Table of contents

• Key takeaways
• Regulations
• Funding
• Product updates
• Partnerships
• M&A
• Appendices

Key takeaways

• The EU enacted the AI Act in May, marking the first comprehensive law in the world to regulate AI. In contrast, the UK's AI Regulation Bill has been temporarily dropped and is to be reintroduced at the State Opening of Parliament on July 17. Meanwhile, the US continued devising regulations, having announced the formation of an AI Safety and Security Board, while China released new GenAI draft legislature. 

• The GenAI ecosystem raised a record-breaking USD 19.6 billion in Q2 2024 over 60 rounds, compared with USD 6.0 billion in Q1 2024 and USD 4.2 billion in Q2 2023. This was heavily driven by CoreWeave’s USD 8.6 billion raise across two funding rounds (USD 7.5 billion and USD 1.1 billion), alongside xAI’s USD 6 billion round, collectively accounting for 74% of the total funds raised for the sector.

• Given the large rounds recorded, funding was heavily skewed toward the GenAI Infrastructure and GenAI Applications spaces. This contrasts with the fundraising pattern in the previous quarter, where startups in the FM space accounted for nearly 50% of total funds raised in the ecosystem. In fact, the average deal size for FMs fell by 43% YoY and 54% QoQ in Q2 2024, possibly due to an oversaturation of funds raised for the industry.

• FMs: The highlight of Q2 was the introduction of Apple Intelligence, a suite of GenAI models to assist users with daily tasks, integrated into Apple’s mobile devices, including the iPhone, iPad, and Mac. This system can complete tasks like writing text, summarizing it, creating images, and simplifying app interactions, enhancing user experience across its ecosystem. Microsoft also launched Copilot+ PCs, equipped with ~40 GenAI models that allow GenAI-powered apps to run without an internet connection. This aligns with the growing trend of embedding AI capabilities directly into consumer electronics, as seen in previous efforts by Samsung and Lenovo. 

• In addition, Q2 saw a flurry of new model launches, including large language models (LLMs) by Cohere (Aya 23), Google (Gemma 2 and PaliGemma), Amazon (Titan Text Premier), Microsoft (MAI-1), Mistral AI (Mixtral 8x22B), and Stability AI (Stable LM 2 12B); fine-tuned models by Mistral AI (Codestral) for coding and Google Deepmind (AlphaFold 3 and Med-Gemini) for medicine; and multimodal models by OpenAI (GPT-4o), Reka AI (Core), and Meta (Chameleon). 

• GenAI Infrastructure: Hardware developments dominated, with AMD announcing PC chips, Intel launching the new AI chip Gaudi 3, and Meta unveiling the next generation of Meta Training and Inference Accelerator (MTIA), hoping to challenge NVIDIA's dominance. Companies also focused on efficient processing and inference, with RaiderChip launching GenAI v1 for low-cost field programmable gate array (FPGA) devices and NVIDIA launching RTX A400 and A1000 GPUs that can run graphics with minimal power consumption. Google also announced a number of launches in the hardware space, including a sixth generation of tensor processing unit (TPU) chips, a custom-built arm processor Axion, and upgrades to its AI supercomputer.

• Other notable updates include 1) platforms to support and accelerate app development (Amazon’s Bedrock Studio, Microsoft’s Azure AI Studio, Google’s Firebase Genkit, and Weights & Biases’ W&B Weave), 2) data platforms (Hyper-Converged Data Platform by DataStax and Data Intelligence platform by Databricks), and 3) synthetic data generation platforms (Fine-Tune Studio by TELUS International and CodecLM by Google Cloud).

• GenAI Applications: Players upgraded their chatbots and other applications like video and image generation ones with new models and features to offer emotional nuance understanding and enhanced vision capabilities. For instance, OpenAI upgraded ChatGPT with GPT-4o, Google enhanced Gemini chatbot with the Gemini 1.5 Pro model, xAI introduced a multimodal Grok-1.5V model with vision capabilities, and video generation solution providers Luma and Runway upgraded their products with new models. OpenAI and Anthropic (Tool Use) also introduced data analytics features to their chatbots. As these tools become more accessible and capable, they will likely disrupt traditional content creation processes across various industries.

• Notable trends in Q2 included 1) partnerships to obtain data for model training such as with OpenAI, Reka AI, and Upstage, 2) infrastructure providers partnering to offer integrated solutions for app development and AI deployment (e.g., Pinecone and Gathr Data, Qualcomm Technologies and TensorOpera, and NVIDIA and Nutanix), and 3) model and AI infrastructure integrations into smart devices, with NVIDIA partnering with Intrinsic to develop autonomous robots and LG unveiling its service robot CLOi GuideBot powered by Google’s Gemini model. Google also partnered with DigiLens to power AR smart glasses. OpenAI also partnered with Apple to integrate ChatGPT into Siri and other apps. 

• GenAI players also reported partnerships with consultancy firms to accelerate enterprises' adoption of GenAI. These include Deloitte and Infosys’ partnerships with Intel; KPMG, McKinsey, and Accenture partnering with Google; and Roland Berger’s partnership with Microsoft. Additionally, PwC became OpenAI's first reseller and largest enterprise customer to date. 

• We tracked nine deals in Q2; the majority were reported in the GenAI Infrastructure space (Databricks acquired Langflow, NVIDIA acquired Run:ai, and Rebellions and Sapeon Korea announced plans to merge). OpenAI reported two acquisitions during the quarter: Multi, a video collaboration platform, and Rockset, an analytics database platform. Legal tech firm Harvey acquired Mirage to expand its AI capabilities, and Brazilian neobank Nubank acquired model developer Hyperplane. 

• GenAI directly integrated into consumer devices: As GenAI models continue to advance and become more accessible, we can expect a significant expansion of their integration across a wide range of consumer devices—from mobile phones and PCs to robotics platforms. Mobile phone developers have been at the forefront of this trend, incorporating GenAI capabilities into their products. This momentum is now extending into the PC market, as demonstrated by the launch of Microsoft's Copilot+ PCs, which feature built-in GenAI-powered assistants. Similarly, in the robotics domain, NVIDIA's partnership with Intrinsic to integrate its platform into autonomous systems and LG's unveiling of its service robot CLOi GuideBot, powered by Google's Gemini model, signal the growing adoption of these technologies. According to ABI Research, the annual shipments of devices capable of running GenAI on-device are projected to surge from nearly 200,000 in 2024 to over 1.8 billion by 2030. This trend is further supported by the introduction of new PC chips from hardware developers, such as AMD's recently announced Ryzen Pro 8040 Series and Ryzen Pro 8000 Series, designed to bring advanced AI processing capabilities to laptops and desktops. As these trends continue to unfold, the integration of GenAI into consumer electronics and robotics may become increasingly widespread. 

• Targeted data acquisitions to improve models while adhering to data privacy regulations: Amidst growing enforcement of intellectual property (IP) regulations, AI developers are increasingly adopting strategic partnerships and licensing agreements to acquire and safeguard data. This approach allows them to enhance the competitiveness of their models while maintaining integrity. For instance, OpenAI continued its trend of partnerships with media companies, building on momentum from previous quarters. This quarter, the company announced collaborations with a range of media entities, including The Atlantic, Vox Media, News Corp, Reddit, Dotdash Meredith, Stack Overflow, Financial Times, and TIME. Similarly, Reka AI, another prominent player in the GenAI space, joined the growing list of companies forging partnerships within the media industry. Additionally, some developers are exploring the use of synthetic data as a means to augment the volume of their training data, demonstrated by the launch of TELUS International's Fine-Tune Studio and Google Cloud's CodecLM to generate data for model training. 

Regulations: EU approves final AI Act, with certain regulations being phased in slowly

Analyst Take: Numerous regulators made strides in implementing AI safety legislation, the most notable being the final approval of the EU’s AI Act. Similar to China, the EU has a stringent stance against GenAI firms, which now need to navigate a complex array of regulations, including the General Data Protection Regulation (GDPR), copyright laws, and competition laws, in addition to the AI Act. This is in contrast to Japan, which relies on existing data privacy laws and non-binding guidelines to regulate AI firms. Meanwhile, the US is taking a more cautious approach, developing recommendations for safe AI use following President Biden’s Executive Order in Q4 2023.

• EU grants final approval for AI Act: The European Parliament gave its final approval in May, making it the first comprehensive law in the world to regulate AI. The law uses a risk-based approach, banning "unacceptable risk" uses like cognitive behavioral manipulation and social scoring. The act also sets rules for "general-purpose AIs" (GPAIs), like OpenAI's ChatGPT, with stricter regulations applying to GPAIs exceeding a certain compute threshold and ones that pose a "systemic risk." Some regulations from this new AI Act are to be phased in over a two-year period.

• UK’s AI Regulation Bill is dropped temporarily amidst other updates: The AI Regulation Bill, introduced to parliament in November 2023, was dissolved due to the country’s General Elections. The bill aimed to establish an AI authority to oversee AI regulations in the UK. However, party members plan to reintroduce the bill at the State Opening of Parliament on July 17. Meanwhile, regulators made progress by signing a Memorandum of Understanding with the US to develop a common approach to AI safety testing and by releasing a new AI safety evaluation platform.

• Department of Homeland Security (DHS) launches new AI Safety and Security Board to guide AI usage: The DHS announced the formation of an AI Safety and Security Board meant to oversee the use of AI within US critical infrastructure. The board's main responsibilities would pertain to developing recommendations and strategies for the safe and responsible use of AI. It will include key figures from tech firms, academia, government agencies, critical infrastructure entities, and civil rights communities.

• China releases new GenAI draft regulations: In May, China’s National Information Security Standardization Technical Committee (NISSTC) released draft GenAI regulations. These focused on securing training data, protecting AI models, and implementing comprehensive security protocols and guidelines for security assessments. This adds to the country’s existing GenAI regulations introduced in Q4 2023.

• Japan publishes AI Operator Guidelines: Japan published AI Operator Guidelines in April, aiming to consolidate three key previous publications into a comprehensive set of non-binding guidelines. These guidelines apply to commercial entities and are designed as a "living document," expected to adapt and evolve over time. They focus on transparency, fairness, and safety, including identifying risks across the AI lifecycle, such as biases, and developing content authentication systems like digital watermarks, among others. Following this, Japan introduced a voluntary international framework for GenAI regulations in May, with plans to provide technical support for GenAI risk reduction by establishing the Global Partnership on Artificial Intelligence (GPAI) Tokyo Center.

• Singapore launches Governance Framework to address GenAI-related concerns: The Deputy Prime Minister of Singapore announced the Model Governance Framework for GenAI to address concerns related to GenAI, emphasizing transparency and ethical governance to combat misinformation. The framework advocates for clearer personal data laws, collaboration on training data sets for low-resource languages, and technical solutions like digital watermarking to verify AI-generated content. Additionally, the Ministry of Communications and Information launched Project Moonshot, a testing tool to tackle security and safety issues related to LLMs.

Other notable industry updates 

• US regulators announced that they would launch antitrust inquiries into Microsoft, OpenAI, and NVIDIA, scrutinizing their partnerships and investments in the GenAI space. Additionally, OpenAI revealed plans to block Chinese users from accessing its AI products, a move likely to exacerbate the US–China divide. This follows an export ban on NVIDIA's LLM training chips to China.

• Stringent EU regulations caused a stir for Big Tech and OpenAI, which included the following: 
  • Apple delaying the launch of Apple Intelligence features in the EU due to regulatory concerns
  • Reported investigations of several partnerships between Microsoft and OpenAI on grounds of antitrust concerns
  • Meta pausing the training of its AI systems based on user data from the UK and the EU in response to regulatory pressure

• Similarly, UK regulators have tightened their belts by launching investigations into AI deals by Microsoft and Amazon on grounds of anti-competitive practices. They are also working toward introducing binding requirements for FMs and investing ~USD 107 million in AI research hubs and frameworks to achieve this.

• Notably, OpenAI formed a new Safety and Security Committee led by Sam Altman and other directors—tasked with developing its processes and safeguards. This is likely in response to the resignation of the company’s Chief Scientist and Co-founder Illya Sutskever, who played a pivotal role in the company’s AI safety initiatives. 

Funding: Q2 2024 funding surges with record-setting deals

Analyst Take: Funding for the GenAI ecosystem reached a fever pitch at USD 19.6 billion, its highest ever since 2021. This surge was backed by CoreWeave’s USD 7.5 billion debt round, marking one of the largest private financings ever recorded. The company is on a funding tear, having raised ~USD 11.5 billion over the last 12 months and reportedly functioning as the preferred provider of NVIDIA’s GenAI chips to Microsoft and OpenAI. Meanwhile, Elon Musk’s xAI raised USD 6 billion, representing the largest singular funding round recorded by the GenAI Applications industry, as the company seeks to compete with OpenAI’s dominant market position.

• In Q2 2024, startups in the GenAI ecosystem raised USD 19.6 billion across 60 rounds, the highest ever recorded. Funding more than doubled QoQ and rose 3.7x YoY. Heavy contributions came from two players: hardware infrastructure provider CoreWeave’s USD 8.6 billion raise across two rounds and chatbot developer xAI’s USD 6 billion round, collectively raising 74% of total funding in the GenAI ecosystem.

• Funding concentration across GenAI Infrastructure and GenAI Applications marked a notable departure from the previous few quarters, where startups in the FM space accounted for nearly 50% of total funds raised in the ecosystem. The change in trend could be attributed to an oversaturation of funding by major players in the space (Anthropic, Mistral AI, Baichuan Intelligence, and Aleph Alpha raised significant funding in the last two quarters), alongside lengthy development times associated with FMs.

• Meanwhile, the hardware infrastructure segment within the GenAI Infrastructure space consistently raised the majority of funding for the industry since Q2 2023. This underscores consistent investor appetite for fueling the growing GenAI boom, at a time when getting a hold of NVIDIA’s top-of-the-line training chips for LLMs is becoming increasingly difficult.

• On the GenAI Applications front, xAI's significant fundraise underscores investor confidence in Elon Musk's ability to challenge OpenAI, given his success in co-founding OpenAI and propelling Tesla to become one of the largest EV manufacturers in the world. 

• Given the large funding rounds, the average deal size for the quarter more than doubled, with the GenAI Applications and GenAI Infrastructure industries recording their highest-ever values. On the other hand, FM deal sizes roughly halved QoQ, reaching their lowest level since Q1 2023.

• Early-stage rounds (30) made up half of the total funding rounds recorded, backed by the GenAI Applications hub (20). Consequently, this industry had the highest number of startups raising funding for the first time since their inception across the GenAI ecosystem. This trend indicates growing investor interest in smaller-stage startups, which aim to capture market share in a space currently dominated by larger firms.

Notable funding rounds and investors

Repeat funding rounds during TTM Q2 2024

Analyst Take: CoreWeave generated the most investor interest in GenAI Infrastructure, backed by its cost-effective cloud service offerings relative to competitors, with other players in the industry only raising funding twice over the last 12 months. Anthropic and Mistral AI also completed three funding rounds, as the companies gear up to take on OpenAI’s dominant position in the space.

• Anthropic, CoreWeave, and Mistral AI each completed three funding rounds during the trailing 12 months (TTM) until Q2 2024. Following closely, four other companies—xAI, Pika, Speak, and Upstage AI—each secured two funding rounds. Additionally, 35 other companies recorded two funding rounds each.

Repeat funding rounds during TTM Q2 2024

Product updates: New model launches and integrations continue to drive activity

FMs

Analyst Take: The GenAI space continues to evolve rapidly, with model developers launching advanced language, multimodal, and fine-tuned models, as they are poised to become the core components of future intelligent systems. Significant developments this quarter were Apple's introduction of Apple Intelligence, a suite of GenAI models to assist users with daily tasks, integrated into their mobile devices, and Microsoft launching Copilot+ PCs. These align with the growing trend of embedding AI capabilities directly into consumer electronics, as seen in previous efforts by Samsung and Lenovo to integrate models into mobile phones. Additionally, there is a noticeable trend of developing compact yet powerful foundation models, aiming to make AI more accessible by minimizing the resources required for deployment, especially on consumer devices. This is evidenced by the launch of several small language models (SLMs), further driving the integration of GenAI into everyday lives.

• We tracked 64 product updates in Q2 2024 among nine disruptors (28 product updates) and 14 incumbents (36 product updates).

• Most updates involved new model launches, with LLMs leading the way. Companies such as Cohere (Aya 23), Google (Gemma 2 and PaliGemma), Amazon (Titan Text Premier), Microsoft (MAI-1), Mistral AI (Mixtral 8x22B), and Stability AI (Stable LM 2 12B) introduced new models. Notably, two Japanese tech companies, NEC and Fujitsu, launched their own models. Continuing momentum from the previous quarter, several SLMs were launched: OpenELM for on-device use by Apple as well as Phi-3 Mini for advanced reasoning and Phi-3-Vision for image analysis by Microsoft.

• Similar interest was seen in fine-tuned models, with models specific for various domains like coding (Mistral AI’s Codestral), medicine (Google Deepmind’s AlphaFold 3 and Med-Gemini), weather prediction (Microsoft’s Aurora), and education (Google’s LearnLM) launched within the quarter. Several multimodal models were also launched, including by companies like OpenAI (GPT-4o), Reka AI (Core), and Meta (Chameleon).

• A notable development during the quarter was the integration of models into devices, including Microsoft announcing new Copilot+ PCs equipped with Windows Copilot Runtime (a collection of ~40 GenAI models) and a local vector-based system that allows GenAI-powered apps (including third-party ones) to run without an internet connection. Apple also announced a personal intelligence system, Apple Intelligence, which includes multiple GenAI models to assist with everyday tasks integrated into mobile devices. 

• OpenAI launched several tools to guide model behavior, including Media Manager to control content use in AI training, a tool to detect images generated from DALL-E, and Model Spec framework to guide AI behavior. It also launched CriticGPT to detect and highlight errors in code developed by other AI services to better evaluate outputs from advanced AI systems.

GenAI Infrastructure

Analyst Take: The GenAI industry is witnessing a surge in demand for computational power. In response, chip developers like AMD, Intel, and Qualcomm as well as tech giants such as Microsoft, Meta, and Google are developing chips, challenging NVIDIA's dominance in the market. While these players work to fulfill the growing requirement for AI-optimized chips, they are also focused on developing more cost-efficient solutions to address the challenge of high training model costs. For instance, Google's Trillium TPU claims a 4.7x increase in compute performance per chip while also being 67% more energy-efficient than the previous generation. Similarly, Intel's Gaudi claims to deliver, on average, 50% better inference and 40% better power efficiency compared to NVIDIA's H100. As demand for computational power continues to surge, the ability to provide more cost-effective and energy-efficient AI processors could be a crucial differentiator for chip manufacturers. 

• We tracked 47 product updates in Q2 2024 among 20 disruptors (22 product updates) and 12 incumbents (25 product updates).

• Hardware infrastructure development accounted for around 30% of total updates. Notably, AMD announced PC chips for edge processing and Instinct MI325X GPU for data center workloads, Intel launched a new AI chip Gaudi 3, and Meta unveiled the next generation of MTIA. Several developments focused on efficient processing and inference for LLMs and accelerated graphics tasks as well, with RaiderChip launching GenAI v1 for LLM inference for low-cost FPGA devices, allowing unquantized LLMs to run on limited memory bandwidths. Radxa also launched AICore SG2300x, which allows users to run models directly on devices, while NVIDIA launched RTX A400 and A1000 GPUs that can run graphics and rendering tasks with minimal power consumption. 

• Google announced a number of launches in the hardware space, including a sixth generation of TPU chips, a custom-built arm processor Axion, and upgrades to its AI supercomputer for improved AI development and training. 

• Several companies launched tools and platforms to support and accelerate app development, including Amazon’s Bedrock Studio, Microsoft’s Azure AI Studio, Google’s Firebase Genkit, and Weights & Biases’ W&B Weave. 

• During this period, companies unveiled data platforms and synthetic data generation platforms to address the need for data solutions to meet the growing demand of GenAI models. For instance, DataStax introduced the Hyper-Converged Data Platform, while Databricks launched the Data Intelligence Platform specifically for asset maintenance and predictive analysis in energy applications. Additionally, TELUS International's Fine-Tune Studio and Google Cloud's CodecLM emerged as notable synthetic data platforms. 

GenAI Applications

Analyst Take: The upgrades to chatbots like OpenAI's ChatGPT and Google's Gemini, which now feature enhanced emotional understanding, vision capabilities, and data analysis features, underscore the industry's focus on delivering more versatile and intuitive user experiences. This push for improved multimodal and analytical capabilities would be crucial, as GenAI becomes more integrated into daily lives and workflows. Similarly, the introduction of new video and image generation models by providers like Luma, Runway, and Ideogram demonstrates the rapid expansion of GenAI's creative applications. As these tools become more accessible and capable, they will likely disrupt traditional content creation processes across various industries.

• OpenAI upgraded ChatGPT with its new model GPT-4o during the quarter, improving the user experience with emotional nuance understanding and enhanced vision capabilities to answer queries related to images or screens. Similarly, its competitors Google and xAI enhanced AI chatbot Gemini with the Gemini 1.5 Pro model andintroduced a multimodal Grok-1.5V model with vision capabilities, respectively. Video generation solution providers Luma and Runway upgraded their products with new models, while image-generating platform Ideogram also updated its image generator with new features. 

• Leading chatbot providers were seen enhancing their chatbots with advanced data analysis capabilities during the quarter to deliver a more versatile user experience. For instance, OpenAI announced a new data analysis feature for ChatGPT users, and Anthropic announced "Tool Use" for its AI assistant Claude, which allows it to autonomously interact with external data sources, application programming interfaces (APIs), and tools. Other notable new features include Pages for customized research presentations by Perplexity and a custom assistants feature by You.com.

• Four players launched new AI assistants, including Meta’s AI assistant built on the Llama 3 model, Baichuan launching an AI assistant built on upgraded LLM Baichuan 4, and Google launching Gemini Code Assist, an AI code completion and assistance tool. 

Partnerships: GenAI ecosystem sees increasing collaborations to drive technological advancements

Analyst Take: Growing collaborations within the companies in the GenAI industry appear to have a significant impact on the ecosystem's development and accessibility. By integrating GenAI models directly into products, companies are making these advanced capabilities available to end users. Similarly, partnerships between infrastructure providers, cloud platforms, and GenAI developers are enabling the creation of comprehensive, integrated solutions that streamline model development, deployment, and access. This not only improves the overall user experience but also lowers barriers to entry, allowing more organizations to leverage GenAI technologies. Overall, these collaborative efforts are driving the GenAI industry, enabling the transformative potential of these technologies to be more widely realized across various sectors.

• We observed 118 partnerships in Q2 2024. 28% of these accounted for partnerships within the GenAI ecosystem. 32% of the partnerships were from the FM hub, while GenAI Infrastructure partnerships accounted for 25%. Half of the partnerships were for product development, while 28% accounted for customer partnerships and 22% were sales partnerships. 

• In the FM space, a notable number of partnerships were reported to obtain data for model training. OpenAI continued its spree from last quarter, partnering with several media companies, including The Atlantic and Vox Media, News Corp, Reddit, Dotdash Meredith, Stack Overflow, Financial Times, and TIME. Other companies also reported similar partnerships, such as Reka AI partnering with Shutterstock and Upstage partnering with Flitto to create multi-language AI datasets. 

• In the Infrastructure space, partnerships were reported to offer integrated solutions for app development and AI deployment. For instance, Pinecone partnered with Gathr Data to integrate fabric and data processing capabilities with vector search technology, Qualcomm Technologies partnered with TensorOpera to support the entire GenAI lifecycle, including computing, training, deployment, and storage, and Cisco partnered with NVIDIA to simplify AI adoption for Ethernet network-based enterprises. NVIDIA also partnered with Nutanix to simplify enterprise GenAI adoption by offering an integrated product to develop GenAI solutions at the edge.

• Model and AI infrastructure integrations into smart devices continued this quarter, with NVIDIA partnering with Intrinsic to integrate its platform into autonomous robotics and LG unveiling its service robot CLOi GuideBot powered by Google’s Gemini model. Google also partnered with DigiLens to power AR smart glasses with its Gemini models, signifying ongoing efforts to incorporate GenAI to enhance functionality and user experiences.

• Around 28% of partnerships were between companies within the GenAI ecosystem. Notable areas of partnership within the space include the following:
  • Integrating models into products: e.g., legal tech company Harvey partnered with Mistral AI to use its models
  • Using infrastructure service to develop models: e.g., Reka AI partnered with Oracle to train its models on Oracle Cloud
  • Co-developing storage, compute, and cloud service solutions: e.g., NVIDIA and Hewlett Packard partnered to offer NVIDIA AI Computing by HPE, and Cerebras Systems partnered with Dell to offer infrastructure to develop large models
  • Improving accessibility of models and products through cloud platforms: e.g., Secoda and Google Cloud partnered to offer data management solutions on its Marketplace; AI21 Labs partnered with Snowflake, Microsoft Azure, and Amazon Bedrock to make its Jamba-Instruct model available on cloud platforms; and Cohere partnered with Amazon to incorporate Command R+ and Command R models to Amazon Bedrock 
  • Running AI models directly on chips: e.g., Hugging Face partnered with Amazon to enable users to run custom AI models on its chip Inferentia2, and Meta partnered with Qualcomm to optimize the performance of Llama 3 LLMs for on-device execution
  • Infrastructure players partnered with cloud providers to offer integrated app development solutions: e.g., Snowflake andNVIDIA, Microsoft and NVIDIA, DataStax and Google Cloud, and MongoDB and Google Cloud

• Another notable development was OpenAI's much-anticipated partnership with Apple to integrate ChatGPT into Siri and other apps.

• GenAI players reported partnerships with consultancy firms to accelerate the adoption of GenAI by enterprises. These include Deloitte and Infosys’ partnerships with Intel, KPMG, and McKinsey; Accenture partnering with Google; and Roland Berger’s partnership with Microsoft. Additionally, PwC became OpenAI's first reseller and largest enterprise customer to date.

Incumbent engagement in the startup ecosystem

• During Q2 2024, there was a low level of investment activity among incumbents relative to the previous quarter. However, NVIDIA stood out as the most active, co-leading the funding round for model developer Twelve Labs. NVIDIA also participated in several other investment rounds, including Mistral AI, Cohere, Scale AI, and Perplexity AI. Salesforce joined NVIDIA in participating in the Cohere and Mistral AI rounds. Notably, the Scale AI round attracted multiple incumbents, such as NVIDIA, Cisco, Intel, ServiceNow, AMD, Amazon, Meta, and Qualcomm. Additionally, Amazon invested in model developer H AI and data curation platform DatalogyAI, while Google invested in Israeli cybersecurity startup Aim Security. 

• Amazon led the incumbent collaboration with disruptors in Q2 2024, partnering with model developers Upstage AI, AI21 Labs, Anthropic, and Cohere to sell their models on its cloud platform. Google partnered with infrastructure providers DataRobot, DataStax, and MongoDB to offer integrated solutions to accelerate AI app development. Another notable partnership was Microsoft and Oracle partnering with OpenAI to extend Azure AI on Oracle Cloud to provide additional capacity for OpenAl. These partnerships demonstrate the dynamic and collaborative nature of the GenAI industry as it continues to evolve, with leading players working together to drive technological advancements and meet growing demand. 

Incumbent partnerships in the GenAI ecosystem (Q2 2024)

M&A: Infra players dominate while OpenAI records two acquisitions

Analyst Take: Similar to the previous quarter, the majority of M&A deals were reported in the GenAI Infrastructure space. However, OpenAI was the highlight during the quarter, which acquired two companies, marking its third acquisition since Global Illumination in August last year. The acquisition of the video collaboration platform Multi suggests OpenAI's intent to expand its offerings potentially integrating video generation and collaboration features into its suite of GenAI products. The Rockset acquisition would improve OpenAI's data retrieval and analytics capabilities. Databricks also continued its acquisition spree with Langflow to offer an all-inclusive GenAI stack with flexible deployment options.

• Keeping in line with its broader strategy of investing in enterprise solutions, OpenAI reported two acquisitions during the quarter: Multi, a video-collaboration platform, and Rockset, an analytics database platform. While the exact purpose of Multi’s acquisition was not disclosed, the company will integrate Rockset's technology to power the retrieval infrastructure backing its product suite. 

• Several M&As were reported in the GenAI Infrastructure space. Data infrastructure and data analytics solutions provider Databricks acquired Langflow to offer an all-inclusive GenAI stack with flexible deployment options, marking the company’s third acquisition for the year. Meanwhile, NVIDIA announced the acquisition of Kubernetes-based workload management and orchestration software provider Run:ai to help customers use their AI computing resources more efficiently. South Korean AI chip developers Rebellions and Sapeon Korea announced plans to merge to develop chips focusing on neural processing units used in AI. 

• One acquisition was reported in the GenAI Applications space; legal tech firm Harvey acquired Mirage to expand its AI capabilities. Notably, Brazilian neobank Nubank acquired Hyperplane, a developer of FMs for banks, to use Hyperplane's AI capabilities to generate insights and enhance its decision-making processes.

Appendices

Appendix 1

Funding by hub

Appendix 2

Startups that raised external funding for the first time in Q2 2024

Appendix 3

Disruptor partnerships in Q2 2024

Incumbent partnerships in Q2 2024

Partnerships within the GenAI ecosystem in Q2 2024