Henry Blodget on the Software Selloff Hysteria and the Problem for OpenAI

Henry Blodget, the veteran technology analyst whose early internet calls remain industry reference points, has published a contrarian analysis arguing that the 2026 software sector selloff — driven by fears that AI will replace traditional enterprise software — fundamentally misreads the dynamics of enterprise technology adoption. Simultaneously, he identifies OpenAI's core strategic weakness: despite dominant AI capabilities, the company faces a "distribution problem" where enterprises prefer AI embedded within existing software rather than accessed through a standalone platform.

For finance professionals watching their portfolio companies' SaaS valuations decline 15-30%, Blodget's framework provides a lens for distinguishing genuine AI disruption risk from market hysteria — and for understanding why OpenAI's $100+ billion valuation may be more fragile than it appears.

The Selloff Narrative (and Why It's Wrong)

The bear case driving the software selloff:

The narrative: "AI will make enterprise software obsolete. Companies won't need Salesforce when they can ask ChatGPT to manage customer relationships. They won't need SAP when AI agents handle procurement and accounting autonomously. Every SaaS company is dead."

Market reaction: Salesforce -22%, Adobe -28%, ServiceNow -18%, Workday -25%, Intuit -15% from Q4 2025 peaks. The sector-wide decline represents $800 billion in erased market capitalization.

Blodget's counter-thesis: The narrative confuses two distinct phenomena:

1. AI improves software (bullish for incumbents). SaaS companies embed AI into their products, making them more valuable and justifying price increases. Salesforce + Einstein AI is better Salesforce, not dead Salesforce. Intuit + AI is better TurboTax, not no TurboTax.
2. AI replaces software (bearish for incumbents). Standalone AI tools make entire software categories unnecessary. This is happening in narrow categories (basic content creation, simple data analysis) but NOT in complex enterprise workflows (CRM, ERP, HCM, financial planning).

The distinction matters because 90% of enterprise software revenue comes from complex workflow management where AI augments rather than replaces. The 10% in "simple tool" categories (basic productivity, templated content, straightforward analysis) faces genuine disruption — but this doesn't justify a 20-30% drawdown across the entire sector.

The OpenAI Problem

Blodget's more provocative thesis: OpenAI's strategic position is weaker than its technology suggests.

The distribution moat problem. Enterprise customers want AI capabilities delivered through tools they already use. A CFO doesn't want a separate AI platform — they want AI inside NetSuite, inside their BI tool, inside their banking portal. This creates a dynamic where:

• Microsoft embeds OpenAI's technology into Office 365 as "Copilot" (Microsoft captures the enterprise relationship)
• Salesforce builds AI as "Einstein" on top of its CRM (Salesforce captures the value)
• SAP adds "Joule" AI assistant (SAP keeps the customer)

OpenAI provides the engine, but the car manufacturers capture the customer relationship and most of the value. OpenAI becomes the Intel Inside — valuable but commoditizable as competing models (Anthropic, Google, open-source) achieve comparable quality.

The unit economics problem. OpenAI's inference costs remain above its revenue per query for many enterprise use cases. Processing a complex financial analysis request costs $0.15-$0.50 in compute; enterprise subscriptions generate $0.08-$0.20 per equivalent query. The math requires either: massive scale improvements (declining), pricing increases (risking customer loss to competitors), or finding use cases where customers will pay premium prices (challenging in a market moving toward commodity AI).

The enterprise sales cycle. OpenAI's consumer products (ChatGPT) grew virally. Enterprise sales require: security reviews, compliance assessments, data residency negotiations, custom deployment, and 6-12 month procurement cycles. OpenAI has 200 enterprise salespeople competing against Microsoft's 60,000 and Salesforce's 30,000. Distribution advantage: incumbents.

The Investment Implication

Blodget's framework suggests:

Buy the selloff in incumbent SaaS. Companies like Salesforce, Intuit, and Workday that are successfully embedding AI into their existing products will emerge stronger — with higher prices justified by AI capabilities and deeper customer lock-in through AI-personalized workflows. The 20-30% drawdown is a buying opportunity.

Be cautious on standalone AI companies. OpenAI, Anthropic, and AI-first startups face the "picks and shovels" problem in reverse: they provide infrastructure but don't own the customer relationship. Valuations that assume direct enterprise revenue capture may be overstated.

Avoid the truly disrupted. Narrow-function software (basic writing tools, simple analytics dashboards, templated design products) faces genuine replacement by AI. These categories will consolidate or disappear — and companies concentrated here deserve lower multiples.

Why This Matters

For CFOs managing software portfolios: Your existing enterprise software vendors (SAP, Oracle, Salesforce, Workday) are likely to become more valuable to your organization as they embed AI — not less. Budget for 15-30% price increases on AI-enhanced versions rather than planning wholesale replacement.

For accounting firms: SaaS client valuations directly affect audit risk (going concern, goodwill impairment) and advisory work (M&A valuation, capital raises). Understanding whether a client is in "AI-enhanced incumbent" territory versus "AI-disrupted niche" determines the risk assessment.

For investors in fintech stocks: Distinguish between fintech companies that embed AI into irreplaceable workflows (Intuit in tax, Stripe in payments, Plaid in data) versus those offering standalone AI products without workflow lock-in. The former will thrive; the latter face commoditization.

For enterprise technology strategy: The "build versus buy" calculus hasn't changed — it's intensified. AI makes existing SaaS products better faster than most companies can build custom alternatives. The cost of building internally (AI engineers, infrastructure, maintenance) exceeds the cost of paying SaaS vendors' AI-enhanced pricing in almost all cases.

For the software industry broadly: Blodget's framework predicts consolidation: AI-enhanced incumbents absorb market share from AI-disrupted niche players. The number of enterprise software vendors will decrease, average vendor size will increase, and switching costs will rise as AI personalization deepens lock-in.

The key takeaway: The $800 billion software selloff confuses AI's role as an enhancer of existing enterprise software with AI as a replacer — most SaaS companies will embed AI to become more valuable and more locked-in, while OpenAI's real challenge is that enterprises prefer AI delivered through existing vendor relationships rather than standalone platforms.