The impact of the 'AI bubble risk' on procurement and how to navigate it

By the end of 2025, business investments in AI-native tools had exploded by an average of 300% YoY.

AI’s potential and power has created a market of hyper-investment. But analysts, financial institutions and even big AI players now point to a “materially stretched” valuation of AI stocks.

Could the AI bubble be close to bursting? And if it does, what impact could this have on paying customers?

What is the “AI bubble risk”?

As we move past the honeymoon phase of AI, the financial world is starting to ask for evidence that these tools are delivering on their promised value.

The warning signs are already emerging:

• Valuations are astronomical: The Bank of England has warned that valuations for AI-focused tech firms are "particularly stretched". Share prices are so high they are vulnerable to a "sharp correction" if investor expectations aren't met.
• The dotcom comparison: Pierre-Olivier Gourinchas, Chief Economist at the IMF, believes that while we’re not yet at the levels of “market frothiness” seen in the dotcom period, there is cause for concern given the striking similarities.
• The “zero return” reality: Despite an estimated $30-$40 billion in enterprise investment into GenAI, a 2025 MIT study found that 95% of organizations are currently seeing no return on their investment.

The “burst” – and why finance and procurement should care

If the AI bubble bursts, the fallout isn’t just a stock market issue; it’s an operational one. End users may not see the risk immediately, but organizations will likely feel the impact:

• Startup failures and service disruption: If investment dries up, niche AI providers will run out of runway. Business-critical workflows tied to these platforms may not fail immediately, but development could stall, support may weaken and long-term reliability could erode. Organizations relying on these tools may face service interruptions, no advanced features or even forced migration to alternative solutions.
• The potential cost surge: Companies may see prices rise, discounts shrink and products paywalled as providers look for ways to boost profits.
• Shadow AI proliferation: Our data shows that 44% of AI-related spending occurs without procurement oversight – significantly higher than the 26% average for software spend overall. In a tightening market, this compounds risk. Duplicated capabilities, compliance exposure and cost escalation only becomes visible after contracts are signed.

Navigating the AI bubble risk

Now more than ever, companies should heed the call for greater financial discipline and stronger governance.

1. Negotiate for flexibility and protection

Structure AI agreements to maintain flexibility if the solution doesn’t meet expectations. Ensure contracts include:

• SLAs and break clauses: Include the right to terminate the contract if the provider fails to meet agreed-upon performance, accuracy or uptime standards – without cost.
• Data portability: Ensure all training data, prompts and outputs remain yours and can be exported in usable formats without friction, delay or extra cost.
• Commercial flexibility: Avoid long-term, seat-based contracts for unproven tools. Opt for consumption-based pricing with clear usage limits, formal review points, and a cap on price increases to manage cost risk.

2. Monitor and optimize AI spend

AI budgets are growing rapidly, but market volatility means inefficiencies can quickly translate into real costs. To protect value and control spend:

• Identify overlap: As established SaaS vendors expand AI functionality, some point solutions may become redundant. Consolidate where possible to reduce duplication and minimize spend.
• Align spend with usage: 7% of AI tools in the average SaaS stack go completely unused and a further 40% underutilized. Procurement platforms with usage analytics can identify this waste and recommend whether to rightsize or terminate contracts.
• Negotiate more favorable pricing: Providers may raise prices at renewal, but many are still willing to offer discounts to retain customers. Leverage utilization data, competitive benchmarks and contractual levers to negotiate more favorable terms and pricing.

3. Formalize AI governance

Eliminate unsanctioned AI tools by ensuring all requests pass through formal intake and approval processes.

• Centralize intake: Prevent ad-hoc purchases on credit cards or other informal channels, and ensure every AI request passes through security and compliance reviews before approval.
• Maintain a vendor repository: Track all approved AI tools, subscriptions and contracts in a central system to enable oversight, renewal planning and spend analysis.

Moving from hype to discipline

Many organizations have exempted AI purchasing from standard procurement, financial and even governance rules in an attempt to capitalize. Tools have been piloted quickly, budgets approved based on hype rather than ROI and formal oversight frequently bypassed.

However, with stark warnings about overvaluations, and investors and businesses alike wanting to see imminent returns, companies are now financially exposed to a possible AI market crash.  

Being diligent, understanding the financial and usage models of AI tools, and ensuring you have control over intake gives you the best chance of mitigating the impact this bubble may have.

Our Buying AI Playbook helps you turn these safeguards into action, exposing the hidden risks in AI licensing and guiding you toward data-backed, value-driven deals.