GCV data shows the AI boom is at the expense of enterprise software startups

AI mania is everywhere you look. The largest frontier model developers, OpenAI, Anthropic and Mistral, are all reportedly in talks to secure more multi-billion dollar funding rounds. Stories of superstar salary offerings to lure away top researchers keep surfacing in the press. New startups are raising billions merely on the strength of their founding teams, even before they deliver a product.

Corporate VC investment in the sector is also rampant. In GCV’s data roundup stories it has become common to see gargantuan spikes in monthly deal value from mega funding rounds for AI companies. Since ChatGPT launched in 2022, startups working on AI development, or building tools that heavily rely on it, have proliferated.

What is often overlooked is the effect this has on other sectors. If we zoom out and take a view of GCV’s funding round data since 2023, it appears that the growth of AI investment has come at the expense of enterprise software deals. As corporates turn their gaze towards an AI-powered future, deal flow has risen steadily, while other business software startups have been increasingly spurned.

Is AI eating software?

First, some caveats. GCV sorts startup deals into categories based on a judgement call about what the product is, which is not always straightforward to determine. There can be some overlap between AI and enterprise software products, and vague marketing descriptions can cloud the picture further. But generally, to be classified as an AI startup, the company will be making AI technology or infrastructure, software designed to help LLMs or agentic AI to function, or a product where the core feature is a generative AI tool. A regular business software offering with a rudimentary chatbot thrown in would not count. Overall, while there will be some marginal cases in the data, the same approach has always been followed when sorting deals, so any pattern will be fairly robust.

The vast majority of corporate VCs in both categories have made one-off investments. These might be an effective marker for each sector’s overall popularity, but they do not in themselves prove that deals for standard software startups are falling because of the AI boom.

But looking at the biggest investors in both sectors gives a far stronger indication.

Alphabet, Salesforce and Microsoft were the top three enterprise software investors. And yet throughout the period, their investments in the sector were, at best, relatively static. Salesforce and Alphabet consistently backed more AI startups than enterprise software outfits in each half-year period. For Microsoft, the trend was even more striking. In the first half of 2023, M12, its VC arm, invested considerably more in enterprise software startups than AI, but then by H1 2024 the position had flipped. In the most recent period, H1 2025, our data shows the unit made no software investments.

This mirrors the corporate parent’s wholehearted embrace of AI in its growth plans. Microsoft has spent enormous sums on data centres, has embedded the technology in its products and was an early investor in OpenAI. At its recent earnings call, the company’s CEO promised data centre spending of $120bn over the next four quarters.

It’s a similar story at Alphabet, which has invested in OpenAI’s rival Anthropic, and is also putting aside large sums for data centre investments. Alphabet has rolled out AI products across its business, including the AI overview feature on Google which summarises search results. Salesforce, meanwhile, has launched an agentic AI platform for users to create their own agents, and has similarly embedded AI in its internal business processes, according to the CEO.

Overtaking patterns like that of Microsoft occurred in the investments from some of the other largest corporate software investors, including Samsung, Cisco, SoftBank and Saudi Aramco. This can be seen in the graphs below.

Ignoring software completely

Some deep-pocketed corporate investors have invested heavily in AI while making few or no investments in enterprise software startups over the period. Six of these were among the top 15 AI investors, including the tech giants Nvidia, AMD, Amazon and Alibaba.

As the inventor and host of AWS, Amazon did more than any other company to birth the cloud computing revolution. This has been a great spur to innovative enterprise software startups like Slack and Stripe, which both scaled on the platform. But as a corporate investor, Amazon’s focus is on an entirely different direction. Along with Meta, Microsoft and Alphabet, the company is focusing relentlessly on AI, driving ahead to build out massive data centres and ploughing billions into OpenAI’s rival Anthropic. Enterprise software developers are nowhere to be seen in the startup investment data. Instead, it has focused on AI startups, such as Luma, an AI video generator, and Algorized, which makes people-sensing technology for visual models.

And because the generative AI boom has allowed US chipmakers like Nvidia and AMD to stockpile mountains of cash, it is no surprise that this has contributed to a burst of strategic AI investment post-2022, with no comparable interest in enterprise software.

Nvidia’s 39 AI investments include in the open-source model developer Together, and the spatial intelligence startup World Labs. By contrast, it only made four investments into software startups over the same period. Meanwhile, AMD invested in Thinking Machines Lab, one of the apparently productless new AI companies that raised $2bn in July. It made no software investments.

Changing gears

Enterprise software startups have been a stalwart of VC investment for a reason. The number of use cases feels unlimited, they scale easily, and the subscription model gives a steady, reliable source of revenue. If corporate VCs are betting on AI startups moving into its place, it is far from clear whether this will be a smooth replacement.

The most obvious difference is the cost involved. Making AI products is rarely cheap, largely due to the cost of processing data and running computations. At the most sophisticated end of the sector, the demand for capital goes beyond what any software startup has ever asked for. The mega rounds raised by frontier labs like OpenAI are so distortive that the average corporate-backed funding round value in our data for AI startups, where disclosed, is $213m. The average enterprise software round is $27.8m. Even if we strip out the rounds that raised over a billion, the software average is unchanged and the AI average is still $51.5m – nearly double.

Then there is the question of revenue generation. An enterprise software provider simply charges for the number of seats on a subscription package. But given how nascent the AI startup sector is, there is not yet an industry-standard payment model that has been proven to work.

If what the data seems to be showing is true, and the growth of AI enthusiasm is drawing funding away from enterprise software startups and into a high-cost sector with shaky revenue, then today’s mania could give way to bitter disappointment.

But to end on an optimistic note, GCV’s exits data for the first two quarters show that AI startups may already have the edge over their enterprise software counterparts when it comes to being acquired. Maybe this is an early sign that the shift will be justified.