Hedge funds are buying these two big tech stocks while selling two rivals
NVIDIA Corporation (NASDAQ:NVDA) trades near $183.68 (-1.56%) ahead of its FQ3 2026 earnings release on November 19, as the market braces for one of the most closely watched reports of the year. The company, now valued at $4.53 trillion, stands at the heart of both Wall Street’s AI optimism and its growing anxiety over valuation, with Nvidia representing 7.6% of the S&P 500’s total weight.
Recent weeks have seen rising volatility, triggered by hedge funds trimming positions, public short bets from Michael Burry, and SoftBank and Peter Thiel fully exiting their Nvidia stakes. Yet, fundamentals remain staggering—Street forecasts $55B in FQ3 revenue (up 56.8% YoY) and $61.6B for FQ4, maintaining momentum across its GB300 and Rubin product ramps.
Nvidia’s management reaffirmed its outlook for a $3–4 trillion AI infrastructure spending wave through the end of the decade, a claim that aligns with cumulative global data center capex projections of $4.8 trillion (2026–2029).
In Q2 FY2026, Nvidia’s revenue hit $46.7B (+56% YoY), with its Data Center division contributing 88% ($41.1B)—a historic concentration of sales. The company projects Q3 revenue of $54B (+61% YoY), driven by hyperscaler demand from Amazon (NASDAQ:AMZN), Meta (NASDAQ:META), and Alphabet (NASDAQ:GOOG).
Meta’s massive Prometheus and Hyperion data campuses (up to 5 GW capacity) and Microsoft’s $3.3B Mount Pleasant expansion illustrate a CAPEX supercycle that directly channels into Nvidia’s GB300, Blackwell, and Rubin chips. Nvidia’s market share in AI accelerators remains near 90%, giving it command of the hardware backbone for this $4T opportunity.
CEO Jensen Huang’s roadmap underscores Nvidia’s evolution beyond GPUs—into full AI infrastructure. The GB300 ramp, producing over 1,000 racks per week, now powers hyperscalers like CoreWeave (CRVW) and Nebius (NBIS). The upcoming Rubin platform, scheduled for volume production in H2 2026, is designed to sustain Nvidia’s compute dominance amid rising competition from AMD’s MI450 and Broadcom’s (NASDAQ:AVGO) custom XPU systems.
Each generational leap compounds Nvidia’s performance lead: B200 (Blackwell) delivers 3x faster training and 15x higher inference throughput than the H100 (Hopper), while Rubin Ultra (2027) and Feynman (2028) are expected to push compute density to unprecedented levels.
CFO Colette Kress guided GAAP gross margins at 73.3% ±50 bps, and non-GAAP at 73.5%, with management targeting mid-70s% margins through year-end—a figure few hardware firms can sustain at this scale.
At 42x forward earnings, NASDAQ:NVDA trades at a premium—43% above the tech sector median—yet below peers like AMD (NASDAQ:AMD) on growth-adjusted metrics. Nvidia’s projected FY27 revenue of $300B and FY28 revenue of $360B imply a forward PEG ratio near 1.1, reflecting continued faith in AI monetization.
However, the slightest earnings miss or cautious guidance could trigger a double-digit selloff, given how crowded the trade has become. Hedge funds including Coatue Management have reduced exposure, while short interest remains modest at 0.98%, suggesting that sentiment—not fundamentals—is the bigger risk.
While Nvidia’s software revenue remains small—$1B in FY2024 (2.1% of Data Center)—its strategic value is immense. The CUDA ecosystem, used by over 4 million developers, remains unmatched in performance optimization. Nvidia’s AI Enterprise license ($4,500 per GPU/year), combined with tools like TensorRT, NeMo, and BioNeMo, enhances throughput and latency, ensuring deep customer stickiness.
By 2029, software sales could reach $5.5B, still under 1.5% of total revenue, but serving as a high-margin moat that anchors Nvidia’s hardware dominance. The decision to bundle most software offerings free of charge reinforces ecosystem lock-in—forcing hyperscalers and sovereign clients to remain within Nvidia’s infrastructure stack.
Geopolitical risk remains a recurring overhang. The GAIN AI Act under the new U.S. administration could mandate domestic supply prioritization and impose a 15% revenue share tax on approved exports. Nvidia’s CFO clarified that this is not yet codified, stating, “I don’t have to do this 15% until I see a regulatory document.”
Nonetheless, H20 shipments to China—potentially $2–$5B per quarter—remain suspended amid Chinese state directives halting purchases “on security grounds.” The halt, combined with Huawei’s ramp-up via SMIC, intensifies competition in the Chinese data center market, though early reports show Huawei’s AI processors underperforming in large model training compared to Nvidia’s H20 and Blackwell B30 units.
Consensus estimates point to FQ3 revenue of $54.9B (+57% YoY) and EPS of $1.26 (+55%), followed by Q4 guidance of $61.5B (+56%). Analysts expect Nvidia to exceed guidance again—its typical beat rate averaging 6–8% per quarter since 2023.
However, investor sensitivity is extreme. With the S&P 500 underperforming and Nvidia carrying the heaviest weight in the index, any deviation in Rubin’s timeline, China exposure, or expense growth beyond 40% YoY could ignite volatility.
Still, hyperscaler CAPEX growth of 65.4% in 2025—led by Meta, Microsoft, and Amazon—anchors visibility. Nvidia’s Sovereign AI revenue alone could surpass $20B in 2026, doubling YoY, even without Chinese contribution.
Technically, NASDAQ:NVDA trades within a descending triangle since late October, forming lower highs near $212 and holding support at $180.50. RSI hovers near 46, signaling neutrality before earnings. A breakout above $192.60 would open momentum toward $205.80, while a miss could drag shares to $172–$165, the latter aligning with the 200-day EMA.
Despite this consolidation, volume distribution shows institutional accumulation in the $175–$185 range, signaling quiet conviction beneath the noise.
Advanced Micro Devices (NASDAQ:AMD) has gained traction via its MI450 GPU and OpenAI partnership for Helios racks, while Broadcom (NASDAQ:AVGO) secures four hyperscaler XPU contracts. Yet Nvidia’s networking segment ($7.3B revenue) and full-stack integration (compute + interconnect) keep margins superior to peers.
In China, Huawei and Cambricon are accelerating domestic chip programs under state-backed mandates, but technical limitations—particularly in memory and interconnect bandwidth—maintain Nvidia’s edge.
Nvidia’s share of global data center capex rose 7x since 2020 to 26.5% in 2024, solidifying its dominance.
After integrating all operational, technical, and macro data, TradingNews maintains a Strong Buy rating on NASDAQ:NVDA, anchored by data center acceleration, sovereign AI contracts, and unmatched software ecosystem control.
The short-term risk of volatility remains high due to extreme market focus, but structural fundamentals—56–62% revenue growth, 73% gross margins, and a $4T+ addressable infrastructure market—define Nvidia as the dominant force in global compute.
With FQ3 earnings imminent and hyperscaler CAPEX still compounding above 20% CAGR, Nvidia’s current consolidation near $183.68 represents a high-risk but high-conviction entry window. The long-term price target remains $317, supported by expanding TAM, network dominance, and sustained innovation in Rubin and AI Enterprise software.
That’s TradingNEWS.com