Yes, questions remain about whether AI chipmaker NVIDIA can maintain its momentum in 2025. NVIDIA’s upcoming Blackwell AI chips have encountered thermal problems when integrated into server racks designed to house up to 72 units. Despite these recent technical hurdles, the company remains an innovation juggernaut.

1. Why NVIDIA is the R&D leader of 2024

The company had an exceptional 2024, which earned it the top spot in this list of public R&D-focused companies. It’s GPUs have played an integral role in launching the current AI wave, which may have surpassed peak hype but is still going strong. In Q4 2024, its revenue soared to $22.1 billion, a 22% quarter-over-quarter jump and a 265% year-over-year increase. Full-year revenue reached $60.9 billion, more than doubling (+126%) the previous year’s total. Its market cap eclipsed $3 trillion in June 2024, briefly making it the world’s most valuable company, ahead of Microsoft and Apple.

The company’s gains were rooted in the firm’s innovation pedigree, which dates back to the early 1990s when its co-founders foresaw the need for novel computing platforms. In 1999, the company introduced the GeForce 256, the first GPU integrating transform and lighting calculations, enhancing 3D graphics performance. In 2006, Nvidia launched CUDA, a parallel computing platform and programming model enabling developers to use GPUs for general-purpose computing tasks beyond graphics. This year, alongside introducing the Blackwell architecture (a Blackwell board is shown on the right) with its B100 and B200 data center accelerators featuring dual-die designs with 208 billion transistors, the company launched its NVIDIA Inference Microservices (NIM). This suite of cloud-native microservices is designed to streamline the deployment of AI models across various platforms. NIM offers optimized inference on a spectrum of AI models, supporting tasks ranging from data processing to large language model customization, inference, retrieval-augmented generation, and implementing guardrails in AI models. The company has also bolstered its market position through strategic partnerships, collaborating with SoftBank for supercomputer deployments and Trend Micro to develop AI-driven security tools for data centers.

Rank	Company	AVG PATENT CITATIONS	Tech Diversity	R&D Intensity	Patent Family Size	PATENT Jurisdiction Count	AVG CITATIONS PER Publication	Market Cap
1	NVIDIA Corporation	16.36	0.44	11.15%	2.62	9	19.10	3482.77B
2	Meta Platforms	144.51	0.59	27.37%	2.35	6	34.47	1398.77B
3	Salesforce	64.68	0.42	13.66%	1.69	5	26.14	310.95B
4	Alphabet Inc.	9.57	0.51	14.00%	2.86	18	17.94	2119.19B
5	Apple Inc.	79.48	0.58	8.18%	2.77	21	9.24	3401.05B
6	Microsoft	17.09	0.48	11.50%	2.28	21	15.99	3085.48B
7	Tencent Holdings	4.01	0.51	9.83%	1.52	11	17.07	476.14B
8	Adobe Inc.	9.92	0.20	18.12%	3.13	4	17.43	221.58B
9	Samsung Electronics	3.33	0.67	10.87%	1.43	25	6.31	301.85B
10	AstraZeneca PLC	3.47	0.51	21.95%	12.95	35	41.33	202.73B
11	Cisco Systems	99.35	0.40	0.15%	3.87	8	6.12	229.04B
12	AbbVie Inc.	5.39	0.50	0.17%	6.02	32	21.75	291.56B
13	Taiwan Semiconductor Manufacturing (TSMC)	6.40	0.42	0.08%	1.79	4	9.73	964.66B
14	Novo Nordisk A/S	63.82	0.47	0.23%	5.66	30	16.96	472.80B
15	Broadcom	4.77	0.60	0.19%	2.43	4	11.06	769.90B
16	Qualcomm	7.72	0.32	0.24%	1.87	26	13.40	178.32B
17	Oracle Corporation	11.37	0.45	0.17%	2.16	11	4.72	509.15B
18	Advanced Micro Devices (AMD)	5.82	0.28	0.28%	2.42	7	3.78	218.92B
19	Abbott Laboratories	132.72	0.57	0.07%	3.80	16	15.77	204.06B
20	SAP SE	14.63	0.29	0.19%	1.64	5	4.73	272.39B
21	IBM Corporation	5.40	0.55	0.12%	2.15	17	12.75	189.54B
22	Bristol-Myers Squibb	8.36	0.51	0.20%	4.57	31	16.04	114.02B
23	Baidu	5.66	0.39	0.17%	1.41	8	14.11	29.65B
24	GSK plc	12.87	0.50	0.19%	3.43	25	16.09	70.17B
25	Merck & Co. Inc.	36.00	0.56	0.22%	3.09	27	15.27	243.63B

2. Meta Platforms show highest citation impact in tech sector

Meta leads all firms in research impact at 144.5 citations and 27.3% R&D intensity while maintaining one of the highest R&D investments. Meta Platforms doubled down on AI this year, pumping 27.5% of its revenue into R&D — the highest figure among the tech companies in the dataset. It released Llama 3.1 405B in July, “the world’s largest and most capable openly available foundation model.” The company also published on Sapiens, a family of high-performance vision models designed for real-world environments. The Ray-Ban Meta glasses gained significant traction, with sales exceeding 700,000 units in their first year. Meta’s research seems to be making waves, too. The firm is getting about 144 citations per patent. They also dropped the Quest 3S VR headset in October for $299.

3. Salesforce demonstrates strong citation performance

Salesforce is investing heavily in AI agents, launching its Agentforce Platform, and rolling out new AI features for its retail and service platforms. The platform allows companies to automate tasks such as resolving customer cases and qualifying sales leads, improving efficiency. While Salesforce spent less on R&D than some others (about 13.7% of revenue), it is achieving a high impact, with about 65 citations per patent.

4. Alphabet ranks third in innovation score

It was something of a mixed year for Alphabet. It launched Gemini, its answer to GPT-4, but ran into controversy as its Gemini AI image generator produced historically inaccurate and culturally insensitive images. Conversely, its cloud business is booming — up 35%, hitting $11.35B in Q3. The company invests about 14% of its revenue into R&D and spreads its bets across computing and communications patents. The company also launched AlphaFold 3, which extends protein structure prediction to include interactions with DNA, RNA, and other molecules. Two Google DeepMind execs, Sir Demis Hassabis and Director John Jumper, jointly won the 2024 Nobel Prize in Chemistry for their work creating AlphaFold.

5. Apple achieves high citations with lower R&D spending

Apple had a busy year. In October, it rolled out “Apple Intelligence,” baking AI features into iPhones, iPads, and Macs. The company also launched the iPhone 16 and iPhone 16 Pro. Despite its historical reputation as an innovator, Apple spent only about 8.2% of its revenue on R&D—significantly less than its competitors. Conversely, its patents still receive an average of about 80 citations. In February, it released the Vision Pro mixed-reality headset.

6. Microsoft maintains balanced innovation metrics across board

In 2024, Microsoft significantly advanced its cloud and AI initiatives by expanding its Copilot AI assistant, powered by OpenAI’s GPT-4 and GPT-4o models, across Microsoft 365, Windows, and Edge platforms, enhancing user experiences through advanced AI capabilities. Microsoft invested heavily in on-device AI processing with proprietary data center infrastructure chips designed to accelerate AI applications and enhance data security to support this expansion. Furthermore, Microsoft strengthened its ecosystem through new cloud partnerships, notably with Dell, which announced AI and cybersecurity advancements, including APEX File Storage and Copilot services for Azure, at Microsoft’s Ignite conference.

7. Tencent innovated in gaming and payments

The Chinese firm concentrated in 2024 on payment innovation and gaming technologies, introducing a palm recognition payment system. It partnered with Visa to pilot the technology in Singapore, allowing Visa cardholders from participating banks to make payments using their palm at select merchants. The company achieved an 8% overall revenue increase to 167.19 billion yuan while investing 9.8% in R&D. Its patents averaged 4.0 citations. The R&D investment held steady in the third quarter, in which Tencent invested 17.9 billion yuan in innovation, accounting for roughly 10.7% of its total revenue. In the third quarter, the company’s revenue grew by 8% to 167.19 billion yuan in the third quarter of 2024

8. Adobe shows focused R&D investment

Like many of the firms on this list, Adobe sharpened its focus on AI in 2024, adding new features to the Creative Cloud suite. A Photoshop upgrade included automatic background removal and generative fill, which enables users to specify how to fill portions of an image with text commands. Adobe also improved its Firefly generative AI model. It also debuted new video generation capabilities, including the Firefly Video Model for AI-powered video generation and editing. In addition, the launch of GenStudio for Performance Marketing aimed to accelerate global advertising and marketing campaigns through generative AI. The company allocated 18.1% of its revenue to research and development (R&D), contributing to an average of 9.9 citations per patent.

9. Samsung Electronics leads in tech diversity

In 2024, Samsung Electronics pushed AI integration across its product lines, launching its ‘AI for All’ vision at CES. The Galaxy S24 series showcased this approach with features like Live Translate for real-time language conversion, Generative Edit for AI-powered image modification, and Circle to Search for visual information queries. Its Neo QLED TVs got an AI boost with the NQ8 AI Gen3 processor, using deep learning to enhance the picture and sound quality in real-time. While Samsung’s patent citations averaged 3.3 per patent, it maintained the highest tech diversity score (0.67) among major companies, covering semiconductors, consumer electronics, and telecommunications advances. Its  R&D spending hit 10.87% of revenue, reflecting its push to compete in AI while maintaining its broad technical portfolio.

10. AstraZeneca expands on multiple fronts

In the first nine months of 2024, AstraZeneca committed 21.9% of its revenue to R&D. That sum helped support the inauguration of the CHF 1.2 billion pRED Innovation Center in Basel, Switzerland, featuring 150 labs to foster scientific collaboration. It demonstrated in its weight-loss drug pipeline with CT-388, a dual GLP-1/GIP receptor agonist, achieving an 18.8% average weight reduction in a Phase 1b trial. AstraZeneca also acquired Carmot Therapeutics for $3.1 billion, gaining access to its innovative metabolic disease therapies, including CT-388.

11. Cisco systems records high citation count

In 2024, the networking giant found itself in a strategic realignment to focus on high-growth areas such as AI and cybersecurity. Yet it strengthened its offerings by integrating AI-powered tech with networking, security, and observability software. It entered the AI server market through a collaboration with NVIDIA. It also acquired Splunk for $28 billion, bolstering its security and observability capabilities. Additionally, Cisco introduced Spatial Meetings for immersive 3D collaboration and reported a 20% year-over-year increase in product orders, reflecting robust market demand. Cisco allocated 15.2% of its revenue to R&D, averaging 99.3 citations per patent.

The list resumes after the sidebar

Tech, pharma (and chemistry) are converging

12. AbbVie fortifies its neuroscience portfolio

In 2024, AbbVie strengthened its neuroscience portfolio by acquiring Aliada Therapeutics for $1.4 billion. The company focused on ALIA-1758, an anti-amyloid beta antibody that uses novel blood-brain barrier technology to target Alzheimer’s disease. Additionally, AbbVie received FDA approval for Vyalev, a 24-hour subcutaneous infusion therapy for advanced Parkinson’s disease, offering a new non-surgical option to improve motor fluctuations.

13. TSMC pioneers 3nm and GAA transistor technology

Taiwan Semiconductor Manufacturing Company (TSMC) remains a major semiconductor manufacturer. In 2024, it introduced its A16 process technology with gate-all-around (GAA) nanosheet transistors and a backside Super Power Rail (SPR) for enhanced performance and efficiency. TSMC also invested in high-NA EUV lithography systems for future production nodes. It expanded its U.S. presence with a $65 billion investment in Arizona, supported by up to $6.6 billion in CHIPS Act funding, to build three advanced fabs. While allocating 7.8% of its revenue to R&D, TSMC achieved an average of 6.4 citations per patent YTD in 2024.

14. Novo Nordisk shows strong citation performance

The Danish GLP-1 innovator has progressed in its weight management portfolio, expanding into heart disease treatments. Its CagriSema trials showed 25% body weight reduction potential, while Wegovy’s launch in China at 1,400 yuan ($193) per pack addressed a market where over half the population is overweight. In March, it acquired German biotech Cardior Pharmaceuticals for €1.025 billion, adding RNA-based heart disease therapies to its pipeline. They also invested in the Gefion AI supercomputer to speed up drug development.

15. Broadcom boasts high-tech diversity

Broadcom diversified its technology portfolio in 2024. The company launched VeloRAIN for enhanced AI workload performance, introduced improvements to VMware Cloud Foundation, and released 400G PCIe Gen 5.0 Ethernet adapters. While generating 4.8 citations per patent, Broadcom maintained a high tech diversity score of 0.595, spanning networking, cloud, and AI applications. The company also entered discussions with OpenAI regarding developing a new AI data center chip.

16. Qualcomm continues to invest heavily in R&D

The San Diegan company continues to expand its mobile and PC processor offerings by introducing the Snapdragon 8 Elite platform, which features the Oryon CPU architecture. The platform delivers substantial performance and power efficiency improvements, supporting AI-driven applications for enhanced user experiences. The company also expanded its reach into the PC market with the Snapdragon X Elite and X Plus processors, which offer superior performance and efficiency and are integrated into notable devices like Microsoft’s latest Surface models. The firm spends 23.8% of its revenue on R&D.

17. Oracle keeps steady metrics

Oracle revamped its healthcare and cloud offerings in 2024. It introduced a next-generation Electronic Health Record (EHR) system built on Oracle Cloud Infrastructure (OCI) to embed AI across clinical workflows, automate processes, deliver data at the point of care, and simplify tasks for healthcare professionals. The company also enhanced its HeatWave platform with generative AI capabilities, including in-database large language models (LLMs), an automated in-database vector store, and scale-out vector processing. Additionally, Oracle expanded its distributed cloud capabilities by making its database services available across major platforms like Amazon Web Services (AWS), Microsoft Azure, and Google Cloud. In the first nine months of 2024, Oracle allocated 16.8% of its revenue to research and development.

18. AMD leads in R&D intensity

While AMD trails rival NVIDIA when it comes to AI, it significantly bolstered its processor lineup in 2024 with the introduction of the Zen 5 microarchitecture, powering the new Ryzen 9000 series and EPYC 9005 series “Turin” server chips. AMD also launched the Ryzen AI 300 Series for laptops, delivering up to 50 TOPS of neural processing performance. The company dedicates 27.9% of its revenue to research and development — the highest R&D intensity in the tech companies surveyed. It had an average of 5.8 citations per patent.  Despite these strategic product launches, AMD announced a 4% reduction in its workforce, aiming to reallocate resources towards high-growth areas like AI and data center products.

19. Abbott Labs achieves second-highest citations

Abbott Laboratories continues to innovate in diabetes care and heart devices. In 2024, the company launched Lingo, an over-the-counter continuous glucose monitor that lets people track their blood sugar in real-time. The company also got FDA approval for its TriClip system, which treats tricuspid regurgitation. Patients showed significant improvements that lasted over a year. In 2024, it also rolled out AVEIR, the first dual-chamber leadless pacemaker. While spending a relatively modest 6.9% of revenue on R&D, Abbott’s innovations are highly impactful. Its patents average 132.7 citations each, among the highest in healthcare.

20. SAP joins AI copilot race

In 2024, SAP deepened its AI integration across its enterprise software suite by introducing “Joule,” a genAI copilot designed to provide contextual insights and streamline tasks. The company also launched the SAP Knowledge Graph, which maps relationships and context within its data landscape. Furthermore, SAP expanded its AI partnerships, including collaborations with Microsoft and Meta, to enhance AI integration and automation capabilities. SAP allocated 19.4% of revenue to research and development in the first 9 months of the year. It had an average of 14.6 citations per patent.

21. IBM maintains broad tech diversity

In October 2024, IBM launched its Granite 3.0 models on the Watsonx platform, offering open-source large language models (LLMs) for enterprise AI applications. These models, ranging from 3 to 34 billion parameters, are available under the Apache 2.0 license. The company also launched a modular quantum computer featuring 156-qubit Heron processors, capable of executing quantum circuits with up to 5,000 two-qubit gate operations. Also, in 2024, the company partnered with Amazon Web Services (AWS), Microsoft, and Meta to enhance AI governance and integrate third-party models into its AI and data platform, Watsonx. Despite a moderate citation impact of 5.4, IBM maintained broad technical diversity with a score of 0.554, encompassing computing, communications, and electronics patents.

22. Bristol-Myers Squibb shows balanced metrics

In 2024, Bristol Myers Squibb (BMS) advanced its oncology portfolio. One highlight was FDA approval for Breyanzi, a CAR T cell therapy, for treating specific types of leukemia and lymphoma in adults. Outside of oncology, BMS developed Penbraya, a pentavalent meningococcal vaccine offering broader protection against the disease. In the first three quarters of 2024, it invested 20% of its revenue in R&D and achieved an average of 8.4 citations per patent. In addition, BMS announced a $1 billion investment in China, including a new R&D facility in Beijing, to strengthen its regional capabilities.

23. Baidu is ascending in AI

In 2024, Baidu launched  I-RAG, a text-to-image generation technology tapping its extensive search capabilities to tackle inaccuracies in AI-generated images. The company also launched Miaoda, a no-code platform utilizing its large language model (LLM) to enable users to develop software applications without coding expertise. Baidu’s Ernie platform now handles 1.5 billion daily user interactions, a notable increase from earlier in the year. In autonomous technology, Baidu deployed over 400 vehicles in Wuhan with plans to expand to 1,000. Despite its momentum, it trails its rivals in patents, averaging 5.7 citations per patent. A Baidu Brain chip with high memory bandwidth and performance is shown here.

24. GSK demonstrates consistent R&D investment

In 2024, GSK bolstered its respiratory and vaccine portfolios. The FDA expanded the approval of Arexvy, GSK’s respiratory syncytial virus (RSV) vaccine, to include adults aged 50 to 59 at increased risk for RSV lower respiratory tract disease. In addition, GSK progressed in developing an eco-friendly version of its Breztri inhaler by transitioning to a next-generation propellant with 99.9% lower global warming potential than current propellants. This new propellant technology is projected to reduce GSK’s greenhouse gas emissions from 6.7 million to 5.4 million tons annually. The company invested almost one-fifth (19.5%) of its revenue in the analysis period (first three quarters of 2024) of its revenue in R&D. It has an average of 12.9 citations per patent.

25. Merck & Co. remains a Big Pharma leader

On the heels of its $10.8B acquisition of Prometheus in 2023, Merck continued establishing partnerships and achieving regulatory milestones this year. In November, the company entered a licensing agreement with LaNova Medicines valued at up to $3.3 billion to develop and commercialize LM-299, an investigational cancer therapy targeting PD-1 and VEGF proteins. Earlier in the year, Merck won  FDA approval for two major products. In March, the FDA approved Winrevair (sotatercept-csrk) for adults with pulmonary arterial hypertension (PAH). By June, Capvaxive, a 21-valent pneumococcal conjugate vaccine, won approval for adults. The vaccine addresses serotypes responsible for approximately 84% of invasive pneumococcal disease in individuals aged 50 and older. Merck invested heavily in innovation, reporting $22.981 billion in R&D expenses for the twelve months ending September 30, 2024. Patent activity also showed growth, with a 0.75% increase in filings and a 0.62% rise in grants in the second quarter compared to the first. Financially, the company posted $16.7 billion in worldwide sales in the third quarter of 2024, a 4% jump from the same period in 2023. With Keytruda’s patent set to expire in 2028, Merck faces potential revenue declines as biosimilar versions hit the market.

---

Methodology: ML helped pin down ‘innovation’

This study looked at a diverse set of mostly R&D-heavy players. Cardinal Health was included owing to its significant revenues. Organizations considered included  3M, AbbVie Inc., Abbott Laboratories, Adobe Inc., Advanced Micro Devices (AMD), Alibaba Group, Alphabet Inc., Apple Inc., AstraZeneca PLC, Baidu, Baxter International, Becton Dickinson, Boeing, Boston Scientific, Bristol-Myers Squibb, Broadcom, Cardinal Health, Cisco Systems, Eli Lilly and Company, Ford Motor Company, GE HealthCare Technologies, GSK plc, General Electric, General Motors, Honeywell International, IBM Corporation, Intel, Johnson & Johnson, Lockheed Martin, Medtronic, Mercedes-Benz Group, Merck & Co. Inc., Meta Platforms, Microsoft, NVIDIA Corporation, Nokia Oyj, Northrop Grumman, Novartis AG, Novo Nordisk A/S, Oracle Corporation, Pfizer Inc., Qualcomm, Roche Holding AG, Philips, SAP SE, Salesforce, Samsung Electronics, Sanofi, Siemens AG, Siemens Healthineers, Stryker, Taiwan Semiconductor Manufacturing (TSMC), Takeda Pharmaceutical, Tencent Holdings, Tesla, Toyota Motor Corporation, and Volkswagen AG.

Amazon was not considered as it does not break down its R&D spending in its financial reports.

The innovation score was calculated as a weighted ensemble of predictions from the Random Forest and Lasso regression models. A tuned Random Forest Regressor was trained on the features (X) to predict the target variable (financial_score), which included revenue and stock performance metrics fetched from yfinance. The model informed the selections of the most essential features correlated with innovation, often a nebulous word. Similarly, a Lasso regression model (The name Lasso is an acronym for “Least Absolute Shrinkage and Selection Operator regression”) was used with cross-validation and trained on the same features, yielding predictions (lasso_pred) using a sparse set of features with non-zero coefficients. These predictions were combined into a final innovation score. A weighted average gave greater emphasis (60%) to Random Forest predictions, which are adept at capturing complex, non-linear relationships without the computational expense of deep learning models, which also tend to fare best with more data than was available in the training set here. Patent figures are from the public patent database in Google Cloud’s BigQuery. Most publication data is from OpenAlex, and a handful of firms’ data is from Elsevier SCOPUS. Financial data are from various sources, including company filings, while stock data is from Yfinance.

Machine learning model correlations with financial performance Metric Correlation p-value citation_impact 0.369874 0.004629 patent_impact 0.228736 0.087001 rd_efficiency -0.256252 0.054347

Regression results Summary statistics Statistic Value Dep. Variable financial_performance R-squared 0.200 Adj. R-squared 0.121 F-statistic 2.548 Prob (F-statistic) 0.0392 No. Observations (Organizations) 57