Navigating AI in Credit: How to Find the Winners in Tech

Artificial intelligence is moving at breakneck speed. For credit investors, the challenge isn't chasing every headline—it's understanding how AI reshapes business models, cash flows and ultimately, creditworthiness.

This month, with insights from Brad Lewis, CFA, technology sector analyst on Barings' global high yield team, we're covering:

1. How we got to this moment in AI
2. The framework Barings uses to parse winners and losers in credit markets
3. What investors may want to watch next

Let's get into it!

The Speed Read:

• AI's acceleration triggered a market inflection, with equities reacting first and credit following as terminal values came into question.
• Volatility reflects uncertainty around long-term business models, not near-term earnings.
• Mission-critical software is far more resilient than point solutions as AI lowers barriers to replication.
• Dispersion is increasing, creating both risk and opportunity for credit investors.
• Process matters: disciplined underwriting and fundamentals matter more than headlines in a rapidly evolving market.

1. AI's Evolution—And Why Markets Hit the Panic Button

AI has been decades in the making. Its roots trace back to the 1950s, with steady progress since then in machine learning, automation and analytics. For much of that time, advances were incremental.

November 2022 changed the tone. The release of ChatGPT marked a clear inflection point, revealing capabilities that felt fundamentally different.

From there, momentum accelerated:

• 2024–2025: Hyperscalers sharply ramped capital spending to build AI infrastructure.
• Leading model providers raised capital at increasingly large valuations, capturing investor attention.
• By 2025, the question was no longer if AI would materially change the landscape for credit investors but just how fast that change would happen.

The Step Change: From Tools to Autonomous Work

The real shift came in late 2025 and early 2026, with the rise of agentic AI—systems capable of performing tasks autonomously rather than simply responding to prompts.

What had been framed as productivity-enhancing tools began to look like technologies that could meaningfully alter workflows and cost structures.

From Equity Volatility to Credit Repricing

As concerns shifted toward long-term business model durability, equity markets moved sharply lower.

Software equities sold off ~37% from peak to trough¹, driven by reassessments of:

• Long-term revenue growth and profitability
• Competitive positioning
• Terminal values

That reassessment didn't stay contained.

Equity volatility bled into credit markets, where terminal values play a central role in underwriting.

Investors began a broad-based repricing of risk.

Technical factors added pressure:

• BDC repositioning and selling in the loan market
• CLOs reducing exposure to lower-rated credits

While high yield bonds had relatively limited software exposure, the loan market absorbed most of the pressure, reflecting its heavier concentration in private equity backed software issuers.

In Short: What began as an equity-market reaction to the speed of AI's evolution quickly became a broader credit market recalibration—marking the moment AI shifted from a long-term story to a near-term risk factor investors could no longer ignore.

2. Separating Winners from Losers: The Moats of Core Software

Fast forward to today, and the key challenge for credit investors is parsing durable software models from those more vulnerable to disruption. This distinction often comes down to which companies have defensive business models, or “moats,” and which don’t.

The Three Moats:

1. Mission-critical software

These are systems that enterprises cannot operate without, meaning:

• If the software fails, the business doesn't run
• The cost of failure is high
• Switching is painful, risky and expensive

Think: general ledgers or enterprise-wide systems of record that are embedded across large portions of an organization. These systems face materially lower replacement risk than standalone applications.

2. Proprietary data and deep domain expertise

These are platforms with decades of accumulated data and embedded industry knowledge that is extraordinarily difficult to replace.

• Many platforms have built this data advantage over 20–30+ years
• The data itself can't be scraped or rebuilt
• When AI is layered on top, these advantages often compound

3. Network effects

This refers to platforms that become more valuable as usage grows.

• Greater adoption reinforces dominance
• Switching costs rise as ecosystems deepen
• Displacement becomes increasingly difficult

Bloomberg is a classic example: the more participants rely on it, the more entrenched it becomes for both the platform and its users.

Key Takeaway: The winning strategy is mission-critical software coupled with AI capabilities

Discipline Over Headlines

There is not a shortage of AI-related headlines, and they're often overstated. While they can drive short-term price moves, they don't always translate into lasting insight.

Rather than reacting to every headline, it is critical to rely on a repeatable, disciplined process designed to hold up across market environments.

The Process:

A repeatable process ensures:

• A consistent framework is applied across both calm and stressed conditions
• Fundamental credit analysis is prioritized over narrative momentum
• Decisions are anchored in history and context rather than fear-driven extrapolation

Risks are continuously reassessed as new information emerges, without abandoning discipline or long-established underwriting standards.

3. Where AI Goes Next: What Investors Should Watch

Amid a sea of headlines and data points, three factors are critical to watch in the months and years ahead:

1. Hyperscaler and frontier model spending trends

• Spending on AI infrastructure has surged
• Spending levels have so far been positively correlated with software volatility in capital markets

If spending on compute and infrastructure continues to rise at exponential rates, increased software disruption becomes even more likely.

2. Greater transparency through public markets

Large language model providers remain opaque. Valuations and capabilities are widely discussed, but business models, cost structures and returns on capital are far less visible.

That could change:

• Potential IPOs of major AI model providers would introduce public market disclosure
• Greater transparency would bring clearer insight into monetization, margins and capital intensity
• Over time, this could help demystify the broader AI ecosystem for investors

For credit markets, increased visibility matters. Understanding who ultimately earns the economics—and at what cost —is essential to underwriting risk in an AI-driven environment.

3. The pace of AI tool innovation

Model releases are frequent and eye-catching. The pace of improvement alone can drive short-term market reactions.

But our focus is elsewhere:

• How are software companies embedding AI into existing products?
• Are new capabilities driving incremental revenue or simply raising costs?
• Do AI features strengthen customer relationships and defensibility, or commoditize functionality?
• And critically: do the economics improve or deteriorate as AI is scaled?

AI doesn't create value on its own. Value depends on how AI is deployed and whether that results in durable returns rather than transient excitement.

Bottom Line:

AI will almost certainly be disruptive, but disruption does not mean uniform destruction. We expect greater dispersion, not wholesale decline.

That creates risk. It also creates opportunity.

For credit investors, success will come down to fundamentals: staying diversified, maintaining discipline and being selective about which companies are positioned to adapt.

Go Deeper:

Listen/Watch: Managing AI Risk in Credit Portfolios

Explore: Barings perspectives across credit and portfolio construction

Subscribe: Follow Barings on LinkedIn and subscribe to Where Credit is Due

The Credits:

Brad Lewis, CFA: Technology Sector Analyst, Senior Director, High Yield

Image

Brad Lewis, CFA: Technology Sector Analyst, Senior Director, High Yield

• 15 years at Barings covering technology
• Wake Forest University graduate
• Brad and his wife have three girls and love to coach softball together

READ MORE FROM BARINGS

Source:
1) Bloomberg, IGV ETF Index. As of September 23, 2025 - April 10, 2026.

Compliance Code: 26-5454788

This commentary is for informational purposes only and does not constitute an offer or solicitation to purchase or sell any financial instrument or service in any jurisdiction. It was prepared without consideration of the investment objectives, financial situation, or particular needs of any recipient. This commentary is not, and must not be treated as, investment advice, an investment recommendation, investment research, or a recommendation regarding the suitability or appropriateness of any security, investment, or investment strategy. It must not be construed as a projection or prediction.