RTX 3050 likely to make a comeback because Nvidia knows you'll still buy one
What no one expected was that the RTX 3050 would still have some sort of relevance more than three years later. The latest version of GPU-Z has added support for a new RTX 3050 A variant. This would make it to be the fifth version of the GPU since its debut in 2022, making it one of the most frequently revised GPUs in Nvidia's recent history.
Too many variants?
Nvidia initially launched the RTX 3050 4GB (GA107) and 8GB (GA106) models in 2022, followed by another 8GB variant using the GA107 GPU. The most recent revision arrived last year as a 6GB model, bringing the total to four distinct configurations under the same name.
Recommended Videos
As for the RTX 3050 A, this variant was initially spotted on a PCI ID database under the name GeForce RTX 3050 A Laptop GPU in 2024. During the time, Nvidia had confirmed that the card would use its Ada Lovelace architecture, specifically the AD106 GPU, replacing the older Ampere-based GA107 chip used in earlier models.
RTX 3050 A RTX 3050 (6GB) RTX 3050 (8GB) RTX 3050 (4GB) RTX 3050 (8GB) GPU AD106 GA107 GA107 GA107 GA106 Architecture Ada Lovelace Ampere Ampere Ampere Ampere Process TSMC 4N FinFET Samsung 8N Samsung 8N Samsung 8N Samsung 8N CUDA cores NA 2,304 2,560 2,048 2,560 VRAM NA 6GB GDDR6 8GB GDDR6 4GB GDDR6 8GB GDDR6 Memory interface NA 96-bit 128-bit 128-bit 128-bit Transistor count 22.9 billion 8.7 billion 8.7 billion 8.7 billion 12 billion
With the latest GPU-Z changelog now explicitly mentioning a new RTX 3050 A, there seems to be a possibility that Nvidia might be relaunching the budget GPU. More importantly, it is unclear whether this version is intended for laptops, desktops, or both.
Unlike its predecessors, the RTX 3050 A is expected to feature the AD106 GPU, based on Nvidia's Ada Lovelace architecture. This is the same silicon used in the RTX 4060 series and some RTX 4070 mobile GPUs. Notably the AD106 GPU is physically smaller than the earlier GA107 and GA106 chips, but since it is built using TSMC's 4N FinFET process, it offers higher transistor density and improved efficiency.
Why is Nvidia doing this?
There's no official statement from Nvidia, but the continued development of a new RTX 3050 variant suggests the company sees sustained demand for entry-level graphics cards. For anyone looking for a sub-$200 GPU from Nvidia, your options will be limited to older or cut-down Ampere cards. An 8GB RTX 3050 based on AD106 could plug that hole, offering good enough performance for 1080p gaming while leveraging Ada Lovelace's power efficiency gains.
Critics might scoff at the idea of dragging out a three-year-old GPU name across five hardware variants. I would too. It completely messes with the product differentiation and is going to confuse buyers who assume all RTX 3050s are the same. But Nvidia has long leaned into broad product ranges under single branding umbrellas.
The real story here might be less about nomenclature and more about how Nvidia continues to repurpose its GPU lineup to cover gaps in its stack, especially in the wake of rising GPU costs and growing pressure from integrated graphics solutions like AMD's Ryzen APUs or Apple's M-series chips.
Nvidia recently did the same earlier this year when it silently launched the RTX 5050 which is essentially a slightly repackaged GPU priced at $250. Rather than offering a fresh, purpose-built GPU for entry-level gamers, Nvidia seems content rebadging older silicon and pushing it into a new generation. Despite being based on the latest Blackwell architecture that powers the latest RTX 50 series GPUs, its specs aren't meaningfully improved over what we've already seen from the many variants of the RTX 3050.
In some ways, Nvidia has launched a next-gen budget card without actually moving the needle forward, leaving consumers to pay more for what is functionally a product that could've launched two years ago. While this strategy might help Nvidia pad its lineup and target price-sensitive markets, it does little to inspire confidence that the company is serious about innovation at the low end.
Try Our AI Features
Explore what Daily8 AI can do for you:
Comments
No comments yet...
Related Articles
Yahoo
2 hours ago
- Yahoo
Nvidia Doubles Down With Massive H20 Order
Nvidia (NASDAQ:NVDA) just placed a fresh order for 300,000 H20 AI GPUs from TSMC (TSM) to keep pace with booming China demand. That order adds to an existing 600,000700,000 H20 chip stockpile and pushes China?bound inventory close to 1 million follows April's U.S. lift on H20 exports under the Trump administration, even though formal export licenses from the Commerce Department are still awaiting approval. Warning! GuruFocus has detected 5 Warning Signs with NVDA. The H20 was built specifically for China with pared?down compute versus Nvidia's H100 and Blackwell lines to comply with 2023 restrictions, and now it's become a linchpin in Nvidia's China play. Analysts say the order underscores Nvidia's pricing power and volume leverage in AI's hottest market. With lead times of nine months to restart H20 production, Jensen Huang has doubled down on TSMC capacity to head off domestic Chinese rivals and blunt any hiccups from export rule changes. Nvidia shares popped on the news as investors bet on strong Chinese uptake to drive another leg of growth. This article first appeared on GuruFocus. Error in retrieving data Sign in to access your portfolio Error in retrieving data Error in retrieving data Error in retrieving data Error in retrieving data
WIRED
2 hours ago
- WIRED
The Nvidia RTX 5060 Can't Quite Beat AMD
For the RTX 5060, Nvidia only offered to send out the card if we agreed to stipulations around which games we would benchmark, which features we would enable, and which other GPUs we'd compare it to. That would've looked a lot better for Nvidia than the reality of the situation, putting an emphasis on multiframe-generation features only available on the newest cards, rather than directly comparing traditional rendering benchmarks. All this to say, I personally went out and bought the RTX 5060 for this review, which wasn't as tough as it is with the higher-end offerings. I've expressed in previous reviews that supply would be an issue, and having now lived it first-hand, it wasn't pleasant, but it was possible. Cards are available, but you may have to jump through some hoops to get them into your cart, especially if there are discounts involved. Photograph: Brad Bourque None of this affects the performance or my review, but I do think it's an indication of how confident Nvidia is about what would historically be its most popular offering. The GPU makers may not want to admit it, but the benchmarks don't lie: PC gaming is officially getting more expensive, and $300 just won't buy you the same performance it would have even a few years ago. Performance here is lacking, but it's only going to get worse as time goes on and more developers start taking 16 GB for granted. While most gamers still play at 1080p, that number is always dropping, and I've said multiple times that you should at least target 1440p for a new build. Otherwise, both the RTX 5060 Ti and RX 9060 XT offer much better value, and as I write this there are OC models for the latter at just over $300. The only reasons you should opt for the RTX 5060 instead is if you can find an exceptional deal, you literally only play Roller Coaster Tycoon 2 (2002) , or you happen to have a 550W PSU you can't or won't upgrade.
Forbes
3 hours ago
- Forbes
How Intel Stock Falls To $10
Intel stock (NASDAQ:INTC) has barely moved this year, up just 2%, as the company continues to struggle with shrinking relevance in its core CPU market and underwhelming progress in its foundry ambitions, despite investing over $50 billion in the space. Revenue has collapsed from $79 billion in 2021 to $53 billion in 2024. While the broader PC market is showing signs of stabilization, Intel's top line is projected to decline again this year - suggesting revenue stagnation could become the norm. Meanwhile, stocks of competitors like AMD and Nvidia are soaring, up 43% and 24% respectively in 2025. Could Intel stock plunge to $10 - half its current value? It may seem extreme, but given the steady erosion of its fundamentals, it's no longer unthinkable. Below, we provide a scenario considering three key metrics, namely revenues, net margins, and price-to-earnings multiple. That said, if you want upside with a smoother ride than an individual stock, consider the High Quality portfolio, which has outperformed the S&P, and clocked >91% returns since inception. SANTA CLARA, CA - JULY 15: An Intel sign is displayed in front of the Intel company headquarters ... More July 15, 2008 in Santa Clara, California. Intel has reported a 25 percent increase in its second quarter earnings with net income of $1.6 billion or 28 cents per share compared to $1.28 billion, or 22 cents per share one year ago. (Photo by) Revenues Could Stagnate Intel's sales have declined considerably of late. Intel revenues declined from $79 billion in 2021 to $53 billion in 2024 as Intel's CPU sales declined due to the cooling off of the PC market post-Covid-19, and also due to market share gains by rival AMD. The rise in mobile devices and increasing demand for AI chips - areas where Intel has a limited presence - have also hurt. While the PC market is recovering with sales projected to grow by low single-digits this year, Intel's revenues don't look like they will stabilize just yet, with consensus estimates projecting a 2% dip in sales this year. There remains a possibility that Intel could see its revenues stagnate in the interim due to a host of factors. While Intel struggles, Will AGI Take Nvidia Stock To $300? Foundry business may not be taking off: There's are increasing signs that Intel's foundry bet isn't taking off the way the company had expected. During its most recent earnings call, Intel appeared to play down hopes of winning major external customers for its 18A process - its most advanced manufacturing tech to date - with the leadership noting that 18A will primarily be used for internal products to begin with. Intel also said in its quarterly filing that it could potentially 'pause or discontinue' its next-generation 14A process if it was unable to win a significant customer. However, the market for foundry services is actually booming. Taiwan's TSMC - the world's largest foundry player - sees its AI-related chips revenue doubling in 2025 and rising at a mid-40% levels over the next five years. Intel, on the other hand, is witnessing very little of the action. CPU market share losses: Intel's CPU business could face further pressure, despite new launches, as the generative AI era could open the doors to more competition as PC makers look to incorporate more smarts into their devices. For instance, both chip-designer ARM and mobile chipset specialist Qualcomm are pushing into the PC space and Microsoft's latest Copilot+ PCs use ARM chips that offer AI features and consume less power. On the server front too, there could be challenges as accelerated computing servers used for generative AI applications typically require just one CPU for eight or more GPUs in AI servers. Moreover, GPU makers such as Nvidia are playing a bigger role in overall server system design, looking to replace dedicated CPUs from the likes of Intel with lower-powered ARM chips instead of Intel's. This could impact Intel's bread-and-butter CPU business. Foundry utilization dilemma: Intel also faces a dilemma of sorts. While competitors like AMD and NVIDIA use TSMC's superior, cutting-edge manufacturing processes, Intel must balance product competitiveness with the financial health of its own costly foundries. Intel has already been sending some chip orders TSMC's way in recent years for crucial components of some of its recent processors. While this outsourcing boosts Intel's product performance, it simultaneously starves its internal manufacturing division of crucial orders needed to cover fixed costs. This forces Intel into a difficult choice: risk using its own possibly less advanced fabs and falling behind rivals in the CPU game, or undermine its internal foundry operations by further embracing TSMC. Intel is clearly on the back foot. While the company is keen to build momentum - employee morale can't be high either. Customers and buyers are more likely to want the 'best' and if the word on the street is that Intel isn't the 'future' - it's less likely to be the choice 'now.' Everything becomes just a tad harder. Intel revenues are projected at about $52 billion for this year per consensus estimates and there is a possibility that sales could remain flat in the coming years, due to the aforementioned factors. Margins Fall Further? Intel's adjusted net margins (net income, or profits after expenses and taxes, calculated as a percent of revenues) have been on a declining trajectory - they fell from levels of over 28% in 2021 (and in years before that) to just about 8.5% in 2023 due to sales declines and considerable losses in the foundry business. The metric fell to negative levels in 2024 as Intel posted losses. While the markets are likely betting that Intel's margins could eventually expand to historical levels as it sets its product and manufacturing roadmap in order, there remains a possibility that margins could remain depressed, remaining at about 5% levels in 2025. Costs associated with the foundry ramp-up could hurt Intel's bottom line. Moreover, Intel's move to outsource production of its Arrow Lake chip to TSMC could potentially reduce the utilization of its own manufacturing facilities. Intel has also not exactly been known for production efficiency. For perspective, in 2023, Intel's foundry business reported an operating loss of $7 billion on sales of $18.9 billion. Separately, higher competition in the CPU space - where new entrants such as Qualcomm and ARM - might also force Intel to resort to some amount of discounting. How This Impacts Intel's Stock Now at the current market price of about $20 per share, Intel trades at about 160x estimated 2025 earnings and about 1.7x consensus revenue. Since Intel was unprofitable in 2024, let's ignore the P/E multiple for this year. In 2023, Intel traded at about 19x. So what explains the difference in Intel's P/E multiple using 2023 and 2025 earnings? It's because investors are betting that things will get better going forward. However, if Intel doesn't deliver in the interim, investor sentiment could go further downhill. If we combine the scenario we detailed above - which assumes no meaningful annual revenue growth between 2025 and 2027 with adjusted net margins falling to about 5% - this means that adjusted net income could fall from about $4.4 billion in 2023 ($1.05 per share) to about $2.5 billion in 2027 ($0.58). Bad times make it easier to imagine worse times - and when that happens, things can spiral causing investors to assign an even lower multiple to Intel re-assessing Intel's recovery path. For example, if Intel's investors assign a multiple of 18x to Intel, following its continued underperformance, this would translate into a stock price of just about $10 per share. What about the time horizon for this negative-return scenario? While our example illustrates this for a 2027 timeline, in practice, it won't make much difference whether it takes two years or four. If the competitive threat plays out, with Intel also continuing to struggle with manufacturing, we could see a meaningful correction in the stock. While you would do well to be careful with INTC stock for now, you could explore the Trefis Reinforced Value (RV) Portfolio, which has outperformed its all-cap stocks benchmark (combination of the S&P 500, S&P mid-cap, and Russell 2000 benchmark indices) to produce strong returns for investors. Why is that? The quarterly rebalanced mix of large-, mid- and small-cap RV Portfolio stocks provided a responsive way to make the most of upbeat market conditions while limiting losses when markets head south, as detailed in RV Portfolio performance metrics.
