AI stocks surge: Nebius +29%, CoreWeave +20%

Thursday, 13 August 2026 at 18:40

AI-aandelen stijgen Nebius +29% en CoreWeave +20%

The latest AI stock rally isn’t being driven by a flashy chatbot, a new language model, or a viral app. This time, it’s all about the AI engine room: data centers, servers, memory chips, and the rental of scarce computing power.

Nebius jumped more than 29 percent Wednesday afternoon. CoreWeave gained nearly 20 percent and Supermicro over 19 percent. Micron and Nvidia also moved clearly higher.

The timing is no coincidence. CoreWeave and Supermicro released fresh numbers Tuesday evening, followed by Nebius with its own quarterly results on Wednesday morning. All three sent roughly the same message: demand for AI infrastructure is still outpacing the capacity most companies can add. Softer U.S. inflation data gave the broader market an extra tailwind.

Five stand-out AI stocks on August 12, 2026

Company	Ticker	Price around 8:00 p.m.	Change	Main catalyst
Nebius	NBIS	$249.83	+29.3%	Revenue up 454 percent
CoreWeave	CRWV	$108.37	+20.0%	Record revenue and $104 billion backlog
Supermicro	SMCI	$37.64	+19.1%	Strong margins and bullish guidance
Micron	MU	$931.20	+7.2%	Renewed rally in AI memory and chips
Nvidia	NVDA	$223.89	+2.9%	Positive signals from customers and infra partners

Prices are snapshots. Final closing levels may differ.

The AI supply chain in one line: Nvidia designs the accelerators, Micron supplies critical memory, Supermicro builds full server systems, and Nebius and CoreWeave rent out the resulting compute power.

1. Nebius soars 29 percent on explosive revenue growth

Nebius was the day’s biggest winner and has a clear Dutch link. The company is based in Amsterdam, listed on Nasdaq, and builds specialized cloud infrastructure for AI companies.

Quarterly revenue hit $582.3 million, up 454 percent year over year. Adjusted EBITDA, a broad gauge of operating performance, swung from a $21 million loss to a positive $236.2 million.

Nebius signed four major AI cloud deals during the quarter, each averaging more than $1 billion. That boosts visibility on future income—and explains investors’ excitement.

Still, not every number was pretty. Under GAAP, the company posted a net loss of $190.4 million. Nebius also invested about $5.7 billion in GPUs, data centers, and related infrastructure. Revenue is scaling at breakneck speed, but it requires extraordinary spending.

That makes Nebius compelling to watch—and highly sensitive to delays, financing costs, and shifting expectations. A single-day jump of nearly 30 percent shows how eager the market is, but also how volatile the stock can be.

2. CoreWeave doubles revenue—still in the red

Alongside Nebius, CoreWeave is one of the best-known “neoclouds”—cloud providers that don’t try to offer everything like Amazon Web Services or Microsoft Azure, but focus primarily on GPU capacity for AI.

CoreWeave posted second-quarter revenue of $2.58 billion, up from $1.21 billion a year earlier. By the end of June, it had a revenue backlog of roughly $104 billion—contracted services to be recognized in future periods.

Early in the third quarter, CoreWeave says it added more than $25 billion in new customer commitments. During the quarter it brought nearly 500 megawatts of active capacity online, bringing the total to roughly 1.5 gigawatts.

Impressive, yes—but CoreWeave also shows why the AI infrastructure race is financially risky.

The company reported adjusted EBITDA of $1.51 billion, yet under GAAP it recorded a net loss of $626 million. Net interest expense alone was $640 million. CoreWeave also lifted its 2026 capex outlook to $35–$39 billion.

Revenue and backlog prove real demand exists. They don’t yet prove every dollar poured into data centers will translate into ample free cash flow.

3. Supermicro surprises with profit and aggressive outlook

Supermicro doesn’t rent out cloud capacity itself, but it builds a crucial chunk of the physical systems that power the cloud. The company combines processors, Nvidia GPUs, memory, networking, and cooling into complete server racks and datacenter solutions.

Fourth-quarter revenue hit $11.1 billion, up from $5.8 billion a year earlier. Gross margin jumped from 9.5 to 17.5 percent. Net income came in just under $1.18 billion.

For fiscal year 2027, Supermicro guides revenue between $65 and $72 billion. Analysts surveyed by Reuters were expecting around $52.5 billion. That gap explains much of the stock’s surge.

Not everything was flawless. Quarterly revenue came in slightly below expectations. The company said some shipments were delayed as customers waited on power provisioning, cooling, and network gear.

That detail matters. The AI bottleneck is no longer just chips. Entire projects now stall over transformers, grid connections, liquid cooling, fiber links, and the physical build-out of datacenters.

Execution is therefore critical for Supermicro. The market is expanding fast, but large orders can easily slip between quarters when a single datacenter component isn’t ready on time.

4. Micron rides the AI memory boom

Powerful AI chips need massive amounts of fast memory. High-bandwidth memory—HBM—is especially critical for training and running large AI models.

Micron jumped more than 7 percent on Wednesday. Unlike Nebius, CoreWeave, and Supermicro, Micron didn’t report new quarterly results that day. The move appears to be part of a broader rally in chip and infrastructure stocks.

The underlying numbers explain why investors are including Micron in the rally. In its most recent quarter, the company posted $41.46 billion in revenue, up from $9.30 billion a year earlier. Datacenter revenue topped $25 billion.

Micron also reported over $1 billion in revenue from its latest HBM4 memory. The company expects the datacenter DRAM and NAND market in 2026 to be more than twice the size of two years earlier.

The risk: memory is a cyclical business. If manufacturers add too much capacity, prices and margins can quickly fall. Micron is trying to limit that with multi-year contracts, price floors, and customer prepayments, but today’s valuation assumes tight supply will persist.

5. Nvidia climbs too—big test ahead

Nvidia rose about 2.9 percent on Wednesday to $223.89. That’s far less than Nebius, CoreWeave, and Supermicro—but the signal may be more important than the single-day move suggests.

Nebius and CoreWeave buy and rent out large volumes of Nvidia hardware. Supermicro builds complete systems around Nvidia platforms. When these companies say order books are swelling and customers want more capacity, the market reads it as a positive omen for Nvidia.

Nvidia posted record quarterly revenue of $81.6 billion, up 85 percent year over year. Datacenter revenue jumped 92 percent to $75.2 billion. The company reports its next quarter on August 26.

That makes Nvidia the next major test of the AI rally. Investors will focus on demand for new Blackwell and Rubin systems, revenue guidance, gross margins, and any signs of customer delays.

China remains a risk. In its prior outlook, Nvidia assumed zero datacenter processor revenue from China. With a market cap around $5.46 trillion, exceptional future growth is already priced in—meaning small disappointments can trigger big moves.

What this trading day really shows

The takeaway isn’t just that five AI stocks climbed. The numbers show an investment wave moving through the entire stack.

Nvidia sells accelerators and networking. Micron supplies the memory those systems require. Supermicro assembles the servers, racks, and cooling. Nebius and CoreWeave stitch it all into datacenters and rent the capacity to AI firms.

That’s stronger proof of real economic activity than a rising valuation for a popular chatbot. Contracts are being signed, servers shipped, and billions spent.

At the same time, risk is cascading through the chain. These companies must borrow and invest heavily before a datacenter generates revenue. They depend on a handful of large customers and face shortages in chips, power, construction capacity, and cooling.

Anyone researching these stocks should look beyond a single day’s move. Key checkpoints include:

how much of the order book is under binding contracts;
the level of concentration in the largest customers;
how much debt and interest are needed to fund growth;
whether gross margins and free cash flow are actually improving;
how much new capacity is already generating revenue versus still under construction.

Wednesday’s rally proves AI infrastructure demand is red-hot right now. It doesn’t prove that every datacenter project, every loan, and every stock is a good bet at any price.

AI can help you compare earnings, annual reports, and filings faster. Just don’t let a chatbot decide what to buy. In our in-depth guide Investing with AI, we explain how to use AI as a research assistant while you continue to verify sources, numbers, and assumptions yourself.