage sector, confirming a short-term sentiment inflection point.
- **How to trade:** On Gate, use USDT to directly buy MU shares, building the position in two entries—first, chase in a small amount within 2 hours after the rebound starts; second, add after a slight intraday pullback.
- **Risk control:** Set a stop-loss 5% below the June 5 low, with the expectation of holding until the end-of-quarter rebalancing window in late June.

**Intel (INTC) — An order-driven reversal bet**

- **Logic:** Google’s order for 3 million AI chips will only be delivered in 2028. Realizing the actual performance is still far off, but the market has chosen to “believe first, verify later.” Intel was previously deeply oversold, and the appearance of a huge order catalyst at low levels makes the short-term odds attractive.
- **Trade plan:** Participate with a small position on Gate using a quick in-and-out approach. After Intel jumped 11% that day, I didn’t chase. The next day, after observing that trading volume wasn’t meaningfully shrinking, I entered at a 3% pullback.
- **Note:** Don’t treat it as a long-term holding—this is more of an event-driven trade.

**Nvidia (NVDA) — Momentum-following allocation**

- NVDA rose only 1.73% on the day, lagging behind the sector. But as the absolute leader in AI computing power, it exists as a “core holding” in Gate’s US stock portfolio. The key idea is not to chase NVDA’s short-term excess returns—instead, use Gate’s fragmented trading features to invest small amounts of USDT every week to smooth out the cost.

---

2️⃣ With the big players joining forces, analysis of the future trend for AI-concept stocks

**Short term (1–3 months): Beware of “order-fueled hype,” divergence is inevitable**

After Intel announced orders for 3 million chips and surged 11%, it’s important to note: that order is not delivered until 2028, so the impact on Intel’s revenue and profit over the next two years is almost zero. This kind of “distant water quenches near thirst” rally is essentially a release of sentiment in an environment with abundant liquidity.

Possible divergence paths over the coming weeks:

- Stocks that truly benefit from current AI capital expenditure (Nvidia, Broadcom, Micron, TSMC) are expected to maintain strength—especially on the storage side, where the HBM capacity tightness is likely to continue at least through 2027.
- Stocks that rely on long-dated orders or are merely “riding the concept.” Once the broader market adjusts (for example, June CPI/PPI coming in above expectations triggering concerns about rate hikes), the pullback could be extremely sharp.

**Medium term (6–12 months): Energy and connectivity become “invisible champions”**

Most of the current discussion focuses on chips themselves, but power bottlenecks in AI data centers and connectivity bottlenecks are becoming even tighter constraints than the chips:

- **Power side:** Traditional grid expansion takes 5–7 years, gas turbine orders are already queued until 2029, and GE Vernova’s backlog reaches $150 billion. Meanwhile, Bloom Energy’s fuel cell contracts for power supply can go from signing to delivering power in just 55 days, and Q1 revenue growth is 130%. These names you can watch on Gate—GE Vernova (GEV), Bloom Energy (BE). Right now, capital attention is far lower than for chip stocks, which creates an expectations gap.
- **Connectivity side:** Optical module manufacturers such as Lumentum have already gained more than 145% within the year, but their valuations remain below those of chip design companies, leaving room for catch-up gains.

**Long term (1+ year): Key signals revealed by Cathie Wood’s moves**

Wood significantly reduced holdings of TSMC before the first 5 months of 2026 and realized profits on AMD, while building a position in Cerebras (wafer-level AI inference chips). This suggests a crucial judgment: AI computing demand is shifting from large-scale model training toward high-frequency, low-cost inference.

This means:

- The “barbaric growth” phase of training chips may be coming to an end, and the growth rate will slow.
- Inference chips, edge computing, and low-power solutions will become the next main battlefield.
- For individual investors, heavily weighting training-chip leaders (such as Nvidia) is still reasonable at present, but you should start paying attention to early opportunities on the inference side—using small position sizes on platforms like Gate to build exposure to sub-sectors that have not yet been fully priced.

**Overall judgment**

I’m positive on the medium- to long-term trend of AI-concept stocks, but over the next 6 months I’m more inclined to focus on segments “beyond computing power”—storage (Micron, SanDisk), power (GEV, BE), and optical connectivity (Lumentum). Their potential upside may be greater than that of chip design leaders that have already surged significantly. At the same time, closely watch May CPI data: if inflation comes in above expectations, tech stocks may face valuation pressure in the short term—at that point, it would be the time for phased entries rather than panic selling.

---

The above is my personal sharing of investment ideas and does not constitute investment advice. Gate’s USDT trading feature for US stocks provides a convenient cross-market allocation tool, but risks in derivatives such as leverage and options are high. For ordinary users, it is recommended to focus mainly on spot trading.