NVIDIA Just Posted Perfect Earnings. So Why Is Nobody Buying?
NVIDIA delivered a flawless quarter — 73% revenue growth, $78B guidance that crushed estimates — and the stock barely moved. The reason? AMD is about to change the game.
•NVIDIA posted record Q4 revenue of $68.1B (up 73% YoY) and guided $78B for next quarter — crushing Wall Street estimates by $6B
•Despite perfection, the stock barely moved — investors see decelerating growth ahead as capex budgets hit their ceiling
•AMD's MI450 launches in H2 2026 with Meta and OpenAI already committed to gigawatt-scale deployments
•Salesforce's sleepy earnings spooked Wall Street, Trade Desk lost 77% of its value in a year, and C3 AI may be heading for insolvency
Perfection Wasn't Enough
NVIDIA just did something almost no company in history has done: posted a virtually perfect earnings report — and watched the market shrug. Q4 fiscal 2026 revenue came in at $68.1 billion, up 73% year-over-year. Diluted EPS hit $1.76, nearly doubling from $0.89 a year ago. And then came the guidance: $78 billion for Q1, a full $6 billion above what analysts expected [1]. That's not a beat. That's a demolition.
And yet Jensen Huang's stock sat there. Flat. Maybe twitched a little. If you're wondering why a company that just printed $44 billion in operating income in a single quarter can't catch a bid, the answer is both simple and uncomfortable: the market has already priced in the miracle. What it's pricing in now is the slowdown.
The Math Problem Nobody Can Ignore
Here's the core issue. NVIDIA's revenue engine runs on one thing: hyperscaler capex. Amazon, Meta, Google, and Microsoft have been spending at breakneck pace to build out AI infrastructure — and nearly all of that money has funneled straight to NVIDIA. But there's a ceiling. Amazon already guided $100 billion-plus in capex for 2026. Meta is in a similar range. These companies are generating enormous cash flows, but even they can't double their infrastructure budgets indefinitely without loading up on debt or gutting shareholder returns.
The math is straightforward: if your biggest customers can only increase spending modestly in 2027 and 2028, your 73% growth rate becomes a 30% growth rate, then a 15% growth rate. The stock market doesn't wait for that to happen — it prices it in the moment the trajectory becomes visible. And right now, the trajectory is visible.
Hyperscaler capex has been NVIDIA's rocket fuel — but budgets have ceilings.
Enter Lisa Su and the MI450
But the capex ceiling isn't even the biggest threat. The biggest threat is a woman named Lisa Su. AMD's MI450 GPU launches in the second half of 2026, and the specs are not messing around: 40 petaflops of FP4 compute, 432 GB of HBM4 memory, and 19.6 TB/s bandwidth per chip [2]. AMD claims 50% more memory capacity than NVIDIA's upcoming Vera Rubin architecture. Whether that benchmark holds up in real-world workloads remains to be seen — but the direction of travel is clear.
More importantly, AMD isn't just building chips and hoping for buyers. They've already locked in commitments. Meta announced a 6-gigawatt deal starting with a 1GW deployment in H2 2026 using a custom MI450 variant built specifically for Meta's workloads [3]. OpenAI committed to deploying up to 6 gigawatts of AMD Instinct GPUs, also starting in H2 2026 [4]. These aren't speculative partnerships — they're signed deals with the two most important AI companies on the planet.
This puts NVIDIA in a classic innovator's dilemma. They can hold their margins and let AMD eat share, or they can slash prices to compete and watch their legendary profitability erode. Either way, the days of NVIDIA having zero competition in data center AI are over. Since April 2025 lows, AMD stock is up roughly 170% versus NVIDIA's 103%. The market rotation has already begun.
Salesforce: Sleepy Beats in a Scared Market
If NVIDIA's problem is that perfection wasn't enough, Salesforce's problem is that "fine" is now a death sentence. Revenue came in at $11.2 billion, up 12% year-over-year. Non-GAAP EPS of $3.81 beat estimates. On paper, these are acceptable numbers [5]. But income from operations grew just 3% because expenses — R&D up 14%, sales and marketing up 16%, G&A up 18% — are outstripping revenue growth. That's the kind of math that makes investors nauseous.
The Anthropic stake is an interesting wildcard. Salesforce owns roughly 1% of Anthropic after investing about $330 million, which generated an $811 million gain on strategic investments this quarter alone. If Anthropic's valuation trajectory continues toward the trillion-dollar range that some analysts project, that stake becomes a $10 billion asset sitting inside a $170-180 billion market cap company. It's not nothing.
Salesforce also announced a $50 billion buyback — an enormous number relative to its current market cap. If Benioff truly believes Agentforce will push the company toward $60 billion in annual revenue and beyond, buying back shares at current prices is a generational opportunity for the company. If he's wrong, it's a catastrophic misallocation of capital. The buyback is essentially a bet on Benioff's own conviction.
Trade Desk: From $90 to $20 in Six Months
Trade Desk might be the most painful chart in tech right now. The stock hit $91 in August 2025. It's now trading around $20 — a 77% decline in roughly six months [6]. Q4 revenue was solid at $846 million (A-minus income statement by most standards), but the revenue growth trajectory tells the real story: 25% in Q1 2025, 19% in Q2, 18% in Q3, 14% in Q4. Q1 2026 guidance? About 10%.
That's a deceleration so smooth you could draw it with a ruler. And the market is extrapolating: if you're at 10% now, where are you in a year? 5%? Flat? The company burned $1.4 billion in buybacks at an average price of $52.60 — more than double where the stock trades today. That's money that could have been spent building or acquiring. Instead, it evaporated. Trade Desk isn't broken, but it's facing a fundamental question about whether programmatic advertising can grow in an AI-disrupted media landscape.
C3 AI: A Cautionary Tale in Real Time
If you want to understand the distance between "AI company" and "company that makes money from AI," look at C3 AI. Revenue dropped from $98 million to $53 million year-over-year. Gross profit collapsed from $58 million to $9 million. Net loss ballooned to $133 million — meaning they lost more than double what they brought in on the top line [7]. Let that sink in: for every dollar of revenue, C3 AI lost $2.50.
The company announced a 26% workforce reduction and $135 million in annual cost cuts, but at this burn rate with $621 million in cash, the clock is ticking. Three years ago, C3 AI was mentioned in the same breath as Palantir. Since then, Palantir returned 1,700% to investors while C3 AI went from $36 to $8. Same industry label. Completely different execution. It's a brutal reminder that the "AI" label doesn't guarantee anything.
Earnings season separates the real from the hype — and this week delivered both extremes.
The Hidden Winner: Revolve Group
Buried under the NVIDIA and Salesforce headlines is a company almost nobody is talking about: Revolve Group (RVLV). The e-commerce fashion retailer posted Q4 net sales of $324 million (up 10% YoY), net income of $18.6 million (up 58%), and diluted EPS of $0.26 — beating estimates by 62% [8]. For 2025, the company generated $1.23 billion in revenue, sits on $303 million in cash with zero debt, and saw free cash flow jump 157%.
The early 2026 momentum is even more impressive: net sales up 16% through the first seven weeks. In clothing retail — an industry where 5% growth is considered healthy — that's exceptional. Active customers hit 2.84 million, the highest quarterly increase in three years. Beauty sales surged 43%. Owned brands now represent 20% of the REVOLVE segment. This is a company executing on every metric while trading at a fraction of the attention that AI stocks command.
The Bigger Picture
This earnings cycle is a masterclass in market psychology. NVIDIA posted arguably the best income statement in corporate history — and the stock did nothing because the market is forward-looking, and the forward view includes competition and decelerating growth. Salesforce posted "fine" numbers and got punished because "fine" doesn't cut it when investors are nervous. C3 AI proved that slapping "AI" on your company doesn't make you Palantir. And Revolve is quietly printing money while everyone argues about GPU architectures.
The lesson here isn't about any single stock. It's about the difference between a great company and a great investment at a given price. NVIDIA is a great company — maybe the greatest earnings machine in history right now. But at a $5 trillion market cap with competition finally arriving, the easy money has been made. The interesting money is in figuring out who benefits next.
On this page
Web · https://www.technewsworld.com/story/assessing-amds-2025-momentum-and-its-ces-2026-reveals-180068.html
