Every operating partner I talk to has the same slide somewhere in their deck: “AI-Driven Value Creation.” It’s usually around slide 14, sandwiched between procurement optimization and management team upgrades. It looks great. Specific use cases. Projected EBITDA impact. A timeline that conveniently maps to the hold period.
Then I ask a simple question: who’s going to build it?
The answers come fast — the portco CTO, a preferred vendor, the team that delivered at the last deal. But compare how systematized those answers are to the rest of the operating playbook. Most mid-market firms have a repeatable framework for pricing, procurement, the 100-day plan. Engineering capacity? It’s typically the least systematized lever in the value creation toolkit — solved through relationships and pattern recognition rather than a deployable infrastructure play that compounds across the portfolio.
The New Math
Bain’s 2026 Global PE Report introduced a phrase worth remembering: “12 is the new 5.” In the 2010s, a typical deal needed roughly 5% annual EBITDA growth to deliver a 2.5x MOIC — buoyed by 50% leverage at low rates and reliable multiple expansion. Today, with borrowing costs in the 8–9% range, leverage ratios at 30–40%, and purchase multiples at record highs, firms need 10–12% annual EBITDA growth to hit the same return. And at seven-year average hold periods, the IRR math makes every quarter of delay progressively more expensive. [1]
Financial engineering’s contribution to value creation has fallen from 51% in the 1980s to roughly 25% today. Operational improvement now drives 47% or more. [2] If you can’t grow EBITDA through operational execution, you can’t reliably generate the returns LPs expect.
And you’re doing this against a brutal backdrop. Buyout funds globally hold 32,000 unsold portfolio companies sitting on $3.8 trillion in unrealized value. [1] Every quarter your value creation plan stalls, the exit window gets more expensive.
For technology-driven businesses — which describes most of the mid-market at this point — the path to EBITDA growth runs through engineering. The question isn’t whether you need engineering capacity. It’s whether you can get it fast enough to matter.
The Capacity Problem
This isn’t just a PE problem, by the way. The engineering capacity gap is structural across the mid-market — any company between $50M and $500M in revenue, competing for the same talent against hyperscalers paying 2–5x more, trying to ship product and absorb AI at the same time. PE portcos face the sharpest version of it, because there’s a clock.
Your mid-market CTO — solid operator, probably promoted from VP Engineering — is being asked to simultaneously ship product roadmap, reduce tech debt, integrate one or two acquisitions, modernize the platform, keep the lights on, and “figure out AI.” Same team. Same budget. If it’s a portco, add a hold period clock.
They can’t hire fast enough. Mid-to-senior engineering roles take 49–62 days to fill on average; specialized AI/ML roles average 54 days, and in regulated industries stretch to six or seven months. [3] They can’t retain, either. Senior engineers at Google, Meta, and Apple earn $300K+ in total comp. A mid-market company offers $120–$160K base with limited upside. AI/ML roles command 67% higher salaries than traditional software engineering positions. [3][4]
McKinsey found tech talent demand running 2–4x greater than supply. [5] IDC reports 67% of digital transformations are delayed due to skill shortages. [6] And PE-backed companies face an additional dynamic: 73% of portco CEOs are replaced during the investment lifecycle [7] — often the right move, but that leadership transition cascades through engineering orgs. Research shows each senior departure increases the likelihood of further attrition by 9–13%.
This isn’t cyclical. It’s structural.
AI Exposes the Gap — It Doesn’t Close It
We’ve all seen the data by now. McKinsey, MIT, BCG — the studies pile up and they all say roughly the same thing: AI adoption is nearly universal, meaningful EBIT impact is rare, and the gap between the two is widening. [8]
The PE-specific picture is no better. Most PE firms cannot show meaningful returns from AI across their portfolio companies. [9] Nearly half are still managing AI at the individual portco level — which, depending on your portfolio composition, may be leaving significant cross-portfolio leverage on the table. [10]
But the stat that matters most isn’t about adoption rates. It’s about what separates the small cohort actually generating value from everyone else. The answer, consistently, is that they redesigned workflows before deploying technology. They’re more than 3x more likely to pursue transformative change rather than incremental automation. [8] They treat AI not as a tool you bolt onto existing processes, but as a forcing function that reveals which processes are worth keeping.
AI isn’t a project you assign alongside twelve other priorities. It’s a transition your organization goes through — mapping workflows, validating use cases against real data, building feedback loops, instrumenting systems for measurement. The companies that start this transition early compound their advantage. The ones that wait are building on an operating model their competitors have already discarded.
The smart money isn’t asking “which AI tool should we buy?” It’s asking “do we have the engineering infrastructure to actually absorb this technology?”
For most mid-market companies, the answer is no.
Stop Thinking About Engineering as Procurement
When companies recognize engineering capacity as the binding constraint, the instinct is to buy more. Hire domestically. Engage a services firm. Maybe both.
These aren’t wrong instincts. But they’re incomplete — because they treat engineering as something you procure, not something you build.
Domestic hiring is a great answer — when you can afford it and the timeline allows it. But at 49–62 day fill times, a 2–5x compensation gap against FAANG, and a mid-market brand that doesn’t pull candidates off the bench, “right answer” and “feasible answer” aren’t the same thing. You’ll spend two quarters recruiting a team that a firm with existing infrastructure staffed in weeks.
Strategic services partnerships are the right model for a lot of work. Bounded scopes, specialized expertise, capacity surges around acquisitions or migrations, teams that can flex with deal flow. A good long-term services partner builds institutional knowledge, understands your architecture, and operates as an extension of your team. This works well when the need is ongoing execution support across a portfolio of changing priorities.
Where the model strains is when engineering capacity isn’t a service you need — it’s infrastructure you’re missing. When the value creation plan requires 30, 50, 75 dedicated engineers for three to five years, building your product, accumulating domain knowledge, shipping on your roadmap — that’s not a services engagement. That’s an org you need to own.
The distinction matters because of what happens at exit. A services relationship, no matter how good, is a vendor contract the buyer needs to evaluate and potentially renegotiate. An owned engineering center is a transferable asset — it increases the multiple because the buyer inherits a functioning engineering organization, not a dependency.
The Math That Changes the Conversation
Let’s make this concrete with two scenarios. Same company — a mid-market software business, $80M revenue, $20M EBITDA, acquired at 12x for $240M enterprise value. All cost estimates are conservative.
Scenario 1: 30-person team. The value creation plan calls for 30 engineers to accelerate the product roadmap. Domestically, that’s 30 × $200K fully loaded = $6M per year. In a dedicated captive center, it’s 30 × $55K fully loaded = $1.65M per year, plus roughly $400K in setup. The delta: $4.35M per year. At a 12x exit multiple, that cost structure advantage is worth $52M in enterprise value — a 22% uplift on a $240M entry.
Scenario 2: 75-person team. Now the plan calls for a full engineering org — product roadmap plus AI integration. Domestically, 75 × $200K = $15M per year. Captive center: 75 × $55K = $4.1M per year, plus $750K setup. The delta: $10.9M per year. At 12x: $131M in additional enterprise value — a 55% impact on entry. And you had the full team operational in ten weeks versus six to nine months of recruiting.
In both scenarios, the captive center is part of the asset at exit. The buyer inherits an operational engineering organization with institutional knowledge, established processes, and 95%+ annual retention rates. That’s a different due diligence conversation than inheriting a stack of vendor SOWs.
The point isn’t that one model is universally superior. A 15-person services engagement might be exactly right. But operating partners should be running this math for every portco where engineering is the value creation lever — and most aren’t.
The Infrastructure Is Already There
A decade ago, standing up a captive engineering center took 12–18 months and a million-dollar bet on real estate and local entity formation. That’s no longer the reality.
India’s GCC ecosystem now spans 2,100+ centers, generating $98 billion in revenue and employing 2.36 million professionals — with more than 500 PE-backed centers in the mix. [11] The sector has evolved far beyond cost arbitrage: 45% of GCC work is now classified as expertise and frontier activity. [11] Mid-market adoption is accelerating fastest — 583 mid-market GCCs, with 35% established in just the last two years. Attrition rates have dropped to 9%, well below the 12–17% range typical of US engineering organizations. [12]
PE-specific return data makes the case concrete: companies that globalized 35–40% of R&D saw total enterprise value increase by 4–8x during their hold periods, with IRR improvements of 300–400 basis points. [13]
The largest PE firms have already figured this out. The top technology-focused funds maintain 100+ person internal operating teams, dedicated AI incubators, and portfolio-wide engineering playbooks. They implement changes in the first 100 days that other firms take a year to execute.
But most mid-market companies and mid-market PE firms don’t have those resources. They need a partner who can stand up the infrastructure — the entity, the facility, the team, the compliance, all the ground-level work — staff it, and run it until it’s self-sustaining. That’s a different offering than outsourcing, and it’s a different offering than consulting.
What We Do
This is the work we do at Trigent and Morphomix.
Trigent builds and operates dedicated engineering centers — captive GCCs with 3,000+ engineers behind them, operational in as few as ten weeks. We handle everything from local entity formation and facility buildout to hiring, onboarding, and ongoing operations — with full IP ownership retained by the company. And when the center reaches maturity, we transfer it. Full ownership of the entity, the team, and the operations passes directly to the company. Build, operate, transfer — what starts as a partner-led buildout becomes a wholly owned engineering organization on your balance sheet.
For companies that need strategic engineering support without the GCC model, we run dedicated teams with the continuity and architectural depth of an in-house org.
Morphomix brings the AI implementation methodology — our FABRIC framework — for mapping workflows, testing use cases against real data, and building the organizational capacity to move from pilot to production. Because engineering capacity without a clear transition plan just gives you more people doing the wrong things faster.
Across our PE-backed engagements, we’ve built dedicated engineering teams of 100+ that delivered 3x release velocity, 40% cost reduction, and 95%+ annual retention — operational within ten weeks of contract signing. We do this for mid-market companies across healthcare, logistics, insurance, and technology — the companies that face the sharpest version of the talent math and need it solved, not studied.
In a world where 12 is the new 5, the firms that treat engineering as infrastructure — not procurement — will compound returns across hold periods. The ones that don’t will be selling against operators who did.
Sources
[1] Bain & Company, Global Private Equity Report 2026 (February 2026). “12 is the new 5” — deals now require 10–12% annual EBITDA growth for 2.5x MOIC; $3.8T unrealized value across 32,000 unsold portfolio companies; average hold periods at ~7 years.
[2] PwC, How Private Equity Operating Partner Roles Are Changing. Operational value creation at 47% (up from 18% in the 1980s); financial engineering down to 25% from 51%.
[3] KORE1, AI/ML Talent Map 2026. AI/ML roles average 54-day fills; 67% salary premium over traditional software engineering.
[4] Second Talent, Global AI Talent Shortage Statistics 2026. AI talent demand-to-supply dynamics and compensation benchmarks.
[5] McKinsey & Company, Tech Talent Gap: Addressing an Ongoing Challenge (2025). Tech talent demand 2–4x greater than supply.
[6] IDC, via CIO.com. 67% of digital transformations delayed due to skill shortages.
[7] AlixPartners. 73% of portco CEOs replaced during investment lifecycle.
[8] McKinsey & Company, The State of AI in 2025 (November 2025). 88% AI adoption; 39% report any EBIT impact; 6% qualify as high performers; high performers 3.6x more likely to pursue transformative change.
[9] BCG, Inside the AI-First Private Equity Firm (January 2026). Most PE firms cannot show meaningful AI returns across portfolio companies.
[10] FTI Consulting, Three Plays for Driving Value Creation in 2025. 40% of PE firms manage AI at the individual portfolio company level.
[11] Zinnov-NASSCOM, GCC Landscape 2026 (May 2026). 2,100+ centers; $98B revenue; 2.36M professionals; 504 PE-backed; 45% expertise/frontier work; 583 mid-market centers.
[12] EY, GCC Pulse Survey 2025. GCC attrition at 9%; 92% of GCC leaders affirm contribution beyond cost savings.
[13] Zinnov, PE Advisory Practice Data. 35–40% R&D globalization correlated with 4–8x enterprise value growth and 300–400bps IRR improvement during hold period.
