How AI VC funding in the 1st half of 2024 can guide your AI startup strategy
Top 10 most invested AI companies and booming of AI in search, coding, education and pharmaceuticals.
As the amount of funding raised by AI startups continues to increase, a divergence in the market has become more pronounced this quarter, where camps debate over the following key positions:
- The “fight” between closed-source and open-source
- If the GPU shortage will continue into the mid-term, and strategies deployed in these camps
- When AI investments will bring forth a return on investment
This article will offer an overview of the AI VC funding landscape in the last 6 months, highlight some of the emerging trends in AI, and offer some realities for AI founders to consider.
The Generative AI VC Funding Landscape
We previously divided the Generative AI landscape into four layers as per the diagram below. For more details on each of the Generative AI and LLM tech layer, please refer to “Navigating LLM & GenAI Stack: Opportunities and Challenges for Asian Startups and Enterprises.”
We’ve categorized the Top 10 AI investments in the above framework, and we will go into detail below.
Layer 0 — the Generative AI Infrastructure Layer
Solutions and companies that help improve AI chips’ efficiency, attracted significant investments. For example, Nvida acquired Israel-based startup Run:AI at a price point between $700 million and $1 billion to improve its chip design and GPU utilization based on a Kubernetes-based GPU orchestrator developed by Run:AI. Its $300M acquisition of Deci AI aims to automates the design of deep learning models to improve performance. This trend underscores the growing demand for specialised hardware to support the increasing computational requirements of advanced AI models.
We also saw Microsoft invest $1.5 billion in Abu Dhabi’s G42 (a cloud service provider in Abu Dhabi) to accelerate AI development and global expansion in April. That said, this deal is now being scrutinised by US lawmakers, given G42’s previous ties with China.
Layer 1 — Foundational Model Layer
Foundational model companies still received the most financing. xAI, Mistral, Zhipu, etc., all received large amounts of financing. The most significant single financing in Q2 came from xAI, an artificial intelligence startup founded by Musk. On May 26, xAI announced the completion of a $6 billion Series B financing, which is one of the most significant financing events in the field of generative AI this year. The post-investment valuation of xAI is as high as $24 billion, officially entering the first echelon.
In early June 2024, Cohere, a Canadian artificial intelligence startup specialising in large language models, successfully raised $450 million in a new funding round. This significant investment saw participation from major tech players, including Nvidia, Salesforce Ventures, and Cisco. This substantial sum may only be the initial phase of a more significant fundraising effort. Reports suggest that Cohere is actively seeking additional capital, aiming for a total valuation of $5 billion.
In China, Zhipu AI (GLM) recently secured $400 million in a new funding round, with Prosperity7, a fund managed by Saudi Aramco’s venture capital arm, as the investor. This investment set a new record for the highest single financing amount for a large domestic model in China, bringing Zhipu AI’s post-investment valuation to approximately $3 billion. The company’s flagship product, ChatGLM, has been upgraded to GLM-4.
Compared to Q1 2024, these Q2 investments represent a significant escalation in both deal size and frequency. In Q1, the largest AI deal was Anthropic’s $450 million round. The dramatic increase in funding amounts and valuations in Q2 reflects growing investor confidence in AI’s potential to deliver substantial returns, despite the uncertain economic climate.
Layers 2 & 3 — Domain Model and Application Layer
The second quarter of 2024 saw significant investment activity in several emerging AI application areas, reflecting the expanding scope of AI’s impact across industries:
AI Search
AI Search emerged as a major investment focus in the first 6 months of 2024, with companies developing various specialised search capabilities attracting substantial funding.
- AlphaSense, an AI-powered market intelligence platform, secured $650 million in financing.
- Hebbia, which focuses on enterprise search for internal business data, completed a nearly $100 million Series B round, reaching a valuation of $700–800 million.
- Perplexity AI, a conversational search engine focused on expanding user’s knowledge instead of just on search results, raised $63 million in Series B+ funding, followed by an additional $10–20 million from SoftBank, achieving a valuation of nearly $3 billion.
- Metaso, a perplexity AI equivalent in China, attracted a few millions monthly users within two month after launch.
The intersection of AI and e-commerce search also gained traction.
- Daydream, an AI-powered product discovery platform, received a $50 million seed round.
- Constructor, a B2B e-commerce search company, secured $25 million in Series B funding at a $550 million valuation. These investments highlight the growing importance of AI in enhancing product discovery and personalisation in online retail.
AI Code Generation
AI code generation startups have continued to attract significant capital.
- Cognition, founded just six months prior, raised $175 million at a $2 billion valuation ahead of product commercialisation.
- Magic, despite not yet releasing a product, was reportedly negotiating a $200+ million round at a $1.5 billion valuation.
- Augment completed a $252 million round, reaching a valuation of $977 million. These substantial investments in early-stage AI coding assistants underscore the high expectations for AI’s potential to revolutionise software development.
AI code generation field are mainly divided into two schools:
- One relies on existing advanced AGI Agents; while
- The other consists of end-to-end solutions using specialised models.
The former dominates the market, primarily in the form of code completion plugin products, widely seen among mainstream large language model providers and independent agent products, such as GitHub Copilot, Cursor, August, and Cognition. According to statistics, the accuracy of code generation ranges between 30% and 40%.
The latter, end-to-end software and application generation, is considered the ultimate technical direction for code generation. Representative companies include Poolside, Magic, and AIGCode. They mainly develop more advanced model architectures based on the transformer architecture, pushing the commercialization of large models towards low-error-rate scenarios, though this approach presents higher research and development challenges.
Startups have indeed caught the trend, but it’s not easy to rise to the top in this industry, especially considering that tech giants have also launched their own AI programming assistants. Amazon has AWS CodeWhisperer, Google has Gemini Code Assist, Meta has Code Llama, and GitHub, Microsoft, and OpenAI have all launched the AI code assistant Copilot. Currently, GitHub Copilot has over 1.8 million paid users, with an ARR approaching $200 million.
Major Chinese tech companies have also launched their own AI code assistants: Baidu has Comate, Alibaba has Tongyi Lingma, SenseTime has Code Raccoon, and Huawei has Pangu Coder. However, compared to overseas, there are fewer startups in China, with notable examples being Zhipu AI’s CodeGeeX and Silicon Heart Technology’s aiXcoder.
AI Pharmaceuticals
AI Pharmaceuticals emerged as another hot investment area. Notable funding rounds included Xaira Therapeutics, Diagonal Therapeutics, Karius, and Jitai Pharmaceuticals. This trend reflects growing confidence in AI’s ability to accelerate drug discovery and development processes.
AI Education — Special Mention
In the AI Education sector, while large funding rounds were less common, there was a proliferation of successful applications. Chinese companies like Gauth, Question.AI, and Answer.AI gained significant traction in international markets, particularly in the US. AI-powered learning machines and educational toys also saw increased interest, with companies like iFlytek, Zuoyebang, and Yuanfudao launching products in this space.
Plan Ahead: Strategic Tips for AI Entrepreneurs
It is clear that we’re entering the product-driven stage of the Generative AI boom since the players on the foundational model level are more or less settled (pending some potential mergers/acquisitions in the next 12–18 months). As a result, token and infrastructure costs will be lowered, even if they are free in extreme cases, to attract more products and users. The constraints of computing power, bandwidth, and token throughput speed are likely to be resolved faster in the future.
The opportunities now lie at the vertical application level, which is still in its early days. In order to take advantage of the nascent growing market, AI entrepreneurs should consider the following:
Product:
- Focus on product-driven development: The AI entrepreneurship trend in the United States has shifted from being “technology-driven” to “product-driven”. Entrepreneurs should prioritise developing AI solutions that address specific market needs rather than focusing solely on technological advancements.
- Prioritise finding Product-Market Fit (PMF): As the first wave of consumer-facing AI applications struggle with growth and financing, achieving PMF has become crucial for attracting capital. Entrepreneurs should focus on validating their product’s market fit before seeking significant funding.
- Consider vertical-specific AI models: While general-purpose large language models attract substantial funding, there’s growing interest in vertical-specific AI models. Entrepreneurs may find opportunities to develop AI solutions tailored to specific industries or use cases.
Market Development:
- Focus on sustainable business models: Despite the funding boom, many AI startups face profitability challenges. Entrepreneurs should prioritise developing sustainable business models and clear paths to profitability.
- Build strong user communities: As infrastructure issues for large models are being resolved, understanding user needs and building early user communities will become crucial differentiators. Entrepreneurs should focus on user engagement and community building from the outset.
- Leverage existing ecosystems: Investors are showing interest in AI startups that can integrate with existing technology ecosystems. Entrepreneurs should consider how their solutions can complement or enhance existing platforms or technologies.
Market Realities:
- Prepare for intense competition: With tech giants like Amazon, Google, Microsoft, Huawei, Alibaba, and Tencent launching their own AI product suites, startups face significant competition. Entrepreneurs should differentiate their offerings and target niche markets or specific developer needs.
- Prepare for potential consolidation: The AI industry may be entering a period of consolidation, particularly in the large model space. Entrepreneurs should consider their long-term strategy, including potential partnerships or exit opportunities.
- Explore international markets: Some Chinese AI education companies have found success in the US market. Entrepreneurs should consider international expansion opportunities, particularly for consumer-facing AI applications. (Read more about Chinese AI startup venturing abroad)
- Prepare for a challenging exit market: While funding is abundant, the exit market remains challenging. Entrepreneurs should have realistic expectations about potential exits and focus on building long-term value.
Conclusion: AI is here to stay but be cautious
Geographically, the United States led in AI investments, with U.S. venture funding reaching $55.6 billion in Q2, a 47% increase from Q1. China also saw vigorous AI activity despite geopolitical tensions and some international investors pulling back, with Chinese companies securing 8 of the ten largest deals in Asia.
Despite the enthusiasm for AI, challenges remain. The exit market for startups remains sluggish, with IPO activity still limited. Additionally, concerns about the high costs and uncertain business models of some AI startups persist, as evidenced by reports of unicorns seeking mergers or conducting layoffs.
Looking ahead, industry experts anticipate continued strong interest in AI investments, particularly in areas like foundational models, AI applications, and AI infrastructure. However, investors may become more discerning, focusing on companies with clear paths to profitability and sustainable business models.
By considering these angles, AI entrepreneurs can better position themselves to navigate the dynamic and competitive AI startup landscape, increasing their chances of securing future funding and achieving long-term success.
If you would like to be kept in the loop and notified when we drop new content,
Subscribe to AI Business Asia Newsletter to receive expert analysis, news, and solutions and foster thought-provoking conversations in artificial intelligence to inspire learning and meaningful connections.