Late-May dealmaking shows capital chasing AI infrastructure and cash-flow assets

Operators and investors,

Recent deal activity points to a clear underwriting pattern in private equity: capital is moving toward assets with either long-duration infrastructure demand or highly defensible cash flow.

That does not mean the broader mid-market is closed. It means buyers are applying a higher standard of proof.

For generalist mid-market companies, the bar is moving away from broad growth narratives and toward measurable cash flow, cost visibility, recurring revenue quality, and operational control. Assets with clean revenue durability and credible margin defense are easier to finance. Assets with aggressive forecasts, weak add-backs, or unclear cash conversion are facing more scrutiny.

This matters for portfolio companies preparing for a refinancing, bolt-on, exit process, or internal re-underwriting exercise in the next 6 to 12 months.

1. AI infrastructure is being underwritten as infrastructure, not software

The Google and Blackstone TPU cloud joint venture is a useful signal.

This carries a lot more value than just “AI exposure.” The transaction structure points to a more specific thesis: compute demand, dedicated capacity, large-scale capital deployment, and long-duration infrastructure economics.

The venture is expected to bring significant data center capacity online and provide access to Google TPUs through a compute-as-a-service model. Blackstone’s role is not a traditional software growth bet. It is an infrastructure allocation tied to constrained supply and durable demand.

The same pattern appears in Anthropic-related financing discussions, where large debt packages are being structured around compute expansion and TPU procurement. The relevant underwriting variable is not product-led growth in the traditional SaaS sense. It is compute availability, contracted demand, financing structure, and the economics of AI infrastructure consumption.

If your portfolio company has material AI usage, compute dependency, or AI-enabled delivery claims, the board should not settle for a vague “AI roadmap.” It needs a 24-month AI unit economics model.

At a minimum, that should include:

• Expected compute spend by workflow

• Impact on gross margin

• Expected reduction in cost-to-serve

• Dependency on third-party AI vendors

• Data infrastructure requirements

• Workflow-level productivity assumptions

• Governance and security exposure

2. Power and infrastructure are becoming direct AI inputs

AI infrastructure cannot be separated from electricity demand.

The AES take-private is relevant because it reflects how large-scale power generation and grid exposure are being evaluated in the current market. Data centers, AI workloads, and electrification all increase the strategic value of generation capacity, transmission access, and related infrastructure.

This changes how certain assets should be analyzed.

Power, fiber, cooling, data center services, grid services, and energy-adjacent platforms should not be benchmarked only against old sector narratives. Their relevance to AI infrastructure demand now affects buyer interest, financing appetite, and valuation logic.

For portfolio companies with exposure to these areas, the operating question is straightforward:

Can the company demonstrate a measurable linkage to durable infrastructure demand?

That means showing:

• Contracted or recurring demand

• Capacity constraints that support pricing

• Margin durability

• Customer concentration risk

• Capex requirements

• Regulatory or permitting exposure

• Sensitivity to energy pricing and utilization

The companies that can quantify these points will have a stronger process than those relying on generic “AI infrastructure tailwind” language.

3. Healthcare services remain attractive where cash flow quality is clear

Healthcare services continue to attract sponsor interest because many subsectors offer revenue visibility, payer diversification, fragmentation, and roll-up potential.

PwC’s 2026 health services deals outlook points to a market where deal value and volume are expected to improve as higher-quality assets come to market. The major headwinds remain reimbursement pressure, regulation, labor cost, and operating complexity. The attractive assets are those with cash generation, visibility into reimbursement, and defensible operating models.

The lesson is not limited to healthcare & buyers are prioritizing companies that can prove:

• Recurring or highly predictable revenue

• Clean retention and renewal data

• Defensible margins

• Limited add-back dependency

• Credible Quality of Earnings support

• Clear operating controls

• Measurable synergies where a buy-and-build strategy is involved

This is where many roll-up stories are being challenged. A consolidation thesis is not enough. Buyers are testing whether the operating model can actually integrate acquisitions, realize synergies, preserve customers, and maintain margin.

Cross-sell assumptions, “one-time” cost adjustments, and platform synergies are being reviewed with less tolerance than in the 2018 to 2021 period.

4. Generalist mid-market assets need stronger proof

Many mid-market assets outside high-conviction sectors are facing a more difficult valuation discussion.

Common friction points include:

• Sellers anchored to prior-cycle valuation expectations

• Buyers are under pressure from labor, software, compliance, and financing costs

• Quality of Earnings reviews rejecting recurring “one-time” adjustments

• The current pipeline conversion does not support revenue forecasts

• Weak visibility into cash conversion and working capital needs

• Unclear differentiation beyond market growth assumptions

In this environment, a broad growth story won’t be enough; the company will need a tighter operating case.

That means proving the business can generate cash under the current cost of capital, not under 2021 assumptions.

5. What portfolio teams should run in the next 90 days

For portfolio companies preparing for any capital event, the next 90 days should focus on the quality of evidence.

a. Pricing power audit

Review every SKU, contract, customer cohort, and renewal motion from the last 18 months.

The goal is to identify where pricing has not kept up with cost inflation, product value, service expansion, or customer usage.

The output should be specific where:

• Pricing is below market

• Discounting lacks approval discipline

• Renewals are not capturing the value delivered

• Legacy contracts are suppressing margin

• 3% to 7% increase is defensible without material churn risk

Pricing is one of the fastest ways to improve enterprise value when the business has real customer value and weak pricing discipline.

b. AI cost-to-serve rebaseline

Move AI out of the innovation category and into the operating model.

Which workflows become cheaper, faster, or less dependent on manual labor?

Model AI impact across:

• Support

• Content

• Engineering

• QA

• Reporting

• Internal operations

• Sales enablement

• Customer success

If AI does not change cost-to-serve, throughput, cycle time, or error rates, it is in experimentation, so work towards turning it into a value creation lever.

c. Recurring revenue reclassification

Reclassify revenue into three categories:

• Contractually recurring

• Behaviorally recurring

• Expected renewal

Many management reports treat these as the same. Buyers will not.

If the company does not recalculate the revenue base first, diligence teams will do it with more conservative assumptions.

The board should know what percentage:

• Of the revenue is contractually protected

• Is habit-based

• Depends on renewal optimism.

d. Cash flow narrative rebuild

The exit or refinancing narrative should start with cash flow quality.

Growth still matters, but it should be sequenced after the proof base.

A stronger process starts with:

• Revenue quality

• Gross margin stability

• Cash conversion

• Churn and renewal data

• Cost structure visibility

• Working capital discipline

• Operating levers already under management control

Then the company can present upside.

Buyers are more receptive to optional upside when the base case is clean.

e. Re-underwriting exercise in a single page

This is the exercise I would run with the CFO and head of strategy this week.

Pull the last six board decks.

Remove every forward claim that is not supported by current contracts, current pipeline conversion, or actual operating data.

Then rebuild the case using:

• Today’s cost of capital

• Realistic margin pressure

• Current software, labor, compliance, and infrastructure costs

• AI impact on cost-to-serve

• Recurring revenue split by quality

• Current cash conversion

• Downside case buyer assumptions

The final output should fit on one slide:

• Who is the most likely buyer?

• What multiple range is defensible today?

• Which three proof points must improve over the next two quarters?

• Which operating metrics need to be fixed before a process starts?

As always, enjoying your reports and market outlooks replying to the weekly emails.

Best,

Mario

📊 I've watched 8 portfolio companies build attribution dashboards before they built the data model. The result is always the same: reports that can't answer board questions. Attribution is an architecture problem disguised as a reporting one - and the agentic AI wins everyone celebrates run on hundreds of hours of methodical data infrastructure work.

📉 My dormant Facebook account got 27x more views than my LinkedIn on the same post - with 1/8th the followers. 94% of my Meta views came from non-followers, while LinkedIn lost its networking layer between 2017 and 2023. Facebook is operating as a network while LinkedIn became media - for a business network, that's unacceptable.

💼 Tech layoffs in 2026 just hit their highest level since Q2 2023, and I'm watching candidates chase the wrong jobs. The flexible-hours, WFH, top-brand role is 1 seat in 10,000 applicants. The on-site, ambitious, hands-on, multi-hat role is 1 in 10 - and 1 in 3 if you bring the skills. Career growth lives in the second column.

Vertical AI for high-stakes professional services keeps gaining traction over general-purpose tools. Crimson raised $2.5M for its litigation-specific AI platform, already deployed on disputes worth over $40 billion across Magic Circle and Am Law 10 firms.

Risk-adjustment coding is becoming a serious legal liability for managed care and its PE-backed partners. Massachusetts sued UnitedHealthcare for allegedly inflating MassHealth risk scores to extract $100M in excess payments.

Specialist AI agents are eating regulated back-office workflows horizontal tools never could. Gradient Labs raised $26M with 900% revenue growth and performance guarantees backed by refunds as a new commercial benchmark.

European tech funding remains hostage to a handful of mega-rounds in digital infrastructure. The last week of May saw €3.1B invested across 60+ deals, but Pure DC's $2.7B raise alone accounted for the bulk, with the UK capturing €2.6B of the total.

Dog Breeding Marketplace: 20-year-old online marketplace specializing in dog sales and breeder listings. Generates revenue via paid breeder listings, premium featured placements, and third-party display advertising from pet-related brands. $117,987

Home Decor PrestaShop Brand: 20-year-old PrestaShop brand specializing in decorative wall stickers, custom photo murals, vinyl rugs, and designer wallpaper. Backed by robust SEO-driven organic demand and streamlined lean operations. $296,316

B2B Facilities SaaS: 9-year-old B2B facilities management platform serving institutional and enterprise clients including public sector bodies, energy groups, and engineering firms. Revenue is generated via scalable m²-based pricing model. $1,325,715

Premium Car Seat Covers Shopify Brand: 3-year-old Shopify brand specializing in premium, custom-fit neoprene car seat covers. Operated by a lean team with strong systems and documented SOPs. $3,448,000

Working with me

🌐 Scaling $30M - $100M+ mid-market companies with value creation through RevOps, data engineering, and WordPress. DevriX provides full RevOps consulting + delivery with GTM enablement for PE-backed portfolio companies, traditional tech, healthcare, finance, and professional service businesses pacing toward revenue growth initiatives. Our standard retainers between $10K and $60K include revenue lifecycle services for marketing and sales leaders, FP&A for financial teams, pipeline enrichment through websites and dozens of lead sources, automations and delivery integrations, CRO and ongoing testing, product delivery and platform integration solutions, and more through our consulting solutions.

🚀 1:1 Advisory retainers. At Growth Shuttle, I run two popular plans: Async Advisory ($3,500/mo) for $3M - $30M founders and executive teams and the smaller Strategic Growth Circle ($997/mo) for $100K - $500K entrepreneurs, agency founders, scale ups. My fractional executive plan is also available here.

📈 Building US LLCs from Europe. I help European and Asian founders scale faster through doola and their “Business in a Box” model. Also suitable for US citizens (given their bookkeeping solution), but in very high demand across Europe.

📊 Post-Merger Integration. We take on M&A initiatives with Flippa. Working closely on PMI retainers for PE companies and fast-growing startups integrating new companies within their portfolios, enabling data pipelines, and securing more deals through my personal network.