Nvidia's Two-Gigawatt Australia Bet Lands in a Market That Has Stopped Cheering
Published on 09/10/2026 at 17:21 | Editorial boerse-global.de
Nvidia's operating machine keeps producing numbers that ought to end the argument. Revenue for the second quarter of fiscal 2027 came in at USD 96.22 billion, up roughly 106 percent year over year and comfortably ahead of the USD 92.17 billion consensus. Data centers delivered USD 89 billion of that total — more than 92 percent of the business. Guidance for the third quarter sits at about USD 108 billion, give or take two percent.
And yet the stock is not celebrating. Shares changed hands at EUR 192.40 the previous session, down 0.9 percent, capping a week in which they shed 2.1 percent. That leaves the price about five percent below its 52-week high of EUR 202.50. A separate reading put the equity at EUR 187.80, 1.7 percent above its 50-day average of EUR 184.70 — a market digesting headlines without euphoria. Annualized volatility of 41 percent tells its own story about how risky investors still consider the position.
A pipeline stretching from Sydney to Jakarta
The latest expansion news landed Thursday: Nvidia intends to scale its Australian data-center capacity for AI workloads to as much as two gigawatts by 2027, more than doubling the country's current load. The build-out runs through the company's DSX platform, announced a day earlier alongside Australian NVIDIA Cloud Partners, and covers land, power and building capacity across multiple generations of that infrastructure. Partners include Firmus, CDC, NEXTDC and AirTrunk. A broader roster of eight Australian collaborators — Firmus, Sharon AI and IREN among them — is attached to the same target, with Sharon AI alone planning to install up to 68,000 Nvidia GPUs. Indonesia is on the map too, where partner Zankore is pushing a USD 3.1 billion project.
The timing fits a pattern rather than standing alone. Nvidia has already disclosed a deal to acquire Hugging Face for about USD 12.93 billion, including a retention program worth up to USD 1 billion for staff, with completion targeted for the first half of 2027 pending regulatory clearance. October brings the first Windows PCs carrying the RTX-Spark chip, built with Lenovo and Acer, while a deeper tie-up with CrowdStrike extends the company into agent-based cybersecurity.
Should investors sell immediately? Or is it worth buying Nvidia?
The skepticism has concrete sources
Gene Munster of Deepwater framed the dilemma neatly this week: strong numbers can still unsettle the market, because an overly bullish guide feeds fears of overinvestment, while a modest raise gets read as growth normalizing. Either way, skeptics find room to work.
Two structural concerns give that skepticism substance. Customer concentration is the first — research indicates a substantial share of second-quarter revenue, measured at the end of July, came from just two unnamed large customers, well above the year-earlier quarter. A book that lopsided leaves Nvidia exposed to demand shifts at individual hyperscalers, however impressive the headline growth. The second is regulatory. According to a New York Times report, the US Department of Justice is investigating Nvidia's licensing arrangement with AI chip startup Groq, a USD 17 billion deal that may have been structured to sidestep antitrust review. A formal request for information has already been issued; a fine is conceivable, while an unwinding looks unlikely. On its own it is not existential — but it adds to a growing pattern of scrutiny aimed at Nvidia's market position.
Costs are climbing while capacity is planned
Supply chain economics are shifting too. Nvidia is reportedly paying premiums to lock in test and probe card capacity with suppliers such as FormFactor and King Yuan through 2027, while lead times for R&D probe cards have stretched from eight weeks to more than twenty. Memory prices, according to one Freedom Capital analyst, are likely to stay elevated for years. That squeezes margins even with data-center operating margins currently running at 71 to 72 percent.
None of this is an immediate revenue lever in either direction. The Australia program is capacity planning for coming years, not a near-term earnings driver, and the Hugging Face purchase remains subject to approval with completion not expected until the first half of 2027. If AI demand grows more slowly than partners assume, the risk is oversized investment.
What actually settles the argument
Execution is the only real arbiter. Should Nvidia bring the Australian capacity online as planned, it would meaningfully diversify its data-center footprint and open demand markets outside the United States. Marrying DSX hardware dominance with Hugging Face's open developer platform could push the company deeper into the software layer of the AI value chain, binding developers more tightly to its silicon — and the CrowdStrike partnership signals ambitions in the high-margin security segment as well.
The near-term markers are concrete: October's RTX-Spark launch with Lenovo and Acer, followed by regulatory progress on Hugging Face. A stumble on either — a blocked acquisition or a visible delay in Australia — would invite the market to interrogate the growth story far more aggressively. For now the fundamental strengths still outweigh the noise: revenue growth, expanding demand and global partnerships speak plainly. But the accumulation of smaller disruptions — customer concentration, the DOJ probe, rising supplier costs — means the narrative is no longer as frictionless as it looked a year ago. Anyone holding the stock should assume the market is watching more closely than the numbers alone would suggest.
Ad
Nvidia Stock: New Analysis - 10 September
Fresh Nvidia information released. What's the impact for investors? Our latest independent report examines recent figures and market trends.
