OpenAI’s launch of GPT-5 on Thursday sparked a sharp sell-off in the software sector. The new AI model immediately raised questions about the future of traditional software.
Analysts described GPT-5 as “impressive” and a major step forward in transforming technology, especially enterprise software. Still, they noted the big question remains: does GPT-5 mark the beginning of the end for software?
Core Value Drivers of Enterprise Software
Brokerage firm D.A. Davidson identified three main value drivers for enterprise software:
-
Distribution – Large vendors have mastered sales processes and know how to keep customers engaged after purchase.
-
Integrations – Connecting software to other systems and data sources often requires customization, permissions, and middleware.
-
Workflows – Business logic differs from company to company, meaning implementations must be tailored to each organization’s needs.
Changing Pricing Models
While many software tools are still priced per seat, analysts argue the real value is in how many roles the software replaces or how indispensable it becomes over time. The growing trend toward usage-based pricing reflects this shift.
The Role of AI Agents
According to the report led by Gil Luria, AI agents could eventually take over tasks from employees, changing how software is valued. However, the team warned that AI works best with unstructured data, while most enterprise data is structured and still better handled by traditional computing. For CTOs, the challenge will be managing both worlds while addressing issues like data access, governance, and compliance.
Infrastructure vs. Applications
Analysts also distinguished between infrastructure software and application software. They expect infrastructure — which organizes, secures, and manages code and data — to grow in importance as AI adoption accelerates. Meanwhile, the role of application software remains uncertain.
Customer hesitation about long-term software commitments remains another challenge for the sector.
Investment Outlook
Despite risks, D.A. Davidson said many software companies are well-positioned for the AI revolution. They believe firms that execute well could deliver strong returns for investors. Current top picks include Microsoft (MSFT), Snowflake (SNOW), Datadog (DDOG), and Jfrog (FROG).
The team also noted that the biggest risk from AI’s rapid growth is not to software itself, but “to society as a whole.”
GPT-5’s Technical Leap
The launch of GPT-5 comes almost three years after ChatGPT first went public. Unlike GPT-4, which relied heavily on more compute and data, GPT-5 introduces “test-time compute.”
This feature allows the AI to allocate more processing power to complex questions, improving reasoning and problem-solving. It provides an alternative path to smarter AI without depending only on larger datasets.
GPT-4’s growth had stalled due to limited new human-generated text and the risk of hardware failures during long training runs. With test-time compute, GPT-5 takes a different approach to scaling intelligence.
