Nvidia’s GTC showmanship revives the AI trade, and traders smell a clean setup
San Jose felt like theatre on Monday. Jensen Huang, Nvidia’s chief executive, spent two hours at GTC 2026 describing “AI factories”, agentic systems and physical AI. Meanwhile, the market treated the keynote as a permission slip to lean back into the AI momentum book.
Nvidia shares traded in the $186 to $199 zone around the event, after a choppy patch that left plenty of traders hunting for a tidy dip. However, the mood shifted as Huang tied new products to a familiar promise. Accelerated computing cuts costs, expands capability and pulls more workloads onto Nvidia’s stack.
He avoided any single headline-grabbing bookings figure. Instead, he stacked breadth on breadth. Huang talked up an ecosystem of roughly 450 partner companies and framed the prize as “a hundred trillion dollars of industry”. Therefore, the pitch was less about one quarter’s demand and more about Nvidia as default infrastructure.
That matters because the market still trades Nvidia like a capex proxy. If hyperscalers keep spending, Nvidia prints. If they pause, the multiple compresses first and questions arrive later. So traders listened for a tell on durability, and the tone stayed bullish.
Targets drift higher, but valuation keeps its teeth
Consensus remains a tailwind. Across widely followed analyst averages, the central case still points to roughly 20% to 40% upside from the high-$180s. Meanwhile, the spread between highs and lows stays wide enough to keep options desks busy.
At the same time, Nvidia is no longer “cheap” in the lazy sense. A forward multiple around the high-20s can look sensible next to growth forecasts. However, it also leaves little forgiveness for any capex wobble or supply hiccup. Therefore, execution still matters more than storytelling, even when the storytelling is good.
Technicals look like a reset, not a breakout yet
Price action around the $180s has started to look like a pivot zone, with traders watching the 20-day, 50-day and 200-day averages converge nearby. RSI readings have sat near neutral rather than stretched. Consequently, the chart reads like a coiled spring, but not one that has definitively launched.
In plain terms, the bulls have a level to defend. Meanwhile, the bears need a clean breakdown to argue the keynote bump was just fumes.
Palantir has tried to surf the same enterprise AI wave, announcing partnerships across data and services, including names like Databricks, Accenture, Snowflake and Lumen. That kind of list supports the adoption narrative. However, traders should treat any stale price level as worthless until the live tape confirms it.
Therefore, PLTR remains more of a momentum watch than a thesis anchor. If Nvidia is infrastructure, Palantir is a workflow bet, and the market will price them differently when volatility rises.
What to watch next
GTC did what it usually does. It refreshed the story, widened the runway and reminded everyone that Nvidia sells the picks, shovels and operating manual. However, the next clean catalyst is still earnings, where “beat and raise” is the only language that keeps the multiple comfortable.
- For dip buyers, the $180 area matters as near-term support, because it anchors the current reset.
- For momentum traders, strength above the recent $190s band would matter more than any slide deck.
- For options desks, the wide target range keeps skew and defined-risk structures relevant.
- For everyone, capex tone from hyperscalers remains the macro switch for the whole AI complex.
By the numbers
- NVDA price zone discussed: $186 to $199
- Partner count cited on stage: about 450
- Upside implied by many 12-month targets: roughly 20% to 40% from the high-$180s
- Forward P/E referenced by bulls and bears: about 27x
- Keynote: March 16, GTC 2026 week
Key takeaways
- Nvidia’s keynote reinforced ecosystem lock-in, which supports the “buy the dip” reflex.
- Valuation still demands clean execution, so position size should match your tolerance for capex shocks.
- Support near $180 is the line traders will fight over; a hold invites upside tests.
- Palantir fits the enterprise AI narrative, but it needs real-time confirmation before chasing.
