A new analysis of business spending reveals a significant divide in the adoption of artificial intelligence. Startups are aggressively integrating AI into their core operations, while larger, established corporations are taking a more measured and cautious approach, according to a report by venture capital firm Andreessen Horowitz (A16z).
The study, conducted in partnership with the financial technology company Mercury, analyzed the spending patterns of over 200,000 businesses between June and August. The findings suggest that emerging companies are building themselves from the ground up with AI, a strategy that contrasts sharply with the incremental AI implementation seen in bigger firms.
Key Takeaways
- A study by A16z and Mercury shows startups are spending heavily on AI, unlike larger companies.
- Horizontal AI platforms for general productivity, like those from OpenAI and Anthropic, are the most popular.
- Most startups use AI to assist employees, but a growing number of tools aim to replace specific, high-cost roles.
- Coding assistance platforms, or "vibe coding," are popular despite some high-profile technical issues.
- Experts raise concerns about a potential AI market bubble, citing massive spending by unprofitable companies.
A Divergence in AI Adoption
The report highlights two distinct approaches to artificial intelligence in the business world. On one side, startups are embedding AI into their foundational structures. On the other, established enterprises appear to be using AI to enhance existing workflows without fundamentally changing their operational models.
"While big companies are getting incremental benefits to existing team structures, at startups we're seeing truly AI-native companies emerge, built around the next generation of software," wrote A16z partners Olivia Moore, Marc Andrusko, and Seema Amble in the report.
This difference suggests that startups view AI not just as a tool, but as a central component of their business strategy. They are leveraging AI from day one, potentially creating new business models that legacy companies may find difficult to replicate.
Dominance of Productivity Tools
The analysis ranked the top 50 AI application companies based on startup spending. The results show a clear preference for tools that offer broad, company-wide benefits.
Horizontal AI platforms, which are designed to improve productivity across various departments, accounted for 60% of the companies on the top 50 list. These tools are not specialized for one specific job but rather serve as general-purpose assistants for a wide range of tasks.
Top Vendors in Startup Spending
According to the A16z and Mercury data, the two most popular AI vendors among startups are OpenAI and Anthropic. Both companies are known for their powerful large language models that power a variety of applications.
The Rise of AI-Assisted Coding
A surprising finding was the popularity of AI-assisted software development platforms, sometimes referred to as "vibe coding." The third most popular vendor on the list was Replit, a company in this category.
Replit's high ranking comes despite a widely reported incident where one of its AI agents was accused of deleting a customer's production database. Nevertheless, its popularity, along with the inclusion of other platforms like Cursor, Lovable, and Emergent, indicates strong demand among startups for AI coding assistants.
A16z's report suggests this trend is more than a fleeting interest, stating that this form of AI-assisted development "is no mere consumer trend."
Assistance Over Replacement for Now
The study also provided insights into how startups are deploying AI in relation to their workforce. The data indicates that most companies are currently using AI to augment human capabilities rather than to eliminate jobs entirely.
When examining role-specific, or vertical, AI applications, the report found a clear trend:
- 12 companies on the list offer tools designed to assist human workers.
- Only 5 companies offer products intended to fully replace human roles.
This suggests that for now, the primary value of AI for many startups is in making their existing teams more efficient and effective.
Targeting High-Cost Professional Roles
While AI as a replacement tool is less common, the roles being targeted for automation are specific and strategic. The five companies focused on replacement offer services in areas that are typically expensive for new businesses. These include:
- Agentic AI for legal services (2 companies)
- AI-powered IT helpdesks (1 company)
- AI software engineers (1 company)
- Automated go-to-market functions (1 company)
The A16z partners predict that this trend of automating high-skill, high-cost positions will grow. They anticipate the emergence of more "end-to-end agentic products" that allow startups to avoid costly, long-term contracts with professionals like lawyers and accountants.
Navigating the AI Hype Cycle
The aggressive AI adoption by startups stands in stark contrast to the caution of larger firms, raising questions about the sustainability of the current market. Many experts point to signs of a potential bubble, fueled by massive investments into companies that have yet to achieve profitability.
For instance, major AI developers like OpenAI and Anthropic are not yet profitable despite their widespread adoption. Furthermore, large-scale infrastructure deals, such as OpenAI's agreement with Oracle for computing power, rely on future growth and investor confidence rather than current revenue.
A Warning from Gartner
Industry analyst firm Gartner has projected that more than 40% of projects involving agentic AI—the type designed for independent decision-making and worker replacement—will be canceled by 2027 due to the technology's immaturity.
This cautious outlook is shared by many large enterprises that have invested in AI but have not yet seen significant returns. The technology is still developing, and its long-term value has not been definitively proven across all industries.
An Uncertain Future
The findings from A16z and Mercury paint a picture of a bifurcated market. Startups are betting their future on being AI-native, believing it will give them a competitive edge over slower-moving incumbents.
This strategy presents two possible outcomes. In one scenario, these nimble, AI-powered startups could out-innovate large corporations, fundamentally disrupting entire industries. In the other, if the predicted AI bubble bursts, these same startups could face significant financial trouble, having tied their survival so closely to an overhyped and unprofitable technology.
The coming years will determine which path the industry takes, revealing whether the startups' all-in approach or the enterprises' cautious waiting game was the correct strategy.