DigitalBridge, JEXI Close NEC Data Center Deal in Japan Push

DigitalBridge Group and Japan Excellence Investment Corporation (JEXI) have closed their acquisition of select data center assets from NEC Corporation, the companies announced Monday, marking a significant expansion into Japan's enterprise colocation market as demand for AI-ready infrastructure outpaces supply across the Asia-Pacific region.

The deal, first announced in December 2025, brings two strategically located facilities in Tokyo and Osaka under the control of the joint venture partners. While financial terms weren't disclosed, the transaction reflects a broader pattern of global infrastructure investors targeting mid-tier data center operators in markets where hyperscale providers have cornered the lion's share of available capacity.

What makes this deal notable isn't just the geography — it's the timing. Japan's data center market has historically lagged behind Singapore, Hong Kong, and Sydney in attracting international capital, constrained by high land costs, complex permitting, and power availability challenges. But those dynamics are shifting as enterprises across Asia scramble to secure colocation capacity for AI workloads that can't wait for new builds to come online.

DigitalBridge and JEXI are betting that NEC's existing customer base — predominantly Japanese enterprises with conservative IT procurement preferences — represents a sticky revenue stream that larger operators have largely ignored in their race to serve hyperscalers. The question is whether that bet pays off before power constraints and cooling limitations force expensive retrofits.

The acquired facilities operate in Japan's two most data-dense metropolitan areas, but they're entering a market environment that's fundamentally different from when NEC originally built them. Tokyo's Inzai facility and Osaka's operation were designed for traditional enterprise workloads — email servers, ERP systems, corporate applications that drew modest power and generated manageable heat.

AI inference and training workloads draw three to five times more power per rack than those legacy applications. That's forcing data center operators across Asia to either limit density, turn away high-power clients, or invest heavily in upgraded electrical and cooling infrastructure — often all three simultaneously.

Japan's power grid adds another wrinkle. Unlike markets such as Northern Virginia or Singapore, where data center operators can relatively quickly secure additional megawatts, Tokyo's utility approval process can stretch 18 to 24 months for meaningful capacity additions. Osaka faces similar constraints, compounded by local opposition to industrial-scale power consumption in urban and suburban zones.

This creates an interesting positioning challenge for the new ownership. If DigitalBridge and JEXI pursue aggressive density upgrades to compete for AI workloads, they'll burn capital and time on infrastructure overhauls while competitors with newer builds capture market share. If they stick with the existing enterprise-focused model, they risk becoming subscale in a market increasingly defined by hyperscale anchors and GPU clusters.

The deal was structured as a selective asset sale, not a full exit. NEC retained certain data center operations and continues to operate its broader IT services and infrastructure business. That's worth noting — it suggests NEC viewed these specific facilities as non-core or sub-scale relative to the capital requirements needed to keep them competitive.

For a company like NEC, which has been shifting focus toward 5G infrastructure, semiconductors, and biometric security systems, maintaining mid-market colocation assets likely represented a capital allocation problem. The facilities generated steady cash flow but required ongoing investment to stay relevant — the kind of asset that looks better on a pure-play infrastructure investor's balance sheet than a diversified technology conglomerate's.

DigitalBridge, by contrast, has built its data center strategy around exactly these kinds of carve-outs. The firm's portfolio includes stakes in Vantage Data Centers, DataBank, and Scala Data Centers — all operators targeting the gap between hyperscale and small-scale retail colocation. JEXI brings local market knowledge and relationships that matter in Japan's consensus-driven business environment, where longstanding customer relationships and trust carry more weight than in Western markets.

The facilities serve what insiders call the "forgotten middle" — companies too small to justify building their own data centers but too large or too risk-averse to rely entirely on public cloud providers. In Japan, that segment remains surprisingly large, driven by data residency requirements, legacy application dependencies, and institutional caution about migrating mission-critical workloads to hyperscaler infrastructure.

Japan's data center market has bifurcated over the past three years. At the top end, hyperscalers — AWS, Google Cloud, Microsoft Azure — have locked down massive capacity blocks in and around Tokyo, often through long-term leases with providers like Equinix and NTT. At the bottom end, small retail colocation providers serve SMBs and niche workloads.

DigitalBridge's thesis appears to be that the middle-market enterprise segment — companies running hybrid infrastructure, maintaining on-premises compliance obligations, or simply moving slower than the cloud-native generation — has been underserved as operators chase hyperscale anchor tenants. That gap is particularly pronounced in Japan, where cultural and regulatory factors make cloud adoption slower than in the U.S. or parts of Europe.

The firm has deployed this playbook before. In Brazil, DigitalBridge-backed Scala Data Centers targets regional enterprises and carriers. In the U.S., DataBank focuses on edge colocation in secondary markets. The common thread: acquire facilities with existing customer bases, improve operational efficiency, and avoid direct competition with the hyperscale-focused giants.

JEXI's involvement adds local credibility and access to Japanese institutional capital. The investment corporation has a track record in domestic infrastructure assets and brings relationships with Japanese banks, insurers, and pension funds — the kind of long-term, low-cost capital that makes data center investments pencil at single-digit returns.

But there's a tension embedded in the strategy. If the facilities remain focused on traditional enterprise workloads, revenue growth will be limited — Japan's enterprise IT spending is growing, but slowly. If the new owners push toward higher-density, AI-ready infrastructure, they'll need to invest heavily in power and cooling upgrades, potentially eroding returns unless they can command premium pricing.

The middle path — selective density upgrades for anchor clients willing to commit to long-term contracts — requires precise execution and customer selection. Get it right, and the facilities become differentiated assets in a supply-constrained market. Get it wrong, and they risk becoming stranded in the middle: too expensive for legacy workloads, not competitive enough for AI.

It's worth asking why this deal happened at all. If Japan's data center market is so constrained and demand is surging, why did NEC sell rather than scale? And why didn't a hyperscale-focused operator snap up the assets?

The answer likely comes down to scale and flexibility. The Tokyo and Osaka facilities, while well-located, aren't large enough to serve as hyperscale anchors. AWS or Google isn't leasing half a building — they're leasing entire campuses, often with options to expand across adjacent parcels. For NEC, competing in that market would have required land acquisition, multi-hundred-million-dollar capital outlays, and a fundamental shift in business model.

DigitalBridge has been methodically building exposure to Asia-Pacific data center markets over the past two years, largely through partnerships and joint ventures rather than outright acquisitions. The NEC deal follows investments in India and Southeast Asia, where the firm has targeted markets with undersupplied enterprise colocation segments and favorable long-term demand drivers.

Japan represents both an opportunity and a challenge within that strategy. On one hand, it's a wealthy, stable market with low political risk and strong rule of law — attributes that matter when you're deploying patient capital into long-life infrastructure assets. On the other, it's expensive, slow to permit, and dominated by entrenched local players with deep government and corporate relationships.

The JEXI partnership addresses the latter concern. By aligning with a Japanese institutional investor, DigitalBridge gains access to local networks, regulatory knowledge, and — critically — domestic funding sources that can provide yen-denominated debt at rates that would be unavailable to a foreign-only investor.

This matters because data center returns are highly sensitive to cost of capital. A facility generating 8% unlevered returns looks very different depending on whether you're financing it with 4% local debt or 7% international credit. JEXI's involvement likely improves the financing structure materially, making the deal viable at acquisition prices that might not work for a standalone foreign buyer.

Every data center deal announcement in 2025 and 2026 has included some reference to AI workload demand. This one's no different — the press release mentions AI, machine learning, and digital transformation as key growth drivers. But the economics of serving AI workloads are complicated, and not every facility is positioned to capture that demand profitably.

AI training workloads require ultra-low latency networking, GPU-dense configurations, and massive power delivery — often 50 to 100 kilowatts per rack versus 5 to 10 kilowatts for traditional enterprise workloads. That means expensive retrofits: upgraded electrical infrastructure, liquid cooling systems, and often entirely new mechanical plants. For existing facilities like the NEC assets, those upgrades can cost as much as building new — without the benefit of greenfield design optimization.

The biggest risk to this investment isn't competition or customer demand — it's electricity. Japan's power grid operates on two incompatible frequencies (50Hz in the east, 60Hz in the west), limiting inter-regional transmission capacity. Nuclear generation remains politically contentious post-Fukushima, and renewable buildout is constrained by limited land availability and offshore wind development timelines.

If Tokyo and Osaka utilities can't or won't approve significant new data center load, the acquired facilities face a hard ceiling on growth. DigitalBridge and JEXI could optimize operations, upgrade cooling, and improve sales execution — but they can't manufacture megawatts. That makes this deal, in some ways, a bet on Japanese energy policy as much as data center market dynamics.

There's precedent for this risk materializing. Singapore imposed a moratorium on new data center development in 2019, citing power and land constraints, and only cautiously lifted it in 2022 with strict efficiency requirements. If Japan follows a similar path — limiting new load approvals or imposing punitive efficiency standards — existing capacity becomes more valuable, but expansion becomes nearly impossible.

That might actually work in DigitalBridge and JEXI's favor. Scarcity creates pricing power, and if permits become harder to get, the facilities they just bought become more defensible. But it also caps the upside — you can't grow revenue if you can't add racks.

The broader signal here is about where infrastructure capital is flowing as the first wave of hyperscale data center expansion matures. In the U.S. and Western Europe, most of the obvious high-quality assets have been picked over, repriced, and often rolled into REITs or publicly traded platforms. The returns available on stabilized, well-located facilities have compressed to levels that only work for very low-cost capital.

Asia, by contrast, still offers pockets of inefficiency — markets where local operators built facilities for one set of workloads, regulatory environments evolved, and now those assets are mispriced relative to replacement cost or strategic value. Japan is one of those markets. So are parts of India, Indonesia, and Vietnam.

Other global infrastructure investors will be watching how this deal performs, particularly the execution around power procurement and customer retention. If DigitalBridge and JEXI can demonstrate that mid-market enterprise colocation in Japan generates stable, inflation-linked cash flows without requiring massive capital recycling, expect similar carve-outs from other Japanese IT services firms.

NEC isn't the only Japanese conglomerate with non-core data center assets. Fujitsu, Hitachi, and others operate facilities that might be candidates for similar transactions if the economics prove compelling. The challenge is structuring deals that work for both sides — sellers want clean exits at reasonable valuations, buyers want quality assets without hidden liabilities or stranded costs.

The JEXI partnership model also provides a template for how foreign capital can enter Japan's notoriously relationship-driven market. Rather than attempting to navigate local business culture and regulatory processes alone, international investors can partner with domestic institutions that bring credibility, networks, and access to local funding. That's not unique to data centers — it's how infrastructure investing has worked in Japan for decades — but it's being applied to digital infrastructure more systematically now than in the past.

Whether this deal ultimately generates strong returns depends on variables that won't fully resolve for several years: power availability, AI workload adoption rates among Japanese enterprises, competitive dynamics as hyperscalers potentially move downmarket, and the operators' ability to execute infrastructure upgrades on budget and on schedule.

What the announcement doesn't tell us is as interesting as what it does. The deal value remains undisclosed, making it impossible to assess whether DigitalBridge and JEXI paid a premium or found a bargain. The facilities' utilization rates, contract durations, and customer concentration are similarly opaque — all factors that would materially affect investment returns.

We also don't know the governance structure. Is this a 50/50 joint venture, or does one party have operational control? Who's responsible for capital calls if power upgrades prove more expensive than expected? How are exit rights structured if one partner wants to sell in five years and the other doesn't?

Those details matter, particularly in cross-border partnerships where legal systems, business practices, and investment time horizons can differ significantly. The best infrastructure deals have clear, well-structured governance from day one. The messiest ones — ask anyone who's been through a contentious joint venture dispute — often involve partnerships where control and decision rights were ambiguous at signing.

For now, the deal stands as a signal of confidence in Japan's data center market and a bet that enterprise colocation demand will remain robust even as hyperscalers capture the majority of new capacity. Whether that confidence proves justified depends on factors largely outside the operators' control: government energy policy, utility investment in grid capacity, and the speed at which Japanese enterprises shift workloads in ways that require modern data center infrastructure.

The clearest near-term indicator of this deal's prospects will be customer retention. If NEC's existing enterprise clients remain under the new ownership and renew contracts as they come due, that validates the thesis that sticky customer relationships matter more than cutting-edge infrastructure in this segment. If clients defect to competitors or shift workloads to hyperscaler cloud, that suggests the facilities are caught in an awkward middle position.

Power procurement announcements will also signal the operators' strategic direction. If DigitalBridge and JEXI announce significant utility agreements or on-site generation projects in the next 12 to 18 months, that indicates a push toward density upgrades and AI-ready infrastructure. If they remain quiet on the power front, expect a steadier, lower-growth approach focused on operational efficiency and margin improvement.

Finally, watch for follow-on acquisitions. If this deal performs well, expect DigitalBridge to pursue similar carve-outs across Asia-Pacific, likely partnering with different local institutions in each market. If it struggles, the firm may retrench to more familiar markets where regulatory processes and customer behavior are better understood.

Japan's data center market is at an inflection point — caught between legacy infrastructure built for a previous era and demand patterns that require fundamentally different technical capabilities. The NEC acquisition is a bet that the gap between those two realities represents an investment opportunity. Whether it does depends on execution, timing, and factors that even the most sophisticated infrastructure investors can't fully control.