While cash is flowing into the AI startup space, alarm bells are ringing over the long-term viability of many up-and-coming firms.
A new report from Silicon Valley Bank (SVB) shows venture capital investment in AI firms has soared in recent years. Analysis from the bank found roughly 40% of all investment raised by funds last year came from those that “list AI as a focus”.
This marks a sharp increase compared to 2021, in which just 10% of venture capital investment focused on AI startups specifically. Indeed, AI companies now make up 45% of US venture capital investment in enterprise software, again marking a significant jump from 9% in 2022.
A key factor in this sharp rise in funding came from ‘megadeals’, according to SVB, which in this case it defines as raises in excess of $100 million.
There were 107 megadeals across Q1 2025, the most since the second quarter of 2022 at the height of the post-pandemic goldrush. Major industry players such as OpenAI and Anthropic were found to be “leading the charge” in this regard.
But despite the bullish outlook among VC firms, cracks could begin to show across the industry. The report noted that this sharpened focus on the technology means startups in other business verticals are finding it increasingly difficult to raise cash.
“Exclude AI investment and the story changes,” the SVB report stated. “There is no meaningful uptick for companies not leveraging AI, with investment from this group essentially flat for the last year.”
What this means is that companies failing to lean into the generative AI trend are missing out on a potential goldmine of investment.
But there’s more to the tech sector than simply AI solutions and software, and the venture capital obsession with AI could risk harming the broader ecosystem, according to Sam Hields, partner at early-stage tech VC, OpenOcean.
"The Silicon Valley Bank (SVB) data confirms what we already knew: AI is the gravitational force in venture capital,” he said. “As capital gets more and more concentrated into AI, we risk sucking the oxygen out of the room for other breakthrough technologies.”
AI ‘zombiecorns’ are becoming a serious problem
While the diversity of funding warrants serious concern, SVB also noted that cracks are beginning to show with regard to the long-term viability of some startups in the space.
The report noted that many startups are woefully underprepared post-funding, largely due to poor revenue growth and unit economics. SVB describes these companies as ‘zombiecorns’.
Infrastructure costs, dwindling runway, and sluggish revenue growth due to an increasingly competitive marketplace represent a confluence of challenging conditions for these companies.
“While new unicorn creation has been impressive in recent years, unicorn exits are a different story,” the report noted.
“The enterprise unicorn herd stands above 300, with few exiting, closing their door, or taking a down round below a $1B post-money valuation.
“With the growth of the herd, so too comes growing demand for liquidity. But with the IPO window effectively shut, and large-scale M&A being tougher to execute with the overhanging macro climate, many run the risk of ending up in no man's land.”
Tom Glason, CEO and co-founder at ScaleWise, said the SVB report highlights a “harsh truth” about the generative AI trend.
“The AI boom has fueled a wave of overfunded startups that look healthy on the surface, but are commercially hollow underneath,” he said. “These so-called ‘zombiecorns’ raise huge rounds but fail to build sustainable revenue or viable unit economics.”
This has become a common recurring theme in the market, Glason suggested, with founders “confusing capital raised with genuine traction”.
Ultimately, capital efficiency and go-to-market execution still largely determine whether a company graduates to IPO or fades out, he added.
Hields echoed Glason’s comments, adding that the hype surrounding the technology has resulted in a trend in which companies rapidly pivot to become an ‘AI company’ which ends up being nothing more than a shiny veneer on top of an existing tech solutions stack.
“Today, folding an LLM into your product is enough to claim an ‘AI badge’. That’s perfectly natural - and, in many cases, it’s trivial to implement. But it won’t deliver durable returns,” he said.
“SVB is right to warn us about the rise of ‘zombiecorns’. If we avoid the temptation to over-index into dead-end, ‘wrapper-ware’ startups and push against the crowding out effect on deeper R&D, we have a chance to channel enthusiasm for AI into solving real, structural problems. That’s where the real alpha comes in.”
MORE FROM ITPRO
Why IT leaders look for gamers when hiring talent
GitHub outage blamed on misconfigured policy as firm pledges resilience improvements
Poor business context is scuppering enterprise AI adoption – here’s why that matters
Manchester University: from AI initiators to AI integrators
‘Real-time data remains the voice that every leader must hear’: This one engineering role could be the key to building a truly data-driven enterprise
This new technique could improve AI output accuracy by 80% – and tackle hallucinations once and for all
Reports: UK could consider regulation amidst growing 'rogue AI' concerns
UK firms are automating roles, but nowhere near ready to outright replace them
Two-thirds of workers are so fed up with ‘AI slop’ that they ‘feel nostalgic for pre-AI work’
How AI certification can help employees climb the career ladder
