2024 AI Lookback
Key themes shaping the AI startup landscape over the last year
Most people take the holiday season, and specifically the last two weeks of the year, to consider the events that unfolded over the last 12 months. Looking to the recent past is a helpful practice. It allows you to acknowledge the good and the bad, and to identify opportunities for improvement. Many of us use this time of reflection to set goals and establish a focus for the new year.
I did the same: I took note of a few prominent themes that emerged, mostly centered around AI, to see how it might inform the future. Obviously, it was a hot and hyped year for the technology. Some of the trends noted below will have a lasting impact as we continue into 2025.
AI funding easily surpassed all other categories in VC deals
According to Crunchbase, AI-related fundings accounted for $100BN, or one-third of venture deals globally last year. While valuations of traditional SaaS startups have largely normalized, hot AI deals are still getting done in the 100x ARR range
There are even several examples of startups that raised rapid, consecutive rounds and experienced enormous in the span of a single year:
Eon — a next-generation cloud backup platform raised their Seed, Series A, Series B, and Series C rounds in 2024, the first year of founding. The business is reportedly pre-revenue, but by year-end had collected nearly $200M in fresh capital and was anointed with a $1.4BN valuation
xAI – developer of Musk’s AI chatbot Grok, raised their Series A, B, and C rounds in 2024, receiving an 80x uplift in valuation as they ended the year worth $50BN. They raised a total of $12.13BN in 12 months
Mercor – an AI-powered hiring company, raised a $32M Series A in September 2024, and quickly followed with a Series B in December 2024 (valuation undisclosed)
Glean – AI-powered enterprise search platform, raised $460M in 2024 across its Series D and E rounds and closed out the year worth $4.6BN
World Labs – a spatial intelligence company building Large World Models (LWMs), raised close to $300M across a Series A and B round which closed 6 months apart, achieving unicorn status at a $1BN valuation
Cyera – a holistic data storage and security platform, raised $600M across a Series C and D round, doubling in valuation to $3BN by the end of the year
The “Innovator’s Dilemma” is inspiring corporate strategy across the incumbent landscape. Many legacy enterprise SaaS behemoths have implemented their own AI solutions in order to avoid the tsunami of disruption that swells in the distance
Salesforce launched Agentforce – a new product that allows its customers to build custom AI agents, capable of connecting to any enterprise data source to take action across sales, service, marketing, and commerce
Klarna partnered with OpenAI to release AI-powered customer support agents – in a notorious press release Klarna claimed that it’s AI assistants now handled two-thirds of all customer service conversations at the company, handling a workload equivalent to 700 agents
Zendesk announced Zendesk AI – AI-enabled customer support through agents and copilots with features across messaging, support ticket resolution, and analytics to make its existing customer support engine more efficient and to improve CX
Box doubles down on Box AI – in 2023, Box announced they were integrating OpenAI’s ChatGPT API with the Box Content Cloud making it easier for enterprises to search through their files to get results and insights quickly. This past year Box released BoxAI Studio where users can select their preferred LLMs to deploy custom agents, and Box Apps which uses AI-powered metadata extraction to power enterprise and content-specific business process automation
Adobe launched Adobe Sensei and Firefly– integrated into the Adobe product suite, Adobe Sensei is generative AI layered on top of products like Photoshop, Premiere Pro and Lightroom to automate tasks and optimize workflows. Adobe Firefly is a family of generative AI models that use machine learning to seamlessly add images and text when creating new content and media.
Oracle integrated AI workflows into its cloud platform – announced AI-driven workflows as part of the Oracle Fusion Cloud Customer Experience platform in an effort to automate the entire customer service lifecycle and compete as the AI customer service startup race intensifies. The offering includes an automated agent, call and chat summarization, and enhanced search for human support agents
AI has been given agency, and agents are all the rage right now
Agent startups have been exploding across coding, sales, customer support, and marketing. Far from an exhaustive list, some notable examples include:
Coding – Cursor, Cognition (Devin), Codeium, Poolside, Zencoder
Customer Support – Sierra, Decagon, Pylon, Ada, Forethought
Marketing – Jasper, Typeface, Maven AGI, Copy.ai, Writer
Sales – 11x, Artisan, Rox, Nooks
Stripe has even announced ambitious plans to have agents handle payments. The company launched an SDK to enable payments in agentic workflows. It’s still early, but this functionality is powerful. It will allow agents to complete an entire purchasing workflow autonomously, essentially opening up the financial services industry to the power of agents
The fanaticism around agents will likely surge to an even greater level when OpenAI unveils their agent “Operator” this month, that is rumored to be able to autonomously control a computer
What we possess in agent application potential, we lack in agent infrastructure
Most agents show promise for internal use cases. A key unlock going forward will be the ability to take actions externally to complete a task. This future state will likely lead to multi-agent systems where agents will be required to interact with one another. At the moment, attempts to achieve full autonomy have resulted in agents that are trained to navigate the web like regular users would. We will likely see a number of agent infrastructure players emerge proposing interesting ways to connect agents to each other
As with every revolution, there has been some initial backlash to change. Increased AI activity across enterprise channels is driving reversion to more traditional methods
For instance, AI SDRs have led to historically high volumes of cold outbound, which in turn inspired many to revert back to warm and in-person lead generation. As a differentiator, sales leads are emphasizing old school tactics like dinners, conferences, and events to land new business
There is also a noticeable emphasis on human involvement in the AI transition. No one wants to buy a product that they think will replace them one day. Many companies are now taking human-centric approaches to combat this, and their sales pitch presents “human-in-the-loop” as the key to unlocking AI’s potential
These were some of the most notable takeaways I had from 2024. On a high-level, they provide insight into the impacts of AI on the fundraising environment, speed of innovation, increased competition from legacy actors and new entrants, and human behavior amid a transformational market shift. Stay tuned for how I think these trends will continue and change in 2025 and beyond.
Thanks for sharing this