Deep Dives

NVIDIA’s recent quarterly earnings were nothing short of historic, underscoring the company’s dominant position in the AI-driven semiconductor market. The Q4 FY2025 results (for the quarter ended Jan 26, 2025) showed record revenue and profits, fueled by surging demand for its AI chips​ (nvidianews.nvidia.com). This report provides a deep analysis of NVIDIA’s financial performance and evaluates its long-term investment appeal. We also examine the key tailwinds propelling its growth, the headwinds and risks it faces, and the broader impact of NVIDIA’s results on peer companies like AMD, Intel, Broadcom, Google, and Marvell Technology (MRVL). Fundamental metrics, valuation multiples, and analyst sentiment are incorporated to give a comprehensive investment outlook.

Financial Performance Highlights (Q4 and Full-Year)

NVIDIA delivered blowout earnings in its latest report. Quarterly revenue hit $39.3 billion, up 78% year-over-year​ (nvidianews.nvidia.com), marking an all-time high. Notably, Data Center revenue – largely driven by AI accelerator sales – reached $35.6 billion in Q4, a 93% YoY increase​ (nvidianews.nvidia.com). This reflects unprecedented demand for NVIDIA’s advanced GPUs used in artificial intelligence workloads. By contrast, the Gaming segment contributed $2.5 billion (down 11% YoY as the product cycle matured)​ (nvidianews.nvidia.com), underscoring how NVIDIA’s center of gravity has shifted toward AI and data centers. Other segments like Professional Visualization and Automotive, while much smaller, showed YoY growth of 10% and 103% respectively in Q4​ (nvidianews.nvidia.com) (nvidianews.nvidia.com) – indicating NVIDIA’s diversification into new applications (e.g. AI for design, robotics, autonomous vehicles).

For the full fiscal year 2025, revenue doubled (up 114%) to $130.5 billion, a record high​ (nvidianews.nvidia.com). This explosive growth filtered down to the bottom line: NVIDIA’s GAAP net income jumped 145% to $72.9 billion for the year, with exceptional profitability. The company’s EBIT margins exceed 60%​ (morganstanley.com) thanks to its near-monopoly in high-end AI chips, and Q4 GAAP net margin was on the order of ~56% (net income $22.1 billion on $39.3 billion revenue)​ (nvidianews.nvidia.com) (nvidianews.nvidia.com). Such margins are virtually unheard of in semiconductors and highlight NVIDIA’s strong pricing power and operating leverage.

Earnings per share (EPS) skyrocketed accordingly. Q4 GAAP diluted EPS was up 82% YoY​ (nvidianews.nvidia.com) (adjusted for stock splits), and full-year GAAP EPS more than doubled (+147%)​ (nvidianews.nvidia.com). NVIDIA is now generating cash at an enormous rate, which strengthens its balance sheet and ability to invest in future growth or return capital to shareholders. Overall, the financial performance confirms NVIDIA’s status as an “AI juggernaut”, with growth rates more typical of a startup than a company of its size​ (morganstanley.com). This kind of momentum is a strong positive indicator for long-term investors, provided it can be sustained.

Long-Term Growth Potential and Business Outlook

NVIDIA’s long-term growth story remains compelling. The company has effectively cornered the market for AI computing – holding an estimated 90%–95% market share in data-center GPUs​ (morganstanley.com) – and it continues to innovate aggressively. In the
earnings call, CEO Jensen Huang highlighted “massive-scale” production ramp of Blackwell AI supercomputers (its latest GPU architecture) and noted “demand for Blackwell is amazing”​ (nvidianews.nvidia.com). This suggests that even after the current cycle of AI investments by tech giants, NVIDIA expects further waves of growth as new AI applications emerge. Huang specifically pointed to “reasoning AI” and “agentic AI” (AI agents that can take actions) and physical AI (robotics) as the next frontiers, poised to “revolutionize the largest industries”​ (nvidianews.nvidia.com). These areas could dramatically expand NVIDIA’s addressable market beyond cloud and internet companies, into sectors like manufacturing, healthcare, transportation, and more.

Importantly, NVIDIA is preparing for a future where enterprise and industrial adoption of AI plays a larger role. Huang mentioned he believes future revenue from enterprises will eventually exceed that from cloud service providers (CSPs)​ (247wallst.com). In line with this, NVIDIA has been rolling out enterprise-friendly offerings – from AI software frameworks to on-premise “personal AI supercomputers” (e.g. Project DIGITS for researchers​ (nvidianews.nvidia.com)) – aiming to lower the barrier for wider AI deployment. The company is also diversifying its product lineup: for example, launching new Grace data center CPUs and Grace Hopper superchips, which complement its GPUs and could capture a slice of the CPU market in AI-centric data centers. In gaming, NVIDIA just announced the GeForce RTX 50-series for 2025 powered by the Blackwell architecture​ (nvidianews.nvidia.com), which should reinvigorate its gaming revenue and showcase AI-driven rendering features that blur the line between consumer and AI tech. In automotive, NVIDIA’s design wins with Toyota and Hyundai for AI-powered systems​ (nvidianews.nvidia.com) point to growing traction in self-driving and smart cars, a potentially significant long-term revenue stream.

Put simply, NVIDIA is firing on multiple cylinders: it dominates current AI compute needs and is positioning itself to dominate future ones. The growth runway still looks expansive – from generative AI and cloud services today, to AI-powered businesses and devices across the economy tomorrow. As long as NVIDIA maintains technology leadership and its robust developer ecosystem (CUDA and software libraries), its future growth potential remains strong. This underpins the case for NVIDIA as a long-term investment, as the company could continue compounding earnings at high rates if AI adoption continues its upward trajectory.

Key Tailwinds for the Next Year

Several growth drivers and tailwinds support NVIDIA’s bullish outlook over the next year (and beyond):
 

• Unprecedented AI Demand – The surging interest in generative AI and large-language models has translated into exceptional demand for NVIDIA’s hardware. The latest quarter’s Data Center sales +93% YoY​(nvidianews.nvidia.com)

  is clear evidence of an industry-wide investment cycle in AI infrastructure. This trend is likely to continue into next year as companies race to build out AI capabilities. Notably, cloud giants are dramatically boosting capex: Alphabet (Google) announced it will spend $75 billion on AI infrastructure in 2025 (servers and data centers), far above expectations ​(reuters.com) (reuters.com). Similarly, other tech firms and enterprises are increasing AI budgets. This “AI gold rush” directly benefits NVIDIA, which supplies the majority of the critical hardware.

• Market Leadership & Ecosystem – NVIDIA’s near-monopoly in advanced AI chips gives it a significant edge. With roughly 90%+ market share in AI GPUs​(morganstanley.com), it enjoys economies of scale, the broadest customer base, and a vast software ecosystem that competitors struggle to match. Its CUDA software platform and developer community create a high switching cost – a powerful tailwind as new AI adopters almost reflexively choose NVIDIA for its proven stack. This leadership position also allows NVIDIA to capture outsized profits (as seen in its margins) and to reinvest heavily in R&D to sustain its technology edge.

• Product Innovation & Roadmap – NVIDIA’s product pipeline is a major tailwind. The company is not standing still: it’s rolling out the new Blackwell GPU architecture (with an aggressive ramp in 2025) and has next-generation chips (code-named Rubin for 2026) in the works ​(morganstanley.com). Each new generation improves performance and efficiency, spurring upgrade cycles. Additionally, NVIDIA’s move into CPU and DPU (data processing unit) products means it can offer more complete solutions and tap into new markets (e.g. replacing or supplementing traditional server CPUs in certain workloads). The Grace CPU and BlueField DPU lines could become meaningful contributors, riding on the AI trend of specialized computing. In gaming and consumer graphics, the RTX 50-series launch in 2025 (with AI-enhanced features like DLSS 4) should reinvigorate that segment​ (nvidianews.nvidia.com). Continuous innovation keeps NVIDIA a step ahead and creates multiple engines of growth.

• Broader AI Adoption – Beyond the current hyperscaler-driven demand, broader adoption of AI across enterprises is a tailwind that is just beginning. Many industries are still in early stages of implementing AI (often experimenting with pilot projects). As solutions mature (e.g. AI customer service, AI-driven analytics, automation through robotics), we can expect a diffusion of AI workloads to more traditional enterprises. NVIDIA is actively courting this next wave – for example, by offering optimized AI software frameworks for healthcare, finance, and manufacturing, and partnering with IT services firms to reach enterprise clients. Huang’s commentary that enterprise use will eventually dominate is telling ​(247wallst.com). In essence, NVIDIA could see a second phase of growth as “non-tech” companies start deploying AI at scale, on top of the ongoing hyperscaler investments. This provides a sustained tailwind beyond the next year, but we may see its early signs in the coming year (e.g. rising sales to commercial/OEM customers, not just cloud players).

• AI as a Competitive Imperative – Another soft tailwind: at this point, AI capability is viewed as critical to competitiveness in tech. This means companies like Google, Microsoft, Amazon, Meta (Facebook) must continue spending on AI (and thus on NVIDIA hardware) to keep up with each other. Even emerging competition (like China’s DeepSeek) in AI models is indirectly a tailwind for NVIDIA – Oppenheimer analysts noted that the rise of DeepSeek will push U.S. players to “step up their efforts in the AI race instead of pulling back”, ultimately driving more demand for Nvidia’s chips ​(investopedia.com). In other words, fear of falling behind ensures that NVIDIA’s biggest customers will likely sustain high levels of investment in AI projects, supporting NVIDIA’s sales.

• Ecosystem Partnerships – NVIDIA has cultivated strong partnerships that amplify its reach. It collaborates with cloud providers (offering NVIDIA DGX Cloud instances on Azure, GCP, Oracle Cloud, etc.), with software companies (optimizing AI frameworks like PyTorch and TensorFlow for its GPUs), and with integrators (like Dell, HPE for enterprise servers). These partnerships effectively make NVIDIA’s technology ubiquitous and easily accessible, acting as a tailwind for continued adoption. Even in adjacent areas like networking, NVIDIA has partnered or chosen to focus where it’s strongest – for example, NVIDIA emphasized that its networking strategy is centered on Ethernet solutions, an area where Broadcom is a key player​ (247wallst.com). This implies NVIDIA will leverage partners (rather than fight on all fronts), which in turn fosters an ecosystem conducive to its growth.

In summary, NVIDIA’s growth drivers are robust: a dominant market position in a rapidly expanding industry, continuous innovation, and a rising tide of AI adoption across sectors. These tailwinds position the company favorably for the next year and well into the future.

Key Headwinds and Risks

Despite its strengths, NVIDIA faces several headwinds and risks that investors should monitor, especially over the next year:
 

• Rising Competition – NVIDIA’s success has inevitably drawn competition, and this is a key long-term risk. Rival chipmakers and big tech firms are racing to develop alternatives to NVIDIA’s GPUs. AMD has positioned itself as the #2 player in AI accelerators, and while far behind, it expects ~$5 billion in AI chip sales in 2024 (vs. NVIDIA’s ~$30 billion per quarter in data center now)​ (nasdaq.com). AMD’s upcoming MI300 series GPUs are aimed at high-end AI tasks and have won some early customers (Oracle Cloud, etc.), which could gradually chip away at NVIDIA’s share. Intel is likewise investing in its Habana Gaudi AI accelerators and has a roadmap (Gaudi3 and beyond) to offer cost-effective AI chips​(reddit.com) . Additionally, each of NVIDIA’s top cloud customers is pursuing in-house AI chips: e.g. Google’s TPU, Amazon’s Trainium/Inferentia, Meta’s rumored AI accelerator, and Microsoft’s “Athena” project. According to Morgan Stanley, NVIDIA’s four largest customers (Google, Amazon, Microsoft, Meta) make up ~40% of its revenue; if any “succeed in their efforts to design a better-priced alternative to the H100 or its next-gen chips,” it would pose a serious threat​ (morganstanley.com). While none of these efforts have dethroned NVIDIA yet, the competitive pressure is increasing. Over the next year, we may see heightened marketing of rival chips (AMD’s MI300X, Google’s TPUv5 etc.) and potential share gains in niche cases. NVIDIA’s dominance is likely to continue near-term, but not unchallenged – any slip in execution or a competitor breakthrough (on performance or cost) could start eroding its moat.

• Regulatory and Geopolitical Risks – NVIDIA is situated in the crossfire of U.S.-China tech tensions. The U.S. government has already imposed export restrictions on NVIDIA’s flagship chips to China, prompting NVIDIA to offer modified versions (A800/H800) for that market. However, there are reports of further tightening ahead – in fact, just before earnings, news emerged that Washington was planning additional curbs on AI chip exports to China​(reuters.com)

  . China has been a significant market for NVIDIA’s data center GPUs, and stricter export bans could limit NVIDIA’s sales or push Chinese customers toward domestic alternatives. This is a key headwind for the coming year: if geopolitical relations worsen, NVIDIA could lose access to high-growth markets or face licensing hurdles. Additionally, regulatory scrutiny in other areas (antitrust concerns, etc.) could arise given NVIDIA’s dominant position. Although NVIDIA’s attempted ARM acquisition was blocked in 2022 on antitrust grounds, the company could face future challenges if regulators perceive its ecosystem control (hardware + software) as anti-competitive. For now, the more immediate risk is export controls and international trade barriers, which remain an overhang on NVIDIA’s otherwise stellar growth prospects.

• Potential AI Demand Volatility – The current spike in AI hardware demand might see periodic volatility. Investors are wary that after a massive build-out, cloud providers might moderate their spending (“digestion period”). For instance, there was a report of Microsoft canceling some data center capacity leases, which raised questions about cloud capex cycles​ (reuters.com). If one or more major customers temporarily pause orders to absorb existing capacity, NVIDIA’s revenue growth could slow more than expected in a given quarter. Moreover, macro-economic conditions could influence AI investment – while AI is a priority, a recession or tighter corporate budgets might delay some projects. There’s also the risk of hype vs. reality: if the returns on AI investments (in terms of revenue or efficiency gains for cloud companies) don’t materialize as fast as hoped, those companies might face investor pressure to rein in spending. Reuters noted “investor skepticism has grown over the billions channeled into AI infrastructure due to slow payoffs” ​(reuters.com). Any signs of AI over-investment or under-utilization of all those GPUs could dampen the breakneck demand and weigh on NVIDIA’s growth trajectory – at least until new use cases catch up. Thus, NVIDIA’s stock could be sensitive to even short-term demand lulls or cautious commentary from big customers.

• Customer Concentration – Relatedly, NVIDIA’s heavy reliance on a few large buyers is a double-edged sword. As mentioned, ~40% of revenue comes from four tech giants​ (morganstanley.com). These customers have enormous bargaining power and are also those pursuing in-house chips. Even if their own chip projects fall short, they can pressure NVIDIA on pricing or shift mix between internal and external solutions. Over the next year, if one of these major customers decides to significantly alter its procurement (for example, Google allocating more budget to TPUs over GPUs, or Amazon prioritizing its Trainium), NVIDIA could see a noticeable impact. So far, demand from each has been so strong that NVIDIA has been selling everything it can produce, but it’s a risk to watch. On the flip side, NVIDIA is working to broaden its customer base (enterprise, international, other industries) to reduce this concentration risk over time.

• Valuation and Expectation Risk – NVIDIA’s stock valuation embeds very high expectations. After its meteoric rise, the stock still trades at premium multiples – recently about 20× sales and over 40× earnings (on a forward basis) during 2024’s rally​(morganstanley.com). Even with the latest jump in earnings (which has somewhat normalized the P/E), NVIDIA is far more expensive than the average semiconductor or S&P 500 company. Such a valuation can be a headwind in itself: it leaves little margin for error. If NVIDIA’s growth even slightly disappoints relative to forecasts, the stock could see outsized downside. The company’s own guidance for next quarter (Q1 FY2026) is strong – revenue ~$43 billion (+9% QoQ)​ (nvidianews.nvidia.com) – but any shortfall or signs of growth deceleration beyond that could trigger profit-taking. Additionally, broader market rotation away from high-multiple stocks (for instance, if interest rates rise) can pressure NVIDIA’s share price irrespective of its fundamental performance. In short, investors are paying for perfection, and that is a risk if anything less than perfection occurs in coming quarters.

• Innovation/Execution Risk – While NVIDIA has an excellent track record, it must execute flawlessly to maintain its lead. Manufacturing cutting-edge chips is challenging – NVIDIA depends on TSMC’s foundries to deliver new GPUs on time and with good yields. Any delays in the ramp of next-gen products (e.g., if “Blackwell Ultra” chips were to slip in schedule or face yield issues) could give competitors an opening or leave demand unmet. There were rumors about potential delays, but management indicated no major changes to the roadmap ​(247wallst.com). Still, it’s a watch item. Similarly, as NVIDIA broadens into more products (CPU, networking, etc.), it ventures slightly outside its core GPU forte, which could present execution challenges or require different sales strategies. Lastly, supply chain constraints – ironically a risk on the upside – must be managed; NVIDIA has navigated shortages well so far by investing in supply, but if AI demand keeps climbing, any bottleneck (in components like high-end HBM memory, substrate capacity, etc.) could limit its ability to fully capitalize on the tailwind.

In summary, NVIDIA’s risks revolve around competition, geopolitical issues, and the sustainability of its growth surge. None of these appear to derail the company’s trajectory in the immediate term, but they inject uncertainty into the long-term outlook. Prudent investors will keep an eye on how these headwinds develop even as they consider NVIDIA’s strong current performance.

Impact on AI-Related Peer Stocks

NVIDIA’s earnings and market position have significant ripple effects across the semiconductor and tech landscape, especially for other AI-focused stocks like AMD, Intel, Broadcom, and Alphabet (Google). Here’s how NVIDIA’s dominance and latest results influence each:

Advanced Micro Devices (AMD)

AMD is NVIDIA’s most direct competitor in GPUs and AI accelerators. NVIDIA’s continued blowout performance underscores just how far ahead it is. For context, NVIDIA’s data center revenue is now about $30+ billion per quarter while AMD has guided for roughly $5 billion in AI chip revenue for the entire 2024 year​ ( nasdaq.com). This gulf illustrates NVIDIA’s lead in both market adoption and product capability. That said, AMD’s AI ambitions are a factor to watch. The company’s high-end MI250 and MI300 series accelerators have started gaining traction – for example, AMD reported that Oracle Cloud and other providers will deploy MI300X GPUs for AI workloads​ ( nasdaq.com). As NVIDIA’s H100 GPUs remain supply-constrained and expensive, some cost-sensitive customers may experiment with AMD’s solutions.

Intel (INTC)

For Intel, the implications of NVIDIA’s ascendance in AI are somewhat twofold – it’s both an indirect challenge and a potential ancillary beneficiary. Intel’s core business is CPUs, not GPUs, and traditionally CPUs handled most computing tasks in data centers. The AI era, however, has shifted a lot of workloads (especially training of AI models and increasingly inference) to GPUs and specialized accelerators – an area dominated by NVIDIA. This shift means cloud providers are allocating more of their capex to GPU-rich servers (and perhaps somewhat less to general-purpose CPU servers). Intel, as the leading CPU vendor, could see some pressure on its data center CPU demand if customers optimize their budgets toward NVIDIA’s hardware. In other words, NVIDIA’s growth is emblematic of a computing paradigm shift that Intel must navigate. We’ve already seen Intel’s data center group facing headwinds (though competition from AMD’s EPYC CPUs is also a factor).

In response, Intel is trying to establish itself in AI accelerators. It acquired Habana Labs in 2019 and released Gaudi2 AI chips, with Gaudi3 on the roadmap. Intel touts these as a more cost-efficient alternative for certain AI workloads. Some cloud players (notably AWS) have offered Gaudi-based instances as an option for AI training, citing cost per training as an advantage. There are reports and benchmarks (including one by Databricks researchers) suggesting Intel’s Gaudi can achieve lower cost than NVIDIA for certain tasks, albeit at lower raw performance​ ( reddit.com ) ( reddit.com). Intel’s strategy, as reported, is not to go directly at NVIDIA’s highest-performance crown, but to target the value segment of the AI market – offering say 80% of the performance at a fraction of the cost​ ( crn.com). If this strategy gains traction, Intel could secure a niche in AI acceleration (especially among price-sensitive customers). However, so far Intel’s AI revenue is minuscule compared to NVIDIA.

NVIDIA’s latest earnings – and particularly its guidance of continued growth – suggest that Intel is still far from making a dent. NVIDIA’s H100 remains the go-to solution for anyone chasing top-notch AI performance. Intel’s Gaudi3, expected in 2024/2025, will be coming into a market where NVIDIA is already moving to its next-gen (Blackwell) – so Intel will perpetually be chasing a moving target. Moreover, Intel lacks the software ecosystem that makes NVIDIA’s platform so convenient. Convincing AI developers to switch to a new architecture is non-trivial; this is a headwind Intel shares with AMD and others.

Broadcom (AVGO)

Broadcom is a key player in the semiconductor industry that intersects with NVIDIA’s world, but in a complementary way. Broadcom’s business includes networking chips (Ethernet switches, network interface controllers), custom ASICs, and other connectivity solutions – all critical components in modern data centers, especially ones built for AI.

NVIDIA’s massive growth in AI is actually a tailwind for Broadcom in many respects. When customers build AI clusters with hundreds or thousands of NVIDIA GPUs, they also need high-speed networking to connect those GPUs (for distributed training of AI models). NVIDIA itself recognized this need and a few years ago acquired Mellanox, giving it a strong position in Infiniband and high-performance Ethernet for AI networks. However, the market is so large that Broadcom remains a major supplier of data center networking gear, particularly in Ethernet switches/routers that many cloud providers use. Jensen Huang’s commentary on networking in the latest call is telling: he said NVIDIA’s networking strength is centered in Ethernet, and even gave a nod that this is “positive for Broadcom, which is the main Ethernet competition for NVIDIA.”​ ( 247wallst.com). In other words, NVIDIA acknowledges Broadcom as a leading competitor in networking, and by focusing on Ethernet (versus proprietary Infiniband), NVIDIA validates Broadcom’s domain. After NVIDIA’s earnings announcement, Broadcom’s stock actually rose, as the market likely inferred that strong AI server sales will translate to strong demand for Broadcom’s networking chips as well​ ( 247wallst.com).

Another facet: Broadcom has been designing custom AI chips (ASICs) for hyperscalers, much like Marvell. It has notable wins, such as helping develop Google’s Tensor processing chips and reportedly working on ASICs for OpenAI and others​ ( marketbeat.com). NVIDIA’s dominant off-the-shelf solutions don’t fulfill every need; some big players still commission custom silicon for specialized tasks, and Broadcom is often a partner for that. For example, if a company like Google wants a custom chip to handle a specific AI workload (perhaps complementing GPUs), Broadcom might get that contract. Broadcom and NVIDIA aren’t head-to-head competitors in this – rather, Broadcom is enabling those who choose not to exclusively use NVIDIA’s solutions. In the long run, if more companies choose custom ASICs over GPUs, Broadcom could benefit at NVIDIA’s expense. But in the next year, it appears the demand is so high that both NVIDIA’s GPUs and Broadcom’s ASICs will be in heavy use. Broadcom’s CEO Hock Tan has highlighted AI as a growth driver for their networking and custom silicon business. In its recent results, Broadcom cited strong demand for AI networking chips (Tomahawk switch chips) and noted a large portion of its semiconductor backlog is related to AI investments. NVIDIA’s results reinforce that trend. ( reuters.com), a good chunk of that will go into networking gear where Broadcom is a supplier. One slight caveat: Broadcom is also trying to sell its own AI accelerator chips (it has an AI chip division after acquiring Broadcom Inc. as part of Avago, though not as visible as NVIDIA/AMD). Jensen’s comment “just because you have a design win doesn’t mean it’ll get deployed” was seen as a subtle jab at custom ASICs, possibly including Broadcom’s, which won design slots at companies like Google, Meta, Apple, etc. but might not see full-scale deployment​ ( 247wallst.com). This highlights a risk that some custom-chip efforts don’t translate into huge orders. Even so, Broadcom appears well positioned. For investors, Broadcom is often viewed as a picks-and-shovels play on AI – less flashy than NVIDIA, but providing the necessary plumbing (networking ASICs, switches, connectivity) that scales with AI data centers. NVIDIA’s blowout quarter indicates those data centers are being built at breakneck pace, which means good news for Broadcom’s core businesses. Indeed, analysts have flagged Broadcom as one of the “AI infrastructure” winners alongside NVIDIA. The close relationship: strong NVIDIA results will often boost Broadcom’s stock and outlook, as happened this quarter.

Alphabet/Google (GOOGL)

Alphabet (Google) stands out in this list as it’s not primarily a chipmaker (though it designs some chips), but rather a major consumer of AI technology and a competitor on the AI software front. Google’s connection to NVIDIA is multi-faceted: it is one of NVIDIA’s largest customers, a potential competitor in hardware, and a collaborator in many areas.

First, consider Google as a customer/partner. Google Cloud offers NVIDIA’s A100 and H100 GPUs to its customers for AI workloads, and Google’s internal teams have historically used NVIDIA GPUs for various research (e.g. DeepMind used NVIDIA GPUs). NVIDIA’s latest earnings and guidance imply that Google likely continues to purchase large volumes of NVIDIA GPUs to expand its cloud offerings. In Google’s own recent earnings (Q4 2024), the company announced an eye-opening $75 billion capex plan for 2025 focused on AI​ ( reuters.com), This is a hugely positive indicator for NVIDIA: even if Google allocates a portion of that to its custom TPUs, the sheer size suggests significant spend on third-party hardware as well, including NVIDIA. Google’s CFO noted that the majority of that capex will go to servers and data centers for AI, and even mentioned that in Q4 they faced capacity constraints on cloud AI offerings (i.e. demand exceeded their current hardware availability)​ ( reuters.com). This essentially means Google needs more GPUs/TPUs fast, which is why they’re investing so heavily. For NVIDIA, Google’s aggressive investment is a tailwind: as long as NVIDIA can supply, Google will likely buy many of its top-tier GPUs to satisfy cloud customer needs and to power its own products (like Bard, etc., which run on either TPUs or GPUs or both).

Now, as a competitor in hardware: Google has for years developed the Tensor Processing Unit (TPU), an AI accelerator tailored for its workloads (both training and inference). TPU is currently at v4 (deployed) and v5 (in development). Google’s strategy has been to use TPUs internally (for things like Search and Google Ads, which are increasingly AI-driven) and to offer TPU access on Google Cloud for customers who want an alternative to GPUs. TPU has been moderately successful – Google claims performance advantages in some scenarios – but it hasn’t dethroned GPUs broadly. One reason is that TPUs are mostly available only via Google Cloud, whereas NVIDIA’s GPUs are available everywhere (AWS, Azure, on-premises, etc.), giving developers flexibility. NVIDIA’s latest generation H100 is widely considered the most versatile and powerful general-purpose AI chip on the market, and many papers and companies optimize for it first. So, while Google’s TPUs compete, they co-exist with GPUs. In fact, Google itself likely uses a mix: for instance, some of Google’s generative AI (like image generation in Slides) runs on NVIDIA GPUs as per Google’s IO announcements, whereas large language model training might run on TPUs. The impact of NVIDIA’s dominance on Google is that Google must continuously justify its investment in TPUs. So far, Google seems committed – in the wake of NVIDIA’s surge, Google is actually increasing its AI chip efforts (they recently hired more chip engineers and are rumored to be working on TPUv5). Google’s CEO Sundar Pichai on the Q4 call defended their high capital spending partly by pointing out efficiency improvements and competition like DeepSeek​ ( reuters.com), implying Google needs to keep pushing its tech (like Gemini AI models and TPU hardware) to stay competitive. In the near term, however, Google’s priority is having enough AI capacity, period. So it’s likely buying from NVIDIA even as it refines TPUs – essentially hedging its bets.

From a stock perspective, how do NVIDIA’s results affect Alphabet? One way is via sentiment on AI leaders. Google, as a leader in AI applications (search, cloud AI, etc.), is seen as a key player in the AI economy but also one that will bear high costs for AI build-out. Indeed, after Google revealed the $75B capex plan, its stock fell ~9% because investors worried about the hit to near-term profitability​ ( reuters.com), but in theory these investments (in part going to NVIDIA hardware) should enable future growth in Google’s AI-driven revenue streams. In any case, Alphabet is a unique peer in that NVIDIA’s success is largely a supplier-customer relationship story, with a side of cooperative competition in chips.

Comparison with Marvell Technology (MRVL) – AI and Semiconductor Market Parallels

Marvell Technology (MRVL), like NVIDIA, has emerged as a significant beneficiary of the AI boom, albeit in a different arena of the semiconductor market. It’s worth cross-referencing NVIDIA’s situation with Marvell’s, as both are frequently cited as top picks for AI exposure and their trajectories shed light on trends in the AI and networking space.

Marvell’s recent performance mirrors the AI-driven surge we see in NVIDIA’s results – of course on a much smaller scale. In its latest reported quarter (Marvell’s Q3 FY2025, which corresponds to the quarter ended Oct 2024), Marvell’s data center revenue jumped 98% YoY and 25% sequentially​ ( in.marketscreener.com), reaching about $1.1 billion (which was ~73% of Marvell’s total revenue)​ ( marketbeat.com), This is evidence that AI infrastructure spending is lifting Marvell’s business significantly. The company noted that strong demand for its custom AI silicon and networking products drove this growth, prompting Marvell’s CEO Matt Murphy to declare they are entering a “new era of growth” fueled by AI​ ( marketbeat.com), Marvell even forecasted continued momentum, guiding for >20% QoQ data center growth in the subsequent quarter​ ( in.marketscreener.com), These growth rates, while remarkable for Marvell, still lag NVIDIA’s meteoric 78% YoY overall growth​ ( nvidianews.nvidia.com) – but importantly, they show Marvell is riding the same megatrend.

The key difference lies in their roles. NVIDIA sells general-purpose, programmable GPUs that serve as the brains of AI systems across many customers. Marvell, on the other hand, specializes in application-specific integrated circuits (ASICs) and networking/connectivity chips. Marvell’s strategy has been to engage deeply with hyperscalers to design custom chips for specific AI or cloud workloads. For instance, Marvell has developed or is developing unique silicon solutions for Microsoft (Project Maia for Azure), Google (Project Axion CPU for Google Cloud), and Amazon (it has been linked to AWS’s Trainium and Inferentia chips)​ (marketbeat.com) (marketbeat.com).

An example scenario: A cloud provider might use thousands of NVIDIA GPUs for training AI models but could use a Marvell-designed ASIC to accelerate a specific part of the workload (say, encryption/decryption of data streams, or a pre-processing step for AI data) more efficiently than a GPU would. Or they might deploy Marvell’s networking ASICs/DPUs to connect and manage those GPU clusters (Marvell sells Ethernet switch chips, Smart NICs, etc., similar to Broadcom and NVIDIA’s own Mellanox division). Thus, Marvell often works alongside NVIDIA’s products in AI data centers, rather than directly against them.

However, there is a point of competitive tension: those custom ASICs do, in a sense, attempt to optimize away the need for some additional GPUs. Every task offloaded to a Marvell (or Broadcom) ASIC could mean a few less GPU cores needed. If hyperscalers like Amazon and Google heavily invest in these alternatives (and Marvell has proven capable of delivering them), over time it could modestly reduce the total addressable market for generic AI accelerators like NVIDIA’s. But given the enormity of current demand, that’s not a near-term worry for NVIDIA. So far, the pie is expanding so fast that both NVIDIA and Marvell are growing strongly without zero-sum issues. In fact, Barclays analysts projected both Marvell and Nvidia to be leaders among AI semiconductor plays in 2025, reflecting how Marvell is viewed as “NVIDIA-adjacent” rather than a direct rival​ (barrons.com).

From an investment standpoint, comparing fundamentals: NVIDIA’s revenue ($130B annually) dwarfs Marvell’s ($6B annually), and NVIDIA’s profitability is higher (NVIDIA’s gross margins ~70+%, Marvell’s gross margin ~60% non‑GAAP​ (marketbeat.com)). NVIDIA is also more expensive in valuation (price‑to‑sales, P/E) given its dominant position, whereas Marvell, while not cheap, is valued on the expectation of fast growth from a smaller base. Notably, Marvell is reorienting its business mix heavily towards data center and AI – in Marvell’s latest quarter, 73% of revenue was data center (up from 40% a year prior)​ (marketbeat.com), showing a dramatic shift similar to how NVIDIA’s revenue mix shifted (data center became ~90% of NVIDIA’s revenue in Q4). Both companies have effectively hitched their fortunes to AI, but with different approaches.

Marvell’s headwinds and tailwinds are also somewhat analogous: it benefits from AI capex cycles (like NVIDIA does), and faces competition from Broadcom in custom silicon (just as NVIDIA faces competition in GPUs). One direct parallel is that Marvell and Broadcom often vie for the same ASIC contracts (marketbeat.com) – for example, Marvell won the Microsoft Azure custom chip, Broadcom won Google’s TPU tie‑up; both are likely bidding for other hyperscaler projects. So Marvell’s competitive landscape is different, and in that realm NVIDIA isn’t the opponent, Broadcom is. In fact, Marvell’s success with custom solutions could be indirectly favorable to NVIDIA: if custom chips handle ancillary tasks, customers can free up or better utilize NVIDIA GPUs for the core training tasks.

 

In summary, NVIDIA and Marvell are both key players in the AI semiconductor wave, but they occupy different niches. NVIDIA is the primary supplier of universal AI processors, while Marvell is becoming the go‑to for bespoke silicon and fast interconnects. The latest earnings from each confirm the AI boom is lifting all boats: NVIDIA with staggering numbers, and Marvell with record‑high growth (Marvell’s Q3 overall revenue even beat guidance by $66M thanks to AI demand)​ (in.marketscreener.com). For investors, Marvell can be seen as a complementary investment to NVIDIA – it provides exposure to AI data center growth, particularly networking and custom silicon, which are somewhat de‑coupled from whether NVIDIA or someone else wins the GPU battle. Both companies stand to benefit as AI adoption grows. Cross‑referencing them highlights that the AI revolution in tech is multi‑faceted: it’s driving needs for not just compute (NVIDIA’s forte) but also connectivity and customization (Marvell’s forte). As long as AI data center build‑outs continue at this pace, it is likely both NVIDIA and Marvell will continue to post strong growth, each in their respective domains​ (marketbeat.com) ​ (nvidianews.nvidia.com).

Valuation and Fundamental Metrics

When considering NVIDIA as a long‑term investment, one must weigh its fundamental metrics and valuation in light of its growth. NVIDIA’s financial results have scaled so rapidly that traditional multiples are in flux. On a trailing basis (after this latest earnings), NVIDIA’s valuation multiples remain elevated but are more grounded than a year ago due to the earnings catch‑up.
 

• Revenue and Growth: NVIDIA is now on a ~$160 billion annual revenue run‑rate (based on the Q4 rate), after doubling its revenue this past year​ (nvidianews.nvidia.com). Such size with ~100% growth is unprecedented in the chip industry. Forward‑looking, even if growth moderates, analysts still expect robust expansion (e.g. early consensus for FY2026 revenue implies ~30–40% growth on top of FY2025’s doubling). This growth profile justifies a higher multiple relative to slower‑growing peers.

• Profitability: NVIDIA’s net income and free cash flow generation are exceptional. With net margins around 50+% and operating margin ~60%​ (morganstanley.com), its earnings base is very strong quality. Over the last 12 months it earned roughly $73 billion GAAP net​ (nvidianews.nvidia.com). Free cash flow (FCF) is similarly high, since NVIDIA’s fabless model means capital expenditures are relatively low (it doesn’t own fabs like TSMC; capex mainly for labs, equipment, some infrastructure). This means a large portion of earnings converts to FCF, giving NVIDIA flexibility for buybacks or dividends (it pays a token dividend of $0.01, but authorized a sizable buyback program). High FCF can warrant a premium valuation – especially compared to peers that might have capital‑intensive models.

• P/E and P/S: Using rough figures, at the time of this analysis NVIDIA’s market capitalization is in the trillions (it has become one of the top 5 most valuable U.S. companies). Its price‑to‑earnings (P/E) ratio on a forward basis is around the high‑30s to 40×, and its price‑to‑sales (P/S) ratio is around ~9–10× (down from >20× a year ago)​ (morganstanley.com). These are much higher than the semiconductor industry average (for instance, Intel trades at ~3× sales, ~15× earnings; AMD ~8× sales, ~30× earnings; Broadcom ~10× sales, ~20× earnings). So NVIDIA carries a valuation premium even relative to other growth semis. The market has essentially been pricing NVIDIA more like a software or high‑growth internet company than a hardware company, due to its explosive growth and margins.

• PEG ratio: If one considers the P/E‑to‑growth (PEG) ratio, NVIDIA’s might not be as extreme as the raw P/E suggests. With earnings growing triple‑digits this year and potentially another ~50% next year (hypothetically), a 40× P/E might be closer to or below 1× PEG – implying the valuation is reasonable for its growth. However, the sustainability of that growth is the question; clearly, growth will taper to more normal levels over a few years. The bullish view is that NVIDIA’s valuation is supported by its unique combination of growth + profitability, while the bearish view is that a lot of future success is already priced in, making the stock vulnerable.

• Balance Sheet: NVIDIA’s balance sheet is strong – it has a substantial cash hoard (tens of billions) and relatively modest debt. This financial strength reduces risk; NVIDIA can easily fund R&D (which runs ~$7–8B annually) and strategic investments without needing external financing. It can also opportunistically make acquisitions or buy back shares. For example, in 2023 NVIDIA announced a $25 billion share repurchase authorization, signaling confidence in its future and willingness to return capital. A solid balance sheet doesn’t directly determine valuation, but it underpins the company’s ability to weather downturns and seize opportunities (a plus for long‑term investors).

• Comparative Valuation: Compared to peers: AMD, which also has growth potential, trades cheaper but also has far less earnings (NVIDIA’s quarterly profit now exceeds AMD’s annual revenue). Intel is cheap but is in turnaround mode. The closest comparables on valuation might be “AI pure-plays” or high‑flying tech names. Some might argue NVIDIA is in a league of its own, so comps are tricky. One could compare it to historical cases: e.g., Cisco in 1999-2000 when it was at the center of a tech revolution (networking boom) – Cisco traded around ~30× earnings at its peak. NVIDIA at 40–50× isn’t far off, though interest rate environments differ. Morgan Stanley highlighted that NVIDIA’s market cap increase has been so large that it “single‑handedly produced the equivalent of 80% of the market cap rise of the dot‑com bubble”​ (morganstanley.com), emphasizing how much optimism is baked in. They also noted the stock was more than 20× sales and 40× earnings mid‑2024​ (morganstanley.com), essentially cautioning that NVIDIA needs to achieve massive commercial adoption of AI and maintain dominance for the valuation to be justified.

• Investor sentiment: Currently, sentiment is very bullish (as covered in the next section). This often correlates with valuation – when nearly everyone is positive, usually the stock reflects that optimism in its price. From a fundamental investor’s lens, one might ask: has NVIDIA’s stock become “priced to perfection”? Possibly, yes – any slip‑up could lead to a de‑rating. On the other hand, the addressable market for AI is so vast that if NVIDIA continues executing, its earnings could grow into the valuation quickly (which arguably happened this year to some extent).

NVIDIA’s fundamentals are stellar, but its valuation is elevated. Long‑term investors need to be comfortable with the idea that they are paying a premium for arguably the “best in class” AI play. The premium is supported by NVIDIA’s growth, margins, and market leadership – a combination few companies share. Valuation risk is the primary fundamental concern, given potential volatility. Yet, if one believes AI will continue to transform industries and NVIDIA will capture a significant portion of that value, then the stock’s current multiples could be justified or even modest in hindsight. It’s a classic high‑risk‑high‑reward profile: exceptional fundamentals against a rich valuation backdrop. Prudent investing might involve balancing NVIDIA with other names (like some of the peers discussed) to mitigate single‑company risk at such high multiples.

Analyst Sentiment and Recent Developments

Wall Street analyst sentiment on NVIDIA remains overwhelmingly positive, reflecting the company’s strong execution and AI leadership. Virtually all major analysts reiterated bullish stances after the latest earnings, with many highlighting NVIDIA as a top pick in the semiconductor and AI space. According to a Visible Alpha survey, 17 out of 18 analysts covering NVIDIA rate it a “Buy” (or equivalent), with only one hold and zero sell ratings​ (investopedia.com). This is an unusually high consensus bullishness. The consensus 12‑month price target heading into the earnings was around $175 (post‑split), which implied roughly 30%+ upside from the pre‑earnings share price​ (investopedia.com). (This would equate to a target around $700 in pre‑split terms, indicating the Street sees significant further appreciation despite the stock’s huge run‑up.)

In the days around the earnings report, we saw numerous price target revisions: for instance, Wedbush and Oppenheimer reiterated ~$175 targets, citing booming AI chip demand​ (investopedia.com). UBS maintained a bullish stance with a $185 target, noting that investor expectations had risen but supply improvements could allow NVIDIA to beat even lofty estimates (they pointed out NVIDIA was able to ship more of its new Blackwell chips than initially thought, upping their contribution to revenue)​ (investopedia.com). There are even some outlier uber‑bull cases – for example, one market pundit (in a Barron’s piece during GTC conference) talked of targets north of $1,000 (pre‑split)​ (barrons.com), envisioning NVIDIA as central to a coming “AI era” that could justify a multi‑trillion market cap. While those are exceptions, they underscore the bullish narrative around the stock.

A few analysts have expressed tempered views, though not outright bearish. The Wolfe Research analyst who downgraded AMD (mentioned earlier) indirectly was cautious on the AI chip frenzy in general – but notably did not downgrade NVIDIA, perhaps recognizing NVIDIA’s superior positioning. Some analysts and portfolio managers worry about the sustainability of demand (the DeepSeek incident and Microsoft’s datacenter pause mentioned in Reuters had given a bit of pause​ (reuters.com). Yet, NVIDIA’s results largely assuaged those worries by demonstrating still‑climbing demand. The stock’s reaction to Q4 earnings was mildly positive – shares were up ~1.7% after hours during the call​ (247wallst.com), likely because much of the good news was anticipated and priced in, but there were no negative surprises (“no news is good news” when expectations are high). The earnings call being “boring” was taken as a relief in itself​ (247wallst.com), meaning NVIDIA delivered what was expected (or slightly more) with no red flags.

Recent strategic updates from the company and others also color the outlook:
 

• DeepSeek and Competitive Response: Analysts have been asking about competitive threats like the low‐cost AI model from China’s DeepSeek. Interestingly, Oppenheimer noted that the rise of DeepSeek could actually be positive for NVIDIA as it might spur U.S. companies to invest even more in AI to maintain an edge​ (investopedia.com). Jensen Huang, on the call, downplayed the impact of such AI model news on hardware demand. The general take is that competition in AI models (software) will lead to more training and experimentation, which still requires NVIDIA’s hardware.

• Product Roadmap Confidence: Jensen Huang and NVIDIA’s management gave strong signals that next‐gen products are on track. Rumors about delays in the upcoming Blackwell Ultra GPUs were dispelled – NVIDIA is targeting an “extremely aggressive ramp” for its 300‐series Blackwell in data centers​ (247wallst.com). This reassured analysts that NVIDIA won’t cede any time advantage to competitors. Also, NVIDIA emphasized ongoing improvements in software and new use cases (like “AI agents” and robotics), which broaden the narrative beyond just cloud spending.

• Enterprise AI Focus: A notable strategic point (already touched on) was Huang stating enterprise AI will eventually surpass hyperscaler demand​ (247wallst.com). This is a medium‐term prediction, but it frames NVIDIA’s go‑to‑market strategy. We might see NVIDIA invest more in enterprise sales, OEM partnerships, and software that caters to non‑tech companies. This could lead to new partnerships or solutions over the next year aimed at enterprise adoption (e.g. certified systems, turn‑key AI appliances, etc.). Analysts likely view this favorably as it shows NVIDIA isn’t reliant solely on a few big buyers long‑term.

• Capital Allocation: NVIDIA’s management didn’t announce anything new on buybacks in this report (they had already authorized a large repurchase in the previous quarter). But the company’s cash flow suggests it could start buying back shares more aggressively, which analysts would generally welcome given the lack of better uses of cash at the moment (no big acquisition on the horizon after ARM’s attempt failed). A reduction in share count or support via buybacks could bolster EPS growth further. It’s a point to watch for future strategic updates.

• Regulatory Developments: Analysts are keeping an eye on U.S.-China policy. Any update there (positive or negative) can affect outlook. For example, if the U.S. delays or softens the next round of export restrictions, that’s a short‑term relief for NVIDIA (and would make analysts incrementally more positive). Conversely, if restrictions tighten, analysts might trim estimates for China sales. As of now, many analysts are aware of this risk but seem to be focusing on the strong demand elsewhere to offset it.

Overall, the analyst community remains very bullish, with price targets generally moving upward in recent months. Since the huge earnings beat in mid‑2023 (Q2), many analysts dramatically raised targets, and targets have continued to inch up with each successful quarter. Sentiment like “hurry up and buy NVIDIA ahead of earnings” was even expressed by some (e.g. Morgan Stanley included NVIDIA in a list of must‑own stocks into earnings) (cnbc.com). Such enthusiasm can be a double‑edged sword (all buys, no sells could mean peak optimism). But it reflects the fact that NVIDIA has consistently outperformed even lofty expectations, building credibility with the Street.

To sum up the sentiment: NVIDIA is viewed as a foundational long‑term AI winner by analysts. They acknowledge near‑term valuation is high but justify it with NVIDIA’s execution and the scarcity of pure AI plays of this caliber. Many have raised their price targets after earnings to factor in NVIDIA’s higher guidance and confidence. Barring an unforeseen downturn in AI demand, the Street’s base case is that NVIDIA will continue to post strong growth and thus remains a “strong buy”. Investors considering NVIDIA should note this consensus but also be mindful that when virtually everyone is on one side of the boat, the stock can be sensitive to any narrative changes. So far, however, NVIDIA’s management has given analysts little to fret about – keeping the longer‑term outlook very bright.

Is NVIDIA a strong long‑term investment? Based on its latest earnings and the analysis above, the answer leans yes – NVIDIA stands out as one of the most compelling long‑term growth stories in the market today, but with the caveat of an already rich valuation and some rising risks. The company’s financial performance is exemplary: explosive revenue growth, widening profitability, and dominance in a transformative tech trend. NVIDIA has effectively become the “arms dealer” of the AI revolution, supplying critical technology to every major player in the industry. Its entrenched market position (with ~90% share in AI accelerators​ (morganstanley.com)) and robust innovation pipeline (new GPUs, CPUs, and software) give it a strong moat and numerous avenues for future growth. Tailwinds like unabated AI infrastructure spending, expanding use of AI across sectors, and its own ecosystem effects are likely to fuel continued growth in the coming years.

However, long‑term investors should remain clear‑eyed about the risks. NVIDIA’s dominance will invite fierce competition – whether from traditional rivals like AMD/Intel or from the very customers it serves (big tech developing custom chips). Regulatory and geopolitical challenges also loom, and the stock’s valuation leaves little room for disappointment. The next year will be critical for NVIDIA to demonstrate that its recent growth spurt is not a one‑time spike but part of a durable trend of AI adoption. Key things to watch will be how sequential growth holds up (after this period of hyper‑growth), how NVIDIA navigates supply vs. demand, and whether any competitors start narrowing the gap.

The broader impact of NVIDIA’s success extends beyond just its stock. It is reshaping the semiconductor industry, pressuring peers to respond (benefiting some like Marvell and Broadcom that supply the ecosystem, challenging others like AMD and Intel to innovate or partner). Even companies like Google, while investing in their own AI chips, are heavily influenced by NVIDIA’s advancements, which in turn affects their strategies and spending.

In a long‑term portfolio, NVIDIA can play the role of a high‑growth core holding, particularly for those bullish on AI and technology. Its scale and profitability also give it a bit more resilience than a small‑cap tech play would have. That said, prudent investors might balance NVIDIA with other semi or tech stocks – for example, pairing it with more value‑oriented names (like an Intel or TSMC) or with “picks and shovels” plays (like Broadcom or Marvell) to mitigate the valuation risk. As of now, analysts and the market momentum are in NVIDIA’s favor, and the company’s execution has been flawless.

NVIDIA’s latest earnings confirm its status as a powerhouse with a long runway, making it an attractive long‑term investment candidate. The investor just needs to be comfortable with the volatility that can accompany a high‑growth, high‑expectation stock. With eyes open to headwinds such as competition and possible cyclical dips, one can regard NVIDIA as a pivotal player of the next decade in tech. If AI truly is the “new electricity” of the global economy, NVIDIA is supplying the generators – and its long‑term prospects will shine as bright as the trend it is driving.

Sources:

• NVIDIA Q4 FY2025 earnings press release​
  • nvidianews.nvidia.com nvidianews.nvidia.com
• Reuters – AI stocks and NVIDIA, export restrictions​
  • reuters.com reuters.com
• Investopedia – Analyst sentiment ahead of earnings​
  • investopedia.com investopedia.com
• Motley Fool/Nasdaq – NVIDIA vs AMD AI revenue and Wolfe Research comments​
  • nasdaq.com nasdaq.com
• 24/7 Wall St. – NVIDIA earnings call highlights (Jensen on enterprise vs CSP, networking/Broadcom)
  • ​247wallst.com 247wallst.com
• Morgan Stanley research – NVIDIA market share, margins, valuation context​
  • morganstanley.com morganstanley.com
• MarketBeat – Marvell’s AI strategy vs NVIDIA and data center growth
  • marketbeat.com marketbeat.com
• MarketScreener – Marvell Q3 FY25 results and guidance
  • in.marketscreener.com
• Reuters – Alphabet Q4 2024 results and AI capex plans​
  • reuters.com reuters.com