2023 was certainly not a dull year. The startup-technological scene was dominated by generative AI. As anticipated, ideas for commercialising tools from OpenAI began to emerge, and giants like Google, Microsoft (a major shareholder in OpenAI), and Amazon joined the race in this category. VC investors allocated over $36 billion for generative AI investments this year, more than twice the amount from the previous year. Everyone wanted a piece of the AI pie, leading to many businesses engaging in AI-washing – using catchy marketing slogans related to AI tools that lacked substance in reality.
We witnessed a continuation of layoffs in technology companies (in 2022, over 164,000 employees were affected, and just when we thought the worst was over, another 261,000 people lost their jobs in 2023). This was tied to the companies’ desire to improve profitability, stemming from the overhiring in 2020-21, and subsequent efforts to enhance results. The dynamic development of AI tools and solutions such as low code and no-code also played a role, improving efficiency and reducing the demand for programmers.
In 2023, cryptocurrencies experienced a resurgence – Bitcoin and Ethereum provided returns of over 155% and 90%, respectively. At one point, Bitcoin reached a value of $44,500 (though it did not surpass its all-time high). In the first week of January, a U.S. regulator approved the first ETF (Exchange Traded Fund) based on Bitcoin, marking a milestone for both the currency and the entire cryptocurrency market.
As we move into the first month of 2024, it seems to promise no less excitement than its predecessor. What events can we expect in the coming months?
Venture capital investments continue to decline
As predicted, 2023 proved challenging for startups seeking funding. Persistent high-interest rates made investors less inclined to allocate capital to alternative asset classes. According to Pitchbook, VC investments in North America for the first nine months of 2023 amounted to $126 billion, only half of the total recorded in 2022.
Globally, the narrative was similar. While European startups secured over $116 billion in funding throughout 2022, the first three quarters of 2023 saw just over $47 billion. Asian startups gathered only $65.3 billion in the first nine months of the previous year, compared to $140.6 billion in the entire 2022.
What awaits the VC market in 2024? Following inflation stabilization, central banks are likely to ease monetary policies, gradually reversing the trend. However, securing funding will still be more challenging than in previous years, and many startups with inefficient business models may not survive this period. Some projects may have to accept lower valuations than in previous rounds.
There’s also a chance of witnessing more exits from investments. With a substantial amount of capital locked in private companies, investors may feel pressured to release it, even at the cost of lower valuations.
The pressure to release capital will also lead to the development of the secondary VC market. Platforms facilitating the trading of shares in private companies (e.g., EquityZen, SeedDesk) already exist. Expect an increase in demand for this service due to the inability to exit investments caused by low company valuations. The democratization of investing in private companies, once reserved for a narrow group of specialized institutional investors, will likely gain many supporters.
Continued Growth in Employment in Climate Tech
Technology companies focused on climate protection were among the few that continued to hire when others were downsizing. There is a chance that this trend will persist in 2024. Social pressure for climate protection and the readiness for commercialization of numerous projects that have received funding in recent years will influence this. These startups will not only need technological employees but also salespeople, marketers, and service personnel, such as technicians for wind turbines or solar panels.
The ongoing AI revolution may also have a significant impact on increasing demand for renewable energy projects. As calculated by Dutch scientist Alex de Vries, by 2027, global energy consumption related to AI could exceed 134 TWh per year – roughly equivalent to the annual consumption of Argentina, Norway, or Iraq. Energy providers are already feeling the strain. Energy demand will only continue to grow, along with the demand for renewable energy.
Artificial Intelligence in the Business Mainstream?
In the early stages of generative artificial intelligence development, it served primarily as a support tool for programmers (enabling features like GitHub Copilot). In 2022, tools such as ChatGPT and DALL-E 2 emerged, allowing individuals without specialized knowledge and experience to instantly generate text and images. In 2023, tech giants reaped the most benefits from artificial intelligence – the stock index of Alphabet, Amazon, Apple, Meta, Microsoft, and Nvidia increased by almost 80%. How will this story unfold further?
In 2024, companies outside the technology sector may join the AI beneficiaries, implementing artificial intelligence to reduce costs and increase productivity. The past few months have been dedicated to developing practical applications of these models, and the next phase involves further investments and implementation. These tools are already used for creating various documents and meeting notes. In the coming months, their applications are expected to expand into more strategic areas, such as analytics and forecasting market trends.
Regulations in Artificial Intelligence
A year ago, we wrote about the likelihood of soon witnessing the first attempts at legal regulations concerning issues related to generative AI. It is increasingly apparent that the need for such regulations is becoming urgent.
An example is the lawsuit filed by publishers of The New York Times against OpenAI at the end of 2023. The creators of ChatGPT were accused of unlawfully using millions of articles published in the newspaper to improve the model, allowing readers to access this content without paying for it – consequently causing a loss of subscription and advertising revenue, measurable financial losses for the publisher.
This is the first such lawsuit filed by a major institution, so the tech world will closely follow its developments. In addition to potential serious consequences for OpenAI, the entire case may expedite efforts to regulate artificial intelligence. The European Union and its regulation on artificial intelligence, known as the AI Act (AIA), seem to be closest to this goal. In December, after a three-day negotiation marathon, representatives of the Council Presidency and the European Parliament reached a preliminary agreement on the draft regulations. It can be expected that they will come into effect by 2025.
Simultaneously, we observe the development of a new category of tools for detecting the use of AI, i.e., identifying whether a particular material was created using artificial intelligence. This technology can find broad applications not only in media but also in education, among other fields. The first solutions of this kind are already available on the market (e.g., Copy Leaks, Undetectable AI, Scribbr), but the development of this category is likely far from over.
Democratization of Artificial Intelligence, Multimodal Models, “Smaller” and “Private” LLM
Everything indicates that in 2024, more open-source models will be released, especially by large technology companies. According to CB Insights, based on GitHub data, a record number of AI projects will be built on open-source models in 2024. A noticeable price race is already emerging among OpenAI, Mistral, Google, and other entities supporting open-source models – one of its effects will likely be increased accessibility and commercialization possibilities of AI for a growing number of entities.
This will, among other things, facilitate access to this technology for a broader group of users: it is predicted that by 2026, 80% of businesses will be using generative artificial intelligence models. For comparison, at the beginning of 2023, this figure was less than 5%.
There will also be an increasing number of multimodal models on the market, combining text with other types of data, such as images, videos, audio, and other sensory data. Google placed a strong emphasis on multimodality with the release of the Gemini Ultra model in December, and many other LLM creators are following suit. The popularization of multimodal models will unlock a whole list of new potential commercial applications.
We might also witness the emergence of smaller, cheaper models that, in some cases, could deliver better results than LLM. Some companies will also turn towards the adaptation of private AI models, fueled by their own data and facilitating security measures. The security of both large and smaller models will also be a hot topic in the coming months.
Investments in AI
Companies involved in generative AI will be seeking ways to increase profitability and scalability. Consequently, VC investments will flow into projects aimed at enhancing model efficiency and utilizing new AI computing hardware.
The opportunities presented by these models and their wide range of applications will lead to the emergence of many new projects in this area. Those complementing the functioning of AI models and enabling profit generation from the AI revolution are likely to attract significant investor interest.
Given the limited number of programmers specializing in AI solutions, particularly in comparison to the growing demand in the market, solutions in the no-code space will be particularly interesting for investors.
It can also be expected that, much like during a gold rush where shovel sellers profited, a good place to find new, profitable ventures will be in the services sector for the entire artificial intelligence industry.
Limited Access to Components
The explosion in the popularity of generative artificial intelligence has significantly increased the demand for GPU systems. In January 2024, H100 processors were already being sold with a margin exceeding 95%. It can be expected that in 2024, we will reach a point where supply will not keep up with demand. This phenomenon will affect all major recipients, including companies like Meta (which, according to reports, plans to purchase 350,000 Nvidia H100 processors), OpenAI, and Tesla. These hardware limitations may catalyze the emergence of new projects aimed at increasing the performance of available units and developing alternatives (such as cheaper or more environmentally friendly ones). On the other hand, it may lead to a more selective and focused development of artificial intelligence and its applications, concentrating on the most promising projects. Therefore, a more pragmatic approach to AI can be expected this year.
Talent Development in the Field of Prompt Engineering
With the democratization of access to generative artificial intelligence, the importance of effective utilization of this technology is growing. It is projected that in 2024, as many as 80% of companies plan to invest in the recruitment and development of specialists in the field of prompt engineering, which involves designing, testing, and optimizing instructions or queries for language models.
Companies investing in developing these skills within their teams believe that through a more intelligent and personalized approach to using generative technology, they can gain a competitive advantage.
Do you agree with our predictions? Is there something that we completely missed? Let us know – Aron and Sylwia are waiting for your messages on LinkedIn.