Forget AI: Cloud startups are back (kinda)

Welcome to Runtime! Today: generative AI is having a moment, but there are signs of life for the rest of enterprise tech startups, some enterprise tech vendors want to see more nuance around potential AI regulation discussions, and the latest enterprise moves.

(Was this email forwarded to you? Sign up here to get Runtime each week.)

Long way to go

It's no secret that the easiest way to get funding for a startup in 2024 is to jump on the generative AI bandwagon. According to new data released by Accel this week, generative AI companies will receive 40.3% of the estimated $79.2 billion that will be invested in U.S., European and Israeli startups this year.

Generative AI companies only accounted for 5.1% of funding during the cloud investment frenzy of 2021, when enterprise tech startups raised $116 billion. But while it will probably be some time before we see that level of activity again, investment in non-generative AI startups is finally growing again after a brutal couple of years.

• Accel expects non-generative AI startups to receive $47.3 billion in funding this year, compared to $37.7 billion in 2023 and $46.1 billion in 2020.
• 2023 was a year to forget for non-AI tech startups, after funding fell off a cliff thanks to a big pullback in tech spending and the general sense that everyone kind of lost their minds in 2021 in assuming the pandemic-inspired cloud boom was a new normal.
• Obviously, AI is driving most of the overall growth in startup funding in 2024, accounting for almost all of the 65% growth in overall funding since 2020 according to Accel.
• But those numbers are distorted by the huge rounds that companies like OpenAI, Anthropic, Mistral, and others need to train future generations of generative AI models; out of $56 billion invested in generative AI companies over the last two years, $37 billion has gone to foundation model companies.

Times remain tough for non-AI cloud companies, which are growing much slower than during the years of the pandemic boom and in many cases being forced to think about profitability for the first time. But Accel identified several trends related to generative AI that are spurring investment in fundamental enterprise stack companies, which could lead to more sustainable funding growth over the next several years.

• The rush to deploy generative AI applications has forced CISOs to think much harder about their data security strategy, as those apps open up new attack surfaces and generate ever-increasing volumes of sensitive corporate data.
• Managing and storing all the data associated with building and running generative AI apps requires many companies to overhaul their data stacks, one reason why the rise of open file formats is creating new opportunities for data companies.
• Generative AI is being used by more than half the companies studied by Accel to develop software, which could require enterprises to invest more in design and deployment tools as they produce more software on their own rather than buying it from someone else.

Still, there's no question that generative AI technology is souring the outlook for both enterprise SaaS startups and large vendors. Before generative AI, the easiest way to get funding was to build a targeted SaaS company around a specific business problem, such as payroll or procurement, and those days may never return.

• Crunchbase expects SaaS and enterprise software funding in 2024 "to come in far below last year’s $17.4 billion annual tally — which was itself the lowest total in years."
• Big SaaS companies like Salesforce, HubSpot, and ServiceNow are pivoting quickly to agentic AI in hopes of sparking growth, making it harder for startups to break through.
• "AI is sucking the air out of the room,” Accel's Philippe Botteri told CNBC, and while that might leave a little oxygen for startups working on infrastructure or platform tools, the third leg of the as-a-service stool is having trouble breathing.

Integrators; mount up

Enterprise tech companies rebuilding their product strategies around generative AI models continue to worry about how the government might regulate future AI development. A group of enterprise tech vendors sent a letter to Congress Thursday arguing they should think more carefully about how proposed regulations would affect "AI integrators."

The letter, signed by representatives from Alteryx, Salesforce, Twilio, Box, Kyndryl, and Peraton, called on the leaders of the House Task Force on Artificial Intelligence to weigh the effects of any proposed regulation on "companies that retrain, modify, fine-tune, combine, and integrate AI models into products and services." The concern is that much of the current debate over AI regulation distinguishes between "developers," companies that make AI models, and "deployers," companies that use those models as part of their product or business strategy.

However, there's a middle group involving companies that "incorporate these pre-existing models into larger systems so they can be used for specific business needs," Salesforce's Hugh Gamble told CIO Dive. Instead of being held accountable for AI safety and transparency, Gamble argued that any legislation targeting integrators focus on "proper testing and responsible use."

Enterprise moves

Dane Stuckey is the new CISO of OpenAI, joining the company after serving in a similar role at Palantir.

Thierry Delprat and Rohan Vaidyanathan are the new vice presidents of engineering and product, respectively, for content intelligence at Hyland.

The Runtime roundup

AWS jumped on the nuclear SMR train, announcing plans to invest $500 million in small nuclear reactors in Virginia, home to a massive complex of AWS data centers, as well as its home state of Washington.

Meanwhile, AWS won't have Nvidia's Blackwell GPU up and running in production cloud servers until early next year, but CEO Matt Garman told Bloomberg it would "be the first to deploy computers based on the new chip in significant quantities."

Permia completed its acquisition of Squarespace for $7.2 billion, as private equity continues to snap up enterprise software companies.

Speaking of which, Zuora was acquired by Silver Lake and Singapore's sovereign wealth fund for $1.7 billion.

Thanks for reading — see you Saturday!