Behind the Software Panic: Why Goldman Sachs Thinks the AI Sell-Off Is a Premium Discount

Wall Street has a habit of shooting first and asking questions later, a tendency vividly illustrated by the brutal capital punishment inflicted on software-as-a-service (SaaS) stocks. Dubbed by some as the "Saaspocalypse," the panic wiped out billions in market value as investors panicked over the existential threat of generative artificial intelligence eating into traditional software revenue models. However, the analysts at Goldman Sachs Research argue that this sweeping, indiscriminate dump is entirely overdone, creating a rare window to buy elite growth companies at an aggressive discount.

The panic peaked when advanced automated developer tools and coding agents emerged, driving a widespread narrative that standard per-user licensing models would rapidly become obsolete. Goldman Sachs CEO David Solomon pushed back against this bleak consensus, noting that the market's response was far too broad. While there will certainly be winners and losers as enterprise operations transform, premium companies with sticky user bases and deeply integrated proprietary data are uniquely positioned to pivot and thrive rather than facing destruction.

Instead of collapsing under the weight of AI disruption, several prominent tech incumbents are aggressively turning the technology into a fresh growth engine. Financial institutions are pointing to businesses like Figma and Atlassian as prime examples of resilient enterprises that have suffered massive, unjustified valuation compression. Far from being replaced by chatbots, these platforms are seeing robust customer retention alongside rapid early adoption of their own paid AI upgrades.

What Most Reports Miss: The Illusion of Obsolescence

Look past the apocalyptic headlines, and it becomes clear that the bear case for enterprise software relies on a fundamental misunderstanding of corporate tech stacks. The market panicked on the assumption that because an AI agent can write code or generate templates, companies will immediately fire employees and slash their corporate software subscriptions. But software platforms provide the essential system of record, compliance framework, and collaborative environment that AI tools require to function effectively within an enterprise.

Rather than acting as software killers, generative tools are operating as powerful product features that justify entirely new monetization layers. Design heavyweight Figma saw its valuation cut in half during the peak of the panic, yet the company managed to post a massive 41% year-over-year revenue jump to $1.1 billion. By acquiring AI-orchestration firm Weavy and introducing consumption-based AI credits on top of its core subscription model, Figma is proving that artificial intelligence is expanding its addressable market rather than shrinking it.

A similar dynamic is playing out with project management giant Atlassian, whose shares tumbled roughly 60% during the market rout over fears that seat-based licensing would collapse. The hard numbers tell a radically different story, with the company delivering a 32% year-over-year revenue surge to $1.8 billion in its third fiscal quarter. More importantly, its customers are actively expanding their user bases while its internal AI credit usage spiked by 20% month over month.

Ultimately, this cycle mimics historical tech panics where a revolutionary infrastructure shift is initially perceived as a threat to the application layer. When cloud computing took off, skeptics predicted the absolute death of legacy systems, yet corporate mainframes and hybridized databases adapted and retained massive value. The current valuation contraction is a classic case of sentiment divorcing itself from core fundamentals, giving disciplined investors an ideal setup to accumulate high-margin software operators before the market realizes they are actually AI beneficiaries.

Reading Between the Lines: The Structural Paradox of AI Monetization

The core tension in Wall Street’s current software narrative lies in a massive contradiction regarding corporate spending power. On one hand, analysts are cheering the record-breaking capital expenditures poured into Nvidia chips and data center infrastructure. On the other hand, the market is punishing software companies under the assumption that enterprise technology budgets are a zero-sum game, expecting corporations to permanently starve their application layer to pay for raw compute. This logic falls apart over a longer horizon because expensive infrastructure is utterly useless to a Fortune 500 company without the user-facing software interfaces required to deploy it productively.

A deeper look into enterprise mechanics reveals that the transition from seat-based pricing to consumption-based AI pricing is not the death sentence skeptics claim. Critics argue that if AI tools make human workers twice as productive, companies will simply buy half as many software licenses. However, history shows that when the cost of a digital asset drops, total consumption skyrockets exponentially. As AI agents lower the barrier to software development and data analysis, the sheer volume of code, assets, and workflows generated will explode, forcing enterprises to rely even more heavily on platforms like Atlassian and Figma to manage the resulting operational chaos.

There is, however, a legitimate reason for measured skepticism that goes beyond the panic over seat counts. The real risk for SaaS companies is not obsolescence, but rather a temporary margin squeeze driven by the high cost of running AI models. Enterprise customers expect AI features to work flawlessly, but they are notoriously resistant to massive price hikes. Software vendors are currently forced to absorb the substantial costs of cloud compute and API calls while they experiment with pricing models, creating a tricky financial bridge where revenue growth might outpace profitability for several quarters.

Ultimately, the Goldman Sachs thesis hinges on time horizons. In the short term, market sentiment will likely remain volatile as legacy software companies figure out how to accurately meter and bill for autonomous AI interactions. But treating this awkward evolutionary phase as a terminal decline ignores the massive moat enjoyed by incumbent platforms. In the enterprise tech ecosystem, owning the user workflow and the underlying proprietary data is everything, and the newcomers attempting to build AI-native alternatives from scratch face an incredibly steep uphill battle to unseat them.

Wall Street has successfully predicted twelve of the last two technology extinctions, proving once again that nothing panics an investor quite like a software company forced to change its billing department's spreadsheets.

Artūras Malašauskas is an AI Systems Integrator with 20+ years of production-grade web engineering experience. He has designed, shipped, and scaled enterprise Python/PHP systems for logistics, SaaS, and public-sector clients. For the past year, he has focused exclusively on AI integrations: deploying open-source LLMs, building generative media pipelines (image, audio, video), and engineering multi-agent workflows for real production environments. His standard: reproducibility, security, cost-efficient inference—no vaporware. He documents and evaluates emerging AI tooling, separating verified capabilities from marketing noise. Technical editor at: muza-ai.eu, ai-verslas.lt, ai-naujinos.lt Connect on LinkedIn