Hedge funds are selling tech stocks like the AI trade is officially over.
According to Goldman Sachs, hedge funds are selling technology stocks at a pace we haven't seen in nearly a decade. Their long to short ratio in tech has also fallen to its lowest level since 2016, which tells you just how aggressively institutions are reducing their exposure. And I think a lot of this has to do with how crowded the AI trade became over the past couple of years. Many of these funds made enormous amounts of money investing in AI companies, but now they're worried about protecting those gains as concerns around valuations, AI spending, and the sustainability of the entire AI infrastructure buildout continue growing. And to be fair, I can understand some of those concerns. We're talking about hundreds of billions of dollars being invested into AI infrastructure, and investors are increasingly questioning when these companies will start seeing meaningful returns on that spending.
But here's what I find interesting. A lot of these companies are still reporting incredible revenue growth, demand for AI infrastructure remains strong, and their underlying businesses haven't necessarily deteriorated nearly as much as their stock prices would suggest. I think we're seeing a combination of legitimate concerns and hedge funds simply wanting to get out of the way before things potentially get worse. And here's the other side of this. When institutions have already sold this aggressively, you have to wonder how much selling pressure is actually left. If earnings continue holding up and sentiment improves, these same funds could eventually have to come back and buy the stocks they just sold. Now, I'm not saying we've reached the bottom, because things could absolutely get worse from here. But I do think we're getting to a point where sentiment is becoming increasingly disconnected from the long term potential of some of these companies. Personally, I'm still bullish on AI over the long term. I just think the market is going through a pretty significant reset right now, and for patient investors, that could eventually create some pretty interesting opportunities. And if you’re trying to figure out what to buy through all this volatility, Milk Road just launched PortfolioOS, a new investing platform built to help you research stocks, track your portfolio, understand what’s driving your positions, and make better investment decisions in one place. It also comes with FREE Milk Road PRO access.
milkroad.com/portfolioos/?ut…
The latest Fed meeting minutes are out and I think what's happening right now is pretty interesting.
Almost every Fed official agreed that additional tightening was necessary, with a couple even suggesting that a rate hike larger than 25 basis points could be justified in December. Some officials were already pushing for a rate hike in October, and several remain concerned that inflation could stay elevated for longer than expected. But here's the weirdest part of all this. Oil prices have actually been coming down. Crude oil is now trading around $82 a barrel, down more than 21% over the past six months. You would normally expect falling energy prices to help ease inflationary pressures, yet the Fed is still discussing the possibility of even more aggressive rate hikes.
To me, this raises an interesting question. Is the Fed seeing inflationary pressures elsewhere in the economy that are much more persistent than investors realize, or are they risking overtightening into an economy that's already showing signs of weakness? Either way, I think the market is in a pretty difficult position right now, especially with all the uncertainty we're already dealing with.
$MSFT and $INTC just expanded their partnership under a long term agreement aimed at Azure infrastructure for growing cloud and AI demand.
The official releases are clear about what actually changed.
Microsoft is the FIRST company to say it will use Intel Foundry's EMIB advanced packaging for Azure first party silicon, and the agreement confirms multi year Xeon supply, including next generation Diamond Rapids server CPUs.
They are also deepening work on Xeon with CXL memory, QuickAssist, and confidential computing. Microsoft says Azure M series v4 virtual machines can show up to 2.3x greater performance with CXL and Intel Flat Memory Mode, and that SQL backups can run up to 2.5x faster on the first QuickAssist enabled virtual machines in the cloud.
$MSFT used Ignite to show it can build its own AI models and still be the place everyone else's models run.
Microsoft announced seven new MAI models today, with the coding lineup at the center. It says MAI Code 2 reaches 98% of the coding quality of GPT-6.1 Sol at about 40% of the price, and 96% of Claude Opus 5.5 at 5% of the price. Those cost figures use promotional pricing.
MAI Code 2 Flash is the high volume model for everyday edits and routine fixes. Microsoft says it is comparable to GPT-6 Luna at 65% of the price, also using promotional pricing. Both coding models are in GitHub Copilot, and Code 2 Flash is in a limited access preview in Microsoft Foundry.
Microsoft also said AT&T built a custom MAI model for 5G network analytics that outperformed GPT-6.1 Sol at 50 times lower cost.
Satya Nadella said GPT-6.2 is coming to Foundry, and Mark Zuckerberg wrote that Muse Wonder, Muse Spark 1.3, and Muse Ember are coming soon to Microsoft Foundry on Azure.
We think this is $MSFT trying to be both a model maker and a model marketplace. That could reduce how much it has to rely on OpenAI, while cheaper specialized models drive more usage on Azure, including models from $META.
NVIDIA has now completely erased all of its gains following yesterday's earnings, with more than $150 billion in market cap wiped out in just the past hour.
I personally thought management did a pretty decent job addressing most of the concerns surrounding the company, especially with Jensen Huang's departure, circular financing, and the sustainability of AI spending. The numbers were incredible, guidance was strong, and demand doesn't appear to be slowing down anytime soon.
But clearly, the market doesn't care.
I guess one thing management could have done better was address the China revenue situation and provide more clarity on what that opportunity could look like going forward.
Still, it's pretty crazy to see a company deliver numbers like NVIDIA did yesterday and have the market react this way. Just goes to show how high expectations have gotten for this company and, more importantly, how much uncertainty still lies ahead.
I know everyone's been getting beaten down by this market lately and if you're down for the year, we have just the thing for you.
We just launched PortfolioOS, and I genuinely think this is one of the coolest things we've built at Milk Road.
Imagine having an AI research assistant that knows exactly what stocks you own, understands your risk tolerance, and can tell you whether buying more NVIDIA, Palantir, or any other stock actually makes sense for YOUR portfolio.
Instead of spending hours researching companies, you can use the same 12 AI research skills our analysts use to analyze valuations, stress-test investment ideas, and identify risks you might have overlooked.
And the best part is that it works directly with ChatGPT, Claude, Grok and other AI models.
I believe AI is going to completely change how retail investors research stocks, and this is a pretty big step in that direction.
For the next 7 days, we're offering lifetime access for 60% off, plus 3 months of Milk Road PRO for free.
It won't magically recover your losses, but it might help you make better informed decisions going forward
milkroad.com/portfolioos/?ut…
Memory is now more than half the cost of an AI accelerator.
According to Marvell, HBM takes up roughly 30% of an AI chip's area, and at 2026 memory prices it makes up MORE than half of its cost.
That is a massive shift. A few years ago, the GPU was the expensive part and the memory was an afterthought.
Marvell's answer is custom HBM. Moving some of the logic into a custom base die can free up to 25% more area for compute.
The tradeoff is heat. Putting logic under the DRAM stack means the base die, the interface and the packaging all have to be designed together.
This is why I keep saying memory is the bottleneck of the AI buildout. Save this one.
$MRVL$MU$NVDA
$MU is dealing with a split in the memory market that is becoming harder to ignore.
According to CFM Flash Market, memory makers cut prices this week on eMMC, LPDDR4X, and some industrial SSDs because demand is weak.
Buyers are mostly drawing down inventory, and actual transactions are limited and the cuts are concentrated in older, lower end products.
eMMC 256GB fell from $52.50 to $48.50, a drop of about 7.6%. Industrial SATA 512GB SSDs fell from $110 to $95, down almost 14%. LPDDR4X 32Gb and 64Gb were also marked down, while every LPDDR5X price on the list held flat.
NAND wafers and DDR5 chips were unchanged this week, and DDR5 modules are still sitting at record highs. PC makers in China and Taiwan remain cautious on new orders.
This lines up with what TrendForce just said. It raised its 2027 forecast for server DRAM, but it cut consumer DRAM and NAND because higher memory prices are hurting phones and PCs.
We think this is the split to watch for $MU, $SNDK, and $WDC. HBM, server DRAM, and DDR5 still look tight.
The weakness is showing up first in the legacy parts that go into phones, PCs, and older industrial systems.
According to Korean media, Apple is close to signing a four to five year DRAM supply deal with Samsung, starting in the first half of 2027.
The deal reportedly guarantees Apple a MINIMUM volume of DRAM, with prices set later based on market conditions.
This is a big deal because Apple almost never does this. It is famous for short contracts and pushing suppliers for price cuts at every renewal.
Apple was reportedly the only Big Tech company without a long term agreement with Samsung. It is also said to have already signed one with SK Hynix.
When the toughest buyer in the world starts locking in supply for five years, it tells you who has the power in memory right now.
$AAPL$MU
Why would Morgan Stanley downgrade memory stocks while admitting demand is still strong?
According to Commercial Times, Morgan Stanley cut its ratings on Winbond, Macronix, Phison and PSMC, and lowered its Nanya target price to NT$550.
Their argument is not that demand is weak today but they want to get out BEFORE any signs of slowing cloud capex show up, and their model has memory supply and demand gradually normalizing by 2028.
We understand the logic, but we think it ignores what the biggest memory makers are saying. SK Hynix says its DRAM and HBM are sold out through 2029, and it has long term agreements covering more than half of its 2027 and 2028 capacity.
These stocks have already been cut hard. $MU is down about 36% from its June high, SK Hynix about 58% and Samsung about 43%.
At some point a lot of bad news is already priced in.
According to Korean media, TrendForce just laid out its 2027 memory outlook, and this is why I have stayed bullish.
NAND prices are expected to start falling in the second quarter of 2027 as China's YMTC aggressively expands capacity.
DRAM is a different story. TrendForce does not expect DRAM prices to turn lower through the end of 2027.
It sees conventional DRAM prices rising 10% to 15% this quarter, then 5% to 10% in the first quarter, 3% to 8% in the second, and 0% to 5% in the third and fourth quarters of 2027.
The number I am watching is HBM4. According to TrendForce, pricing is expected to rise from at least $2.05 per Gb this year to $3.00 to $5.50 per Gb next year. At the high end, that is more than a 100% jump. Talks around HBM4E are at $5 per Gb or higher.
This is the squeeze I keep talking about. HBM uses about three to four times as much wafer capacity as conventional DRAM, and TrendForce says its share of DRAM wafers at Samsung, SK Hynix and $MU is set to go from 23% this year to almost 30% next year.
When HBM and server memory take more of the wafer, conventional DRAM stays tight even if NAND starts to weaken.
That is a BIG signal for $MU, Samsung and SK Hynix.
If you own ANY AI stocks, you need to read this.
“Our momentum is real. Our demand is accelerating. Our supply is growing. Our customers and our partners are with us. And employees are committed.” That was how NVIDIA's interim CEO Colette Kress ended what might be one of the most important earnings calls in the company's history. NVIDIA came into this quarter facing some serious concerns. Jensen Huang's departure, circular financing, the sustainability of AI spending, declining margins, export restrictions, and whether the company could continue executing without its founder. And I think management did a pretty good job addressing most of them.
First, the numbers were absolutely incredible. NVIDIA reported $116 billion in quarterly revenue, more than doubling year over year and growing 20% sequentially. Data center revenue reached $108 billion, up 21% quarter over quarter, while Vera Rubin generated nearly $20 billion in its first quarter of production shipments, making it the fastest product ramp in NVIDIA's history. Management also guided for $130 billion in Q4 revenue and raised its fiscal 2028 revenue growth outlook from 70% to 75%. Think about that for a second. We're talking about a company already generating more than $100 billion in quarterly revenue that still expects 75% annual growth in fiscal 2028. One of the biggest concerns was obviously Jensen Huang's departure. Interestingly, Jensen was only mentioned twice by name during the entire call. It almost felt like management wanted to distance the conversation from Jensen and shift investors' attention back to the business. The board explicitly stated that Jensen would not return as CEO and refused to discuss the ongoing investigation. Instead, Colette emphasized that NVIDIA's employees, leadership team, customers, and suppliers remain fully committed. She even mentioned that TSMC's leadership had been in NVIDIA's offices this week. The company also confirmed that Vera Rubin Ultra remains on track for next year, with Feynman expected in 2028. Now, despite the board saying Jensen isn't coming back, I personally still think there's a small chance we could eventually see him return in some capacity. I wouldn't bet on it, but I wouldn't completely rule it out either.
Another major concern was circular financing, particularly NVIDIA's financial relationships with neoclouds and other AI infrastructure companies. Investors have been worried that NVIDIA is essentially financing customers who then purchase its GPUs. Management clarified that many of its guarantees only take effect once customers have secured financing and their infrastructure is ready for deployment. If those conditions aren't met, certain commitments can be reduced or eliminated. Colette also confirmed this wasn't a new policy, just additional disclosure explaining how the arrangements work. That doesn't eliminate the risk, especially with NVIDIA carrying $118 billion in guarantees and $32 billion in neocloud related commitments, but it does provide more transparency. And speaking of neoclouds, I thought this was one of the most bullish parts of the entire call for companies like $NBIS and $CRWV. NVIDIA's AI Cloud, Industrial, and Enterprise segment generated $47 billion in revenue, growing 18% sequentially and nearly tripling year over year. Management emphasized that neoclouds are becoming increasingly important to the AI ecosystem, helping enterprises, governments, and AI labs access computing capacity.
NVIDIA revealed that Hopper GPUs are being recontracted at up to twice their original rental prices, while B300 rental rates have increased more than 60% since introduction. Management also confirmed that its compute remains fully utilized across every hyperscaler and neocloud it serves. That's a pretty big deal because one of the biggest bearish arguments against neoclouds has been that their expensive GPUs would rapidly lose value as newer hardware became available. Instead, we're seeing older GPUs retain significant economic value while newer hardware commands higher rental rates. $NBIS jumped more than 12% following those comments, and I think it's pretty clear why. If these pricing and utilization trends continue, the economics for well positioned neocloud operators could be considerably better than many investors expected. Another concern was whether the AI infrastructure boom was approaching its peak. NVIDIA pushed back with some pretty incredible numbers. Cloud industry backlog increased from $2 trillion to $2.5 trillion in just three months, while hyperscaler cloud revenue growth accelerated for the sixth consecutive quarter. Meanwhile, token volumes on OpenRouter increased 35 times year over year, with AI agents now accounting for approximately 80% of tokens. Token prices have also declined nearly 60% since May, encouraging significantly greater usage. Essentially, AI is becoming cheaper to use, but consumption is growing so quickly that demand for computing infrastructure continues to accelerate.
Margins were another major topic. NVIDIA expects Q4 gross margins to decline to 71%–72%, but management explicitly described that quarter as the bottom, with margins expected to recover to 72%–73% in fiscal 2028. Colette explained that NVIDIA had honored existing purchase orders despite rising supply costs and expects updated pricing to help stabilize margins. That's reassuring considering the concerns around memory costs, advanced packaging, and competition from custom AI chips. Management also addressed supply constraints, confirming that NVIDIA has secured additional memory and advanced packaging capacity. Vera Rubin is being deployed roughly twice as fast as Grace Blackwell, with hyperscalers bringing the platform into production in approximately one month. NVIDIA is also prioritizing customers that have secured financing, sufficient power, and data centers ready to deploy its hardware. That's particularly important because demand means very little if customers can't actually install and operate the infrastructure.
There were also questions about a potential slowdown in 2028 due to political uncertainty, electricity constraints, and opposition to data center construction. Management acknowledged those risks but emphasized that customers are still requesting more supply than NVIDIA can currently deliver. The company also established a National Security and Export Control Oversight Committee to strengthen engagement with the US government. Importantly, its guidance already excludes future licensed Hopper shipments to China, although the latest export restrictions are still being evaluated. Another encouraging development was capital allocation. NVIDIA returned $28 billion to shareholders during the quarter, including $22 billion in buybacks and $6 billion in dividends. It also increased its repurchase authorization by $150 billion, leaving $227 billion available at quarter-end. Management intends to execute that authorization by the end of fiscal 2028 while continuing to invest in supply chain expansion and selected infrastructure partnerships.
Overall, I think NVIDIA delivered exactly the kind of earnings call investors needed following the uncertainty of the past week. Management addressed nearly every major concern, raised its long-term growth outlook, and showed that demand for AI infrastructure remains exceptionally strong. Of course, Jensen's departure is still a major uncertainty, circular financing risks haven't completely disappeared, and customers will eventually need to demonstrate sufficient returns on their enormous AI investments. But when you have a company generating $116 billion in quarterly revenue, guiding for $130 billion next quarter, and still expecting 75% revenue growth in fiscal 2028, it's pretty difficult to argue that the underlying business is falling apart. Whether that's enough to reverse the recent stock decline remains to be seen, but I think NVIDIA gave the bulls plenty of ammunition tonight.
$NVDA JUST REPORTED EARNINGS, AND THEY ABSOLUTELY CRUSHED EXPECTATIONS.
Revenue came in at $115.8 billion versus expectations of $108.8 billion, a beat of roughly $7 billion. EPS came in at $2.68 versus expectations of $2.47, while data center revenue reached a staggering $107.6 billion, beating estimates of $101.87 billion. Adjusted gross margins also came in at 74.9%.
These are some incredible numbers, especially considering all the uncertainty surrounding NVIDIA over the past few weeks following Jensen Huang's departure, concerns about circular financing, and questions about the sustainability of AI spending. But here's the thing. Beating expectations was only half the battle. The market also wants reassurance that demand remains strong, customers are continuing to spend aggressively, and NVIDIA can maintain its growth trajectory without Jensen at the helm.
For me, the next quarter guidance and management commentary are going to be just as important as these results. NVIDIA needed to deliver a strong quarter, and so far, it looks like they've done exactly that. Now let's see if they can convince Wall Street that the AI trade still has plenty of room to run.
Nvidia's supply commitments jumped from $279 billion last quarter to $390 billion this quarter.
Nvidia says that increase is mainly tied to MULTI YEAR purchases of memory.
Think about what that means. The biggest buyer of AI chips in the world is locking up memory for years at today's prices, because it doesn't think supply will catch up anytime soon.
Nvidia also expanded deals with Samsung Electro-Mechanics and Kinsus for substrates, Amkor and SPIL for packaging, plus TSMC and Foxconn.
Revenue came in at $115.8 billion, up 103% from a year ago, and Nvidia guided to $130 billion next quarter.
This is the clearest signal yet for $MU, Samsung and SK Hynix. Demand is still running FAR ahead of supply.
$NVDA$MU
The most important customer story in $NVDA's earnings may be the one sitting next to the hyperscalers.
Nvidia's Data Center revenue hit $107.6 billion and inside that, Hyperscale was $60.1 billion, up 71% from a year ago, while AI Clouds, Industrial, and Enterprise reached $47.5 billion, up 195% year over year, according to official CFO commentary.
Those customers are growing faster than they can easily finance. Nvidia said it has $32 billion of neocloud agreements through 2032, which take effect only if customers secure financing, deploy the capacity, and Nvidia accepts it. It also disclosed $118 billion of maximum guarantees for land, power, shell, and hardware financing, including up to $105 billion of credit support for a 4.25 gigawatt SB Energy campus in Ohio under 20 year leases to OpenAI. Nvidia said it has not made any payments on those guarantees.
After the report, Yahoo Finance said $CRWV, $NBIS, $AMD, $AVGO, and $TSM all rose in after hours trading.
Jensen Huang might actually be making a comeback to NVIDIA and things are getting pretty interesting behind the scenes.
According to a new exclusive report from Semafor, some NVIDIA board members are continuing to push for Jensen Huang to return as CEO, despite the company publicly stating that it has no plans to bring him back. Apparently, there's a faction within the board that believes reinstating Jensen would be the right decision, which suggests the board might not be as united as NVIDIA wants investors to believe. Jensen himself has already said he would be honored to return if the board decided it was in NVIDIA's best interests. So the biggest obstacle right now isn't necessarily Jensen's willingness to come back, but rather whether the directors opposing his return can be convinced to change their minds.
Meanwhile, NVIDIA is reportedly considering other candidates, including Arm CEO Rene Haas and Marvell CEO Matt Murphy. However, Haas reportedly has concerns about taking the job while there's still a public campaign to bring Jensen back. And honestly, I can't blame him. Imagine becoming CEO of NVIDIA while half the industry is calling for your predecessor to return. Another interesting detail is that Jensen still owns roughly 3.5% of NVIDIA, which could potentially give him some influence over the board through shareholder voting mechanisms. That doesn't guarantee anything, but it means he's not entirely powerless in this situation.
Personally, I think bringing Jensen back would be a massive positive for NVIDIA, especially given everything that's happened over the past week. The company just delivered incredible earnings, but there are still questions surrounding its leadership, the ongoing investigations, and whether it can maintain the same execution without the person who helped build it into what it is today. I'm not saying Jensen is definitely coming back, especially with the investigations still hanging over everything, but I think the possibility is becoming harder to dismiss. And if more directors start supporting his return, things could change pretty quickly.
If you own ANY AI stocks, you need to read this.
“Our momentum is real. Our demand is accelerating. Our supply is growing. Our customers and our partners are with us. And employees are committed.” That was how NVIDIA's interim CEO Colette Kress ended what might be one of the most important earnings calls in the company's history. NVIDIA came into this quarter facing some serious concerns. Jensen Huang's departure, circular financing, the sustainability of AI spending, declining margins, export restrictions, and whether the company could continue executing without its founder. And I think management did a pretty good job addressing most of them.
First, the numbers were absolutely incredible. NVIDIA reported $116 billion in quarterly revenue, more than doubling year over year and growing 20% sequentially. Data center revenue reached $108 billion, up 21% quarter over quarter, while Vera Rubin generated nearly $20 billion in its first quarter of production shipments, making it the fastest product ramp in NVIDIA's history. Management also guided for $130 billion in Q4 revenue and raised its fiscal 2028 revenue growth outlook from 70% to 75%. Think about that for a second. We're talking about a company already generating more than $100 billion in quarterly revenue that still expects 75% annual growth in fiscal 2028. One of the biggest concerns was obviously Jensen Huang's departure. Interestingly, Jensen was only mentioned twice by name during the entire call. It almost felt like management wanted to distance the conversation from Jensen and shift investors' attention back to the business. The board explicitly stated that Jensen would not return as CEO and refused to discuss the ongoing investigation. Instead, Colette emphasized that NVIDIA's employees, leadership team, customers, and suppliers remain fully committed. She even mentioned that TSMC's leadership had been in NVIDIA's offices this week. The company also confirmed that Vera Rubin Ultra remains on track for next year, with Feynman expected in 2028. Now, despite the board saying Jensen isn't coming back, I personally still think there's a small chance we could eventually see him return in some capacity. I wouldn't bet on it, but I wouldn't completely rule it out either.
Another major concern was circular financing, particularly NVIDIA's financial relationships with neoclouds and other AI infrastructure companies. Investors have been worried that NVIDIA is essentially financing customers who then purchase its GPUs. Management clarified that many of its guarantees only take effect once customers have secured financing and their infrastructure is ready for deployment. If those conditions aren't met, certain commitments can be reduced or eliminated. Colette also confirmed this wasn't a new policy, just additional disclosure explaining how the arrangements work. That doesn't eliminate the risk, especially with NVIDIA carrying $118 billion in guarantees and $32 billion in neocloud related commitments, but it does provide more transparency. And speaking of neoclouds, I thought this was one of the most bullish parts of the entire call for companies like $NBIS and $CRWV. NVIDIA's AI Cloud, Industrial, and Enterprise segment generated $47 billion in revenue, growing 18% sequentially and nearly tripling year over year. Management emphasized that neoclouds are becoming increasingly important to the AI ecosystem, helping enterprises, governments, and AI labs access computing capacity.
NVIDIA revealed that Hopper GPUs are being recontracted at up to twice their original rental prices, while B300 rental rates have increased more than 60% since introduction. Management also confirmed that its compute remains fully utilized across every hyperscaler and neocloud it serves. That's a pretty big deal because one of the biggest bearish arguments against neoclouds has been that their expensive GPUs would rapidly lose value as newer hardware became available. Instead, we're seeing older GPUs retain significant economic value while newer hardware commands higher rental rates. $NBIS jumped more than 12% following those comments, and I think it's pretty clear why. If these pricing and utilization trends continue, the economics for well positioned neocloud operators could be considerably better than many investors expected. Another concern was whether the AI infrastructure boom was approaching its peak. NVIDIA pushed back with some pretty incredible numbers. Cloud industry backlog increased from $2 trillion to $2.5 trillion in just three months, while hyperscaler cloud revenue growth accelerated for the sixth consecutive quarter. Meanwhile, token volumes on OpenRouter increased 35 times year over year, with AI agents now accounting for approximately 80% of tokens. Token prices have also declined nearly 60% since May, encouraging significantly greater usage. Essentially, AI is becoming cheaper to use, but consumption is growing so quickly that demand for computing infrastructure continues to accelerate.
Margins were another major topic. NVIDIA expects Q4 gross margins to decline to 71%–72%, but management explicitly described that quarter as the bottom, with margins expected to recover to 72%–73% in fiscal 2028. Colette explained that NVIDIA had honored existing purchase orders despite rising supply costs and expects updated pricing to help stabilize margins. That's reassuring considering the concerns around memory costs, advanced packaging, and competition from custom AI chips. Management also addressed supply constraints, confirming that NVIDIA has secured additional memory and advanced packaging capacity. Vera Rubin is being deployed roughly twice as fast as Grace Blackwell, with hyperscalers bringing the platform into production in approximately one month. NVIDIA is also prioritizing customers that have secured financing, sufficient power, and data centers ready to deploy its hardware. That's particularly important because demand means very little if customers can't actually install and operate the infrastructure.
There were also questions about a potential slowdown in 2028 due to political uncertainty, electricity constraints, and opposition to data center construction. Management acknowledged those risks but emphasized that customers are still requesting more supply than NVIDIA can currently deliver. The company also established a National Security and Export Control Oversight Committee to strengthen engagement with the US government. Importantly, its guidance already excludes future licensed Hopper shipments to China, although the latest export restrictions are still being evaluated. Another encouraging development was capital allocation. NVIDIA returned $28 billion to shareholders during the quarter, including $22 billion in buybacks and $6 billion in dividends. It also increased its repurchase authorization by $150 billion, leaving $227 billion available at quarter-end. Management intends to execute that authorization by the end of fiscal 2028 while continuing to invest in supply chain expansion and selected infrastructure partnerships.
Overall, I think NVIDIA delivered exactly the kind of earnings call investors needed following the uncertainty of the past week. Management addressed nearly every major concern, raised its long-term growth outlook, and showed that demand for AI infrastructure remains exceptionally strong. Of course, Jensen's departure is still a major uncertainty, circular financing risks haven't completely disappeared, and customers will eventually need to demonstrate sufficient returns on their enormous AI investments. But when you have a company generating $116 billion in quarterly revenue, guiding for $130 billion next quarter, and still expecting 75% revenue growth in fiscal 2028, it's pretty difficult to argue that the underlying business is falling apart. Whether that's enough to reverse the recent stock decline remains to be seen, but I think NVIDIA gave the bulls plenty of ammunition tonight.
While the market has been selling, I've been buying.
You can see exactly what I've been buying, selling and holding, along with my full portfolio for just $1. milkroad.com/pro/?utm_medium…
While the market has been selling, I've been buying.
You can see exactly what I've been buying, selling and holding, along with my full portfolio for just $1. milkroad.com/pro/?utm_medium…
$NVDA: "Our momentum is real. Our demand is accelerating. Our supply is growing. Our customers and our partners are with us. And employees are committed. We've got this."
$NVDA JUST REPORTED EARNINGS, AND THEY ABSOLUTELY CRUSHED EXPECTATIONS.
Revenue came in at $115.8 billion versus expectations of $108.8 billion, a beat of roughly $7 billion. EPS came in at $2.68 versus expectations of $2.47, while data center revenue reached a staggering $107.6 billion, beating estimates of $101.87 billion. Adjusted gross margins also came in at 74.9%.
These are some incredible numbers, especially considering all the uncertainty surrounding NVIDIA over the past few weeks following Jensen Huang's departure, concerns about circular financing, and questions about the sustainability of AI spending. But here's the thing. Beating expectations was only half the battle. The market also wants reassurance that demand remains strong, customers are continuing to spend aggressively, and NVIDIA can maintain its growth trajectory without Jensen at the helm.
For me, the next quarter guidance and management commentary are going to be just as important as these results. NVIDIA needed to deliver a strong quarter, and so far, it looks like they've done exactly that. Now let's see if they can convince Wall Street that the AI trade still has plenty of room to run.
$NVDA: "These partners are instrumental in financing the build-out of AI infrastructure. As we advance this work, we'll remain thoughtful in how we participate and disciplined in how we commit capital. We are proud of the role that we have played so far in supporting the ecosystem. As the market matures and additional sources of capital become available, we are rigorously evaluating future investments, commitments, and guarantees."
$NVDA JUST REPORTED EARNINGS, AND THEY ABSOLUTELY CRUSHED EXPECTATIONS.
Revenue came in at $115.8 billion versus expectations of $108.8 billion, a beat of roughly $7 billion. EPS came in at $2.68 versus expectations of $2.47, while data center revenue reached a staggering $107.6 billion, beating estimates of $101.87 billion. Adjusted gross margins also came in at 74.9%.
These are some incredible numbers, especially considering all the uncertainty surrounding NVIDIA over the past few weeks following Jensen Huang's departure, concerns about circular financing, and questions about the sustainability of AI spending. But here's the thing. Beating expectations was only half the battle. The market also wants reassurance that demand remains strong, customers are continuing to spend aggressively, and NVIDIA can maintain its growth trajectory without Jensen at the helm.
For me, the next quarter guidance and management commentary are going to be just as important as these results. NVIDIA needed to deliver a strong quarter, and so far, it looks like they've done exactly that. Now let's see if they can convince Wall Street that the AI trade still has plenty of room to run.
$NVDA: "NeoClouds remain an increasingly important part of the global AI infrastructure ecosystem, providing additional customers with access to NVIDIA compute, and expanding the reach of our platform. We have supported the development of this ecosystem, and we will continue to be disciplined and selective in our support."
$NVDA JUST REPORTED EARNINGS, AND THEY ABSOLUTELY CRUSHED EXPECTATIONS.
Revenue came in at $115.8 billion versus expectations of $108.8 billion, a beat of roughly $7 billion. EPS came in at $2.68 versus expectations of $2.47, while data center revenue reached a staggering $107.6 billion, beating estimates of $101.87 billion. Adjusted gross margins also came in at 74.9%.
These are some incredible numbers, especially considering all the uncertainty surrounding NVIDIA over the past few weeks following Jensen Huang's departure, concerns about circular financing, and questions about the sustainability of AI spending. But here's the thing. Beating expectations was only half the battle. The market also wants reassurance that demand remains strong, customers are continuing to spend aggressively, and NVIDIA can maintain its growth trajectory without Jensen at the helm.
For me, the next quarter guidance and management commentary are going to be just as important as these results. NVIDIA needed to deliver a strong quarter, and so far, it looks like they've done exactly that. Now let's see if they can convince Wall Street that the AI trade still has plenty of room to run.