The End of the Seat Based Era
The software sector is screaming. Goldman Sachs is listening. The investment bank just launched a custom basket of stocks designed to survive the AI onslaught. This is not a vote of confidence in the industry. It is a triage list. For three years, investors treated every SaaS company like an AI winner. That delusion is dead. Today, the market recognizes that artificial intelligence is a cannibal. It eats code. It eats workflows. Most importantly, it eats the per-seat licensing model that built Silicon Valley.
The new Goldman basket targets companies with deep moats. These are firms whose value resides in proprietary data rather than simple interface logic. Per a recent report from Bloomberg Markets, the divergence between AI-enabled infrastructure and legacy application software has reached a record wide gap. The message from 200 West Street is clear. If your software can be replaced by a prompt, your valuation is headed to zero.
The Architecture of Disruption
Software is becoming a commodity. Large Language Models now generate functional code in seconds. This reality has shattered the traditional barriers to entry. In the last 48 hours, the Nasdaq Software Index saw a 3.4 percent drawdown as rumors swirled regarding a new autonomous agent framework from OpenAI. This framework reportedly automates mid-tier enterprise resource planning tasks. These are the exact tasks that companies like Salesforce and Workday have charged billions to manage.
Goldman’s strategy involves a pivot toward vertical software. These are companies that own the specific data of a niche industry. Think of specialized medical records or complex construction logistics. These datasets are not easily scraped by general purpose models. They require the context that only a decade of integration can provide. The bank is betting that these “data-heavy” firms will be the only ones left standing when the dust settles in 2026.
Visualizing the Software Valuation Gap
The following chart illustrates the performance of Goldman’s AI-Resilient Basket versus the broader Software-as-a-Service (SaaS) index over the last twelve months. The data shows a clear decoupling that began in late 2025.
The Valuation Trap
Investors are currently fleeing companies with high price-to-sales ratios. The “Rule of 40” used to be the gold standard for software health. Now, it is a lagging indicator. A company can grow at 40 percent while its core product is being hollowed out by open-source alternatives. Goldman’s analysts are looking at a new metric: the Cost of Replacement. If a developer can build a functional equivalent of your software using an AI agent for less than $50,000, your $10 billion market cap is a fantasy.
The table below highlights the current market multiples as of February 13. Notice the compression in the legacy sector compared to the resilience in infrastructure and specialized data firms.
| Sector Segment | Forward P/E (Feb 2025) | Forward P/E (Feb 2026) | YoY Change |
|---|---|---|---|
| AI-Resilient (Goldman Basket) | 32.4x | 38.1x | +17.6% |
| General Application SaaS | 24.8x | 14.2x | -42.7% |
| Cloud Infrastructure | 41.2x | 44.5x | +8.0% |
| Cybersecurity | 35.5x | 37.2x | +4.8% |
Institutional capital is rotating. It is moving away from the “middle of the stack.” The middle of the stack is where humans used to sit to move data from one window to another. AI does this for free. According to data from Reuters Finance, venture capital funding for generic SaaS startups has dropped 60 percent year over year. The smart money is terrified of the “wrapper” problem. If your business is just a pretty UI wrapped around an API, you do not have a business. You have a temporary feature.
The Cybersecurity Exception
Not all software is doomed. Cybersecurity remains a fortress. This is because AI increases the threat surface. More code means more bugs. More agents mean more unauthorized access points. Goldman’s basket heavily weights firms that provide automated defense. In this sub-sector, AI is a tailwind. The complexity of defending an enterprise in 2026 requires the very tools that are disrupting other industries. This is the paradox of the current market. The fire is burning the house down, but it is also making the fire extinguisher business incredibly profitable.
The regulatory environment is also shifting. The SEC has recently increased scrutiny on AI disclosures. Companies can no longer simply claim they are “integrating AI” to boost their stock price. They must now prove that AI is not cannibalizing their existing revenue streams. This transparency is forcing a painful revaluation. It is separating the innovators from the pretenders.
We are watching the total reconfiguration of the digital economy. The next milestone is the March 15 earnings deadline for the major software players. Watch the Net Retention Rate (NRR) closely. If NRR begins to slip despite high usage, it means customers are finding cheaper, agent-based ways to do the same work. The data point to watch is the 0.85 correlation between AI agent adoption and SaaS churn. If that holds, the purge has only just begun.
