Flex Ltd. (NASDAQ: FLEX) has agreed to acquire EPC Power for $4.4 billion in a transaction designed to position its Cloud and Power Infrastructure business at the centre of the shift toward 800-volt direct-current architectures inside increasingly power-dense artificial intelligence data centres. EPC Power is expected to generate approximately $800 million of revenue in calendar 2026, grow organically by about 40% in 2027 and reach an EBITDA margin of approximately 30% next year, giving Flex a business whose forecast growth and profitability materially exceed the current margin profile of the wider group. Flex expects the transaction to close during the fourth quarter of 2026 and then plans to separate its Cloud and Power Infrastructure business into an independently traded company during the first quarter of 2027. The result is an unusually important acquisition because EPC Power is not simply joining Flex. It is being bought immediately before becoming part of a new publicly traded AI infrastructure company.
EPC Power has more than 15GW of deployed power-conversion systems across 62 countries and expects annual U.S. manufacturing capacity to exceed 30GW during 2027. Its technology spans rectifiers, DC-to-DC conversion, grid-forming systems and planned solid-state transformer development, allowing Flex to extend a portfolio that already includes data-centre compute infrastructure, power equipment and cooling. The acquisition also gives investors a clearer view of what Flex wants the future spin-off to become: not merely a contract manufacturer serving AI customers, but an integrated infrastructure supplier controlling more of the path from the electrical grid to high-density GPU racks.
How expensive is Flex’s $4.4 billion acquisition relative to EPC Power’s revenue and future EBITDA?
Using EPC Power’s expected 2026 revenue of approximately $800 million, the $4.4 billion transaction value equals about 5.5 times current-year sales. That is a substantial multiple for a physical infrastructure and power-electronics company, although EPC Power’s expected 2027 growth and margin profile make the forward economics considerably stronger than the current revenue comparison alone suggests.
If revenue grows by the approximately 40% Flex expects in 2027, EPC Power would generate roughly $1.12 billion of revenue next year. Applying the company’s targeted 30% EBITDA margin produces approximately $336 million of illustrative EBITDA, which places the $4.4 billion purchase price at about 13.1 times that forward EBITDA estimate.
That calculation closely matches the strategic logic embedded in the deal. Flex is paying a premium for a business expected to expand revenue rapidly while achieving profitability materially above the wider manufacturing group’s present operating margins.
The risk is concentrated in the assumptions. A 40% growth forecast can look very different if data-centre projects are delayed, if customers slow power-system procurement or if manufacturing expansion creates execution problems. Likewise, a 30% EBITDA margin depends on product mix, pricing and factory utilization developing as expected.
The deal therefore becomes much more attractive if EPC Power reaches both targets simultaneously. If revenue continues growing but margins remain materially below 30%, the effective purchase multiple stays higher for longer.
Why are AI data centres moving toward 800V power architectures in the first place?
The computing density inside new AI systems is increasing fast enough that conventional data-centre electrical architectures are becoming less efficient. Every conversion step between the grid and a GPU rack loses some energy, creates heat and adds equipment that occupies valuable floor space.
Moving toward higher-voltage direct-current systems can reduce current for the same amount of delivered power, improving distribution efficiency and potentially allowing electrical infrastructure to become smaller or easier to manage at very high rack densities.
EPC Power is specifically developing systems around 800V data-centre architecture, including rectification and DC conversion. Flex says these technologies can provide clean 800V power for modern GPUs while supporting grid stabilization and backup power.
The architectural shift becomes more significant as accelerator platforms move toward rack-scale systems where a single installation consumes dramatically more power than conventional enterprise servers. Electrical design is consequently moving closer to the centre of computing performance because a customer cannot deploy the newest AI hardware if its power-distribution system cannot support the rack density.
This mirrors the transition already taking place in cooling. Liquid cooling was once a specialist feature used for particular high-performance systems. It is now increasingly treated as standard infrastructure for large AI deployments. Flex is betting that 800V power conversion will undergo a similar move from specialized engineering choice to common design requirement.
Why does owning EPC Power change Flex’s position against Vertiv and other AI infrastructure suppliers?
Flex has historically been known as an advanced manufacturing partner, but its Cloud and Power Infrastructure operation increasingly competes around integrated data-centre systems rather than manufacturing services alone. It already provides power, cooling and compute infrastructure, and EPC Power adds another proprietary technology layer in front of the rack.
That makes Flex strategically closer to companies such as Vertiv, Schneider Electric and Eaton, which are also trying to increase the amount of infrastructure they provide around every AI data centre. The competitive battlefield includes electrical systems, cooling, controls, switchgear, energy storage and increasingly the interface between utility power and GPU-level requirements.
EPC Power is particularly useful because power conversion can act as an architectural anchor. Once a customer designs a facility around a particular electrical architecture, downstream equipment must work around those decisions. Becoming involved at that stage can create cross-selling opportunities for Flex’s cooling, compute and manufacturing capabilities.
The acquisition follows a broader industry pattern. Vertiv recently agreed to buy UtilityInnovation Group for approximately $1.45 billion upfront to move further toward microgrids and behind-the-meter power. Flex is attacking a related problem from another point in the electrical chain by focusing on conversion and the transition to 800V architecture.
Both deals indicate that electricity is no longer being treated as an external utility input into AI computing. It is becoming a technology stack in its own right.
How important is the Cloud and Power Infrastructure business inside Flex before the planned spin-off?
Flex’s Cloud and Power Infrastructure segment generated approximately $2.2 billion of revenue in the fiscal first quarter of 2027, increasing 35% year over year as demand grew across power, cloud and cooling programmes. Adjusted operating income reached approximately $214 million, producing a 9.7% adjusted operating margin.
That means CPI already contributes more than one-quarter of Flex’s quarterly group revenue. Flex reported total first-quarter sales of $7.93 billion, up 21%, with group adjusted operating income of $534 million and adjusted EPS of $1.00.
EPC Power can alter the mix further because its expected 30% 2027 EBITDA margin is substantially higher than the current CPI adjusted operating margin. The measures are not directly comparable, since EBITDA and adjusted operating income exclude different items, but the direction is clear: Flex expects the acquisition to improve the growth and profitability profile of the business it intends to separate.
The timing makes this strategically important for future valuation. Investors often assign different multiples to high-growth AI infrastructure companies than to diversified contract manufacturers. Creating a standalone CPI entity can allow the market to value power and cooling growth independently from Flex’s remaining businesses.
EPC Power effectively becomes part of the equity story Flex will present to future SpinCo shareholders. That raises the importance of integration speed because the acquisition may have only a few months inside Flex before the planned separation occurs.
How will Flex finance a $4.4 billion acquisition before separating the business?
Flex says it is evaluating financing alternatives and expects to fund the transaction through a combination of debt and equity. Citi and Bank of America are providing committed financing supporting the acquisition, but the company has not yet disclosed a final mix showing exactly how much permanent debt and new equity will ultimately be issued.
This financing question matters because the transaction value is large relative to Flex’s existing annual free cash flow. Fiscal 2026 free cash flow was approximately $1.06 billion, while first-quarter fiscal 2027 free cash flow was only $41 million after working-capital investment and transaction-related costs.
A heavy debt structure could increase leverage immediately before the planned spin-off, while substantial equity issuance would dilute current Flex shareholders. The eventual capital structure needs to leave both Flex and the future CPI company with enough financial flexibility to invest after separation.
Management is also operating during a period of strong revenue expectations. Flex recently increased fiscal 2027 revenue guidance to $33.7 billion to $35.2 billion, with adjusted operating margin targeted at 7.0% to 7.2% and adjusted EPS of $4.42 to $4.74. That forecast does not incorporate the planned CPI spin-off.
The acquisition therefore is not being undertaken from a position of operating weakness. The issue is whether management can finance a major strategic expansion without transferring too much value to creditors or new shareholders just before the business structure changes.
What do EPC Power’s 15GW deployed base and 30GW manufacturing target say about execution risk?
EPC Power says it has more than 15GW of systems deployed across 62 countries, providing a meaningful operating record rather than leaving Flex dependent on an early-stage technology platform. More important for the next phase, U.S. manufacturing capacity is expected to exceed 30GW annually in 2027.
That planned capacity is enormous relative to the historical deployed base and helps explain how EPC Power could support 40% organic revenue growth next year. It also creates manufacturing risk because capacity only produces attractive returns if customer orders arrive quickly enough to keep facilities utilized.
AI infrastructure demand currently supports aggressive expansion, but equipment providers across power, cooling, servers and networking are all increasing capacity simultaneously. A shortage can eventually become excess supply if too many customers push projects out or if technology shifts faster than factories can adapt.
EPC Power’s advantage is that its products serve both data-centre and utility applications, providing some market diversification. Grid-forming technology and energy-storage systems can support applications outside AI computing, potentially reducing dependence on one end market.
The acquisition price nevertheless assumes a strong growth trajectory. Flex is paying for what EPC Power is expected to become in 2027, not merely what its first 15GW of deployments established historically.
Why has Flex stock not rallied sharply despite the strategic AI infrastructure acquisition?
Flex shares closed at $109.51 on September 4, the latest completed U.S. session because markets are closed for Labor Day on September 7. The stock gained 1.5% on September 4 and 1.8% on September 3, the acquisition-announcement session, but remained approximately 0.9% below the August 28 close of $110.50.
Compared with the August 7 close of $121.35, the shares were down approximately 9.8% over about one month. The recent 52-week range extended from roughly $53.07 to $166.86, putting the September 4 close well below the year’s high despite the strong underlying AI infrastructure growth.
The moderate reaction makes sense because the acquisition simultaneously strengthens growth and introduces financing uncertainty. Investors can see the strategic attraction of a company expected to produce $1 billion-plus of 2027 revenue at high margins, but they do not yet know how much debt or equity Flex will use to pay the $4.4 billion consideration.
The planned spin-off adds another variable because current shareholders are effectively buying into two transactions at once: a large acquisition and a corporate separation.
If EPC Power achieves its 40% growth and 30% EBITDA margin targets while the new company emerges with a manageable capital structure, the purchase could materially improve the valuation case for CPI. If financing costs rise or integration distracts management during the separation, the same deal could make an already complex transaction harder to execute.
Flex is paying approximately 13 times a very attractive version of EPC Power’s 2027 earnings.
The next year will show whether that version arrives.
Discover more from Business-News-Today.com
Subscribe to get the latest posts sent to your email.
