Part One of this series, Artificial Intelligence: A Balanced, Biblically-Grounded Perspective for the Faith-Based Investor, asked a foundational question: how should a Christian think about artificial intelligence at all? We traced its history from Alan Turing's early theories to the release of ChatGPT and the AI-saturated economy of 2026, weighed its genuine benefits against its genuine dangers, and grounded our posture toward it in Scripture - namely, that AI is a human tool, under the sovereignty of God, to be stewarded neither with fear nor with uncritical enthusiasm, but with discernment.
That paper answered the question of posture. This paper answers a narrower, more practical one: now that AI is embedded in the companies we own, the funds we hold, and the indexes that anchor most retirement accounts, what does that mean for the faith-based investor's actual portfolio?
The question matters because AI is no longer a side bet. As we will show below, AI-linked companies now account for a plurality of the value of the U.S. stock market, and the infrastructure being built to power AI (chips, data centers, electricity) appears to be absorbing capital at a pace with few historical precedents. Whether an investor has ever knowingly bought a single “AI stock” is almost beside the point; if they own an S&P 500 index fund, a target-date retirement fund, or nearly any diversified mutual fund, they already own a substantial and growing position in the AI economy.
“Know well the condition of your flocks, and give attention to your herds, for riches do not last forever, and does a crown endure to all generations?” — Proverbs 27:23–24
A shepherd who could not describe the state of his flock would not be considered faithful, however large the flock had grown. In the same way, a Christian investor whose portfolio is quietly saturated with AI exposure has a responsibility to understand that exposure - not to divest from it reflexively - but to know it, evaluate it, and steward it with open eyes.
This piece proceeds in four parts. First, we examine what the AI investment boom actually looks like in 2026, and what it means for portfolio risk and return. Second, we map the different ways AI shows up in a portfolio, from direct holdings to embedded exposure inside “ordinary” companies. Third, we highlight the specific issues (moral and financial) that investors should watch for in AI-exposed companies. Fourth, we assess which sectors face the greatest AI-related screening considerations, and how Inspire Insight applies its existing Biblically Responsible Investing framework to this fast-moving landscape.
Begin with a simple fact: the performance of the broad U.S. stock market in 2026 is, to a remarkable degree, the performance of a small group of AI-linked companies. By the end of 2025, the ten largest companies in the S&P 500, most of them AI infrastructure providers, chipmakers, or hyperscale cloud platforms, represented nearly 41% of the index's total weight, more than double their share just a decade earlier, according to RBC Wealth Management. Separate analyses that define “AI-exposed” more broadly put the figure even higher, with AI-linked mega-caps accounting for an estimated 40–45% of total S&P 500 market capitalization by mid-2026.
The effect on returns has been just as concentrated. From May 2024 through June 2026, the S&P 500 gained approximately 42%. One third-party analysis estimated that the gain would have been roughly less than half if excluding a specified group of AI-related stocks, illustrating the outsized contribution of AI-linked companies to the index’s performance, according to data reported by Yahoo Finance. The latter figure is methodology-dependent and is not an official S&P 500 calculation. RBC's analysts note that the market-capitalization-weighted index has outperformed its equal-weighted counterpart by roughly 32% over the past three years, a gap that now exceeds the comparable outperformance seen in the late 1990s, in the years just before the dot-com crash.¹
None of this means a repeat of 2000 is necessarily inevitable; markets can remain concentrated for long stretches, and the AI companies driving this cycle have, unlike many historical dot-com-era darlings, real revenue and real earnings. But it does mean that “the market” and “the AI trade” have become difficult to separate. An investor who believes they are diversified simply because they hold a broad index fund should understand that a significant share of that fund's value now depends on a single, unproven technological and economic thesis playing out roughly as expected.
The scale of capital being committed to AI infrastructure helps explain the concentration above. Estimates of 2026 capital expenditure among the largest hyperscale cloud providers - Amazon, Microsoft, Alphabet, Meta, and Oracle - range from roughly $630 billion to more than $800 billion, depending on methodology and which categories of spending are included, per Futurum Group and AL Capital Advisory. Whatever the precise figure, nearly every analysis agrees on the trajectory: this represents an increase of roughly 60% or more over 2025's already-record spending of approximately $388 billion.
Individually, Amazon has guided toward roughly $200 billion in 2026 capital expenditure, Alphabet toward $175–185 billion, Meta toward $115–135 billion, and Microsoft toward $120 billion or more, with the overwhelming majority of each figure tied to AI compute and data center buildout. To fund this, some of the largest technology companies in the world are now devoting a majority of their operating cash flow to capital expenditure: CreditSights estimates Amazon Web Services alone is directing 57% of its revenue toward capex, with Meta near 52% and Microsoft near 48%, levels of capital intensity that are, by historical standards for technology companies, without precedent.
This capital buildout has also produced a financing structure that has drawn increasing scrutiny: so-called “circular” AI deals, in which chipmakers invest directly in the AI labs that are also their largest customers, who in turn use that capital, along with additional debt and vendor financing, to purchase more chips and compute from the very companies that funded them. Nvidia, for example, has invested billions of dollars directly into OpenAI and Anthropic, both major purchasers of Nvidia chips, and reports through mid-2026 indicate Nvidia was weighing a financing arrangement that could backstop tens of billions of dollars of OpenAI's data center buildout, according to Axios. Microsoft, meanwhile, books much of OpenAI's compute spending as Azure revenue, and cloud providers more broadly have been financing AI startups that turn around and spend that same capital on cloud services.
Prominent market commentators have drawn direct comparisons to the vendor-financing arrangements among telecom equipment makers in the late 1990s, deals that boosted reported growth for a time before unraveling when customers could no longer pay, a parallel Jim Cramer has argued explicitly, as reported by CNBC. Others describe the current arrangement as a “virtuous circle” in which suppliers, builders, and customers are simply coordinating to meet a genuine, exploding demand for computing capacity.
For this paper, we do not attempt to adjudicate which view is correct; reasonable, well-informed people disagree, and the truth likely will not be clear until well after the fact. What we would highlight instead is the posture Scripture commends toward this kind of uncertainty.
“The plans of the diligent lead surely to abundance, but everyone who is hasty comes only to poverty.” — Proverbs 21:5
Diligence here does not mean predicting whether AI valuations are justified; no one can do that with confidence. It means understanding how concentrated one's own exposure has become, and making that concentration a deliberate choice rather than an accidental one.
Given the scale described above, it is worth mapping concretely where AI exposure actually appears in a typical investment portfolio. We would suggest four layers.
The most visible layer is the companies built specifically to design, manufacture, or deploy AI: semiconductor and AI-accelerator companies, hyperscale cloud providers whose earnings increasingly hinge on AI compute demand, and specialized AI-native software companies. Many of the most consequential AI labs, OpenAI and Anthropic among them, remain private, which means most retail and even institutional BRI investors cannot buy them directly; their influence instead reaches portfolios indirectly, through the public companies that hold equity stakes in them or supply them with chips and compute. The first semi-foray into this type of exposure would be SpaceX (SPCX), which launched in mid-2026 and has deep developmental ties through its close-knit subsidiary, xAI, which is the developer of the prominent AI platform Grok.
A second, fast-growing layer is thematic AI exchange-traded funds, which package dozens of AI-related holdings into a single ticker. These range from broad developed-market AI funds to narrowly targeted vehicles: some track a patent-based index of AI research and development activity, others apply a strict revenue-purity screen requiring that at least half of a holding's revenue come from generative AI, and still others concentrate on a single sub-theme, such as the memory chips AI models depend on, per ETF.com and Kiplinger. The largest of these thematic funds now manage tens of billions of dollars and have posted extraordinary short-term returns, alongside correspondingly higher volatility and concentration risk; expense ratios for these funds typically run from 0.35% to 0.75%, several multiples of a broad index fund. Investors should understand that these funds are constructed around a theme, not a Biblical screen, and that the individual holdings inside them have not necessarily been evaluated for the concerns discussed below.
The third, and largest, layer is the least visible: AI capability being embedded inside companies that are not “AI companies” in any conventional sense at all. A regional bank using AI for fraud detection, a hospital system using AI for diagnostic imaging, a retailer using AI for logistics, an insurer using AI for underwriting - none of these show up in an “AI fund”; many of them have a high probability of now carrying AI-related operational dependencies, opportunities, and risks. As adoption becomes close to universal across large public companies, the practical question for an investor is less “do I own AI?” and more “which of my holdings' use of AI deserves a closer look?”
The fourth layer is the one most investors give the least thought to: the default allocation of retirement accounts. Because AI-linked companies now represent such a large share of total market capitalization, any market-capitalization-weighted index fund (and by extension any target-date fund or default 401(k) allocation built on one) now carries a meaningfully larger AI concentration than it did even three years ago, without the investor making any active decision to increase it. This is the exposure layer most likely to go completely unexamined, and precisely for that reason, the one most worth a faithful steward's attention.
Part One documented, at a general level, the moral hazards that accompany AI: deepfake exploitation, the erosion of human dignity, and the temptation toward idolatry of innovation. This section translates those concerns into specific issues a Biblically Responsible investor should watch for at the company level, alongside several financial and stewardship-related risks that carry their own moral weight. It is important to note that while many of the following considerations are covered by Inspire Insight’s screening categories (like the Sexually Explicit, Exploitation, or State Owned Enterprise categories), some instances referenced are included for investor awareness purposes as commentary on the current AI landscape in general, rather than a full explanation of Inspire’s screening methodology. Additionally, the following commentary presents awareness of these subjects through the lens of biblical worldview and is not intended to offer any specific investment/screening advice for any company referenced.
Part One documented in detail how SpaceX's AI subsidiary, xAI, and its Grok platform contributed to a –81 Inspire Impact Score for SpaceX, driven in significant part by Sexually Explicit and Exploitation violations tied to Grok's image and video generation features. While this might be a ‘newer’ issue in the publicly traded ecosystem, this is not an isolated case in the grander AI landscape. As generative AI tools compete for user engagement, several major platforms have introduced “uncensored” or NSFW modes as a deliberate product and retention strategy, and a 2025 systematic review published in ScienceDirect documented the same pattern of AI-enabled, non-consensual content spreading across both dedicated and mainstream platforms, with victims reporting profound experiences of violation and psychological trauma. Investors should treat a company's design choices around content moderation, age verification, and “uncensored” feature toggles as material, not peripheral, to its Biblical screening, precisely because these are business decisions, not accidents.
AI companies have also accumulated significant, and in some cases enormous, legal liability tied to how their models were trained. In 2026, Anthropic agreed to pay authors $1.5 billion, covering more than 482,000 books at an implied rate of roughly $3,000 per work, to settle claims that it used pirated copies of copyrighted books to train Claude, one of the largest copyright settlements in U.S. history, according to Fortune. Litigation against OpenAI and others remains ongoing, with courts in early 2026 compelling the production of millions of anonymized user conversation logs as part of discovery, per Norton Rose Fulbright.
“You shall not steal.” — Exodus 20:15
“The wages of a hired worker shall not remain with you all night until the morning.” — Leviticus 19:13
These commands long predate generative AI, but they apply directly to how a company obtains the raw material, whether text, image, or labor, that it monetizes. Investors should treat unresolved IP and training-data litigation as a material, and specifically moral, balance-sheet risk, not merely a legal footnote.
The clearest economic disruption from AI so far has been to labor markets. Employers cited AI in an estimated 101,743 U.S. job cuts through the first half of 2026 alone, nearly double the 54,836 cited for all of 2025, and in May 2026, AI was cited in 40% of all announced layoffs that month, the highest monthly share on record, according to tracking compiled by Founder Reports. Roughly 41% of employers surveyed say they plan to reduce headcount because of AI automation, and research suggests as much as 80% of the U.S. workforce could see at least some portion of their tasks affected by large language models, with the heaviest exposure concentrated among college-educated, white-collar workers.
Scripture does not condemn efficiency or productivity gains; the Parable of the Talents commends the servant who multiplies what he has been given (Matthew 25:14–30). But it does insist that workers are image-bearers, not disposable inputs, and it reserves particular warning for those who profit while treating labor carelessly:
“Behold, the wages of the laborers who mowed your fields, which you kept back by fraud, are crying out against you.” — James 5:4
For the investor, this does not mean avoiding companies that use AI to improve productivity. It means paying attention to how a company manages the resulting workforce transitions - through retraining, redeployment, and honest communication, or through abrupt, indifferent cuts - as a real signal of corporate character, not a soft ESG afterthought.
AI-powered surveillance, facial recognition, and predictive-policing tools have proliferated globally. Researchers have documented public facial-recognition surveillance systems deployed in at least 78 countries, and human rights organizations have raised alarms about AI surveillance vendors selling directly to authoritarian governments, as reported by WebProNews and the Business & Human Rights Resource Centre. This is not a hypothetical, abstract-future concern; it is an active, present-day one, and it maps directly onto categories Inspire Insight already screens for. A company whose AI hardware, software, or data infrastructure is majority-owned by, or contracted to, a government with a documented record of human rights abuses may already trigger Inspire's existing State Owned Enterprise considerations, independent of anything specific to AI.
The concentration and circular-financing dynamics described above are, themselves, an issue worth watching at the individual holding level, not just the market level. A company whose revenue growth depends heavily on a small number of related-party customers, or whose largest customers are also its largest investors, carries a different risk profile than one with diversified, arm's-length revenue. Investors reviewing individual AI holdings should look for customer concentration disclosures and vendor-financing arrangements with the same scrutiny they would apply to any company with related-party risk.
Part One noted AI's promising applications in clean energy and climate modeling; the other side of that ledger deserves equal attention. Global data center electricity consumption is projected to nearly triple by 2030 to roughly 945 terawatt-hours (more than the combined annual electricity use of Pakistan, Bangladesh, and Nigeria, three nations home to more than 650 million people), while U.S. data center electricity use alone is projected to more than double, from 183 terawatt-hours in 2024 to 426 terawatt-hours by 2030, according to the World Economic Forum. The associated water footprint is comparably significant; by 2030, AI data centers could draw between roughly 731 and 1,125 million cubic meters of water annually, comparable to the household water use of six to ten million Americans, per the Lincoln Institute of Land Policy. These costs do not seem to be evenly distributed: while a new data center's tax revenue typically benefits only its host community, the electricity infrastructure required to serve it is often paid for across the entire regional service area, at a time when roughly 30% of U.S. households already report some form of energy insecurity.
“Do not withhold good from those to whom it is due, when it is in your power to do it.” — Proverbs 3:27–28
Genesis 2:15 calls humanity to “work” and “keep” the garden - cultivation, not extraction without regard for cost. A company's environmental and community engagement record around data center siting, water use, and local electricity rates is not a peripheral “green” concern; it is a direct expression of whether that company deals justly with its neighbors. Encouragingly, this is already precisely the kind of behavior Inspire's positive materiality categories - Energy Management, Water Conservation, GHG Emissions, and Environmental Risk Mitigation, among them - are designed to reward.
Not every sector carries the same AI-related considerations. The following is not an exhaustive list, but it highlights where the issues raised above concentrate most heavily.
This is, unsurprisingly, where AI exposure and AI-related screening and BRI investment considerations are most concentrated. Semiconductor and hyperscale cloud companies carry the labor, concentration, and energy-related considerations described above; consumer-facing AI platforms and social media companies carry the heaviest content, consent, and human-dignity considerations - precisely the categories that produced SpaceX's –81 Inspire Impact Score. Investors should expect this sector to require the most active, ongoing monitoring of any in the stock universe.
AI-enabled defense technology has become one of the fastest-growing corners of venture and government investment. Autonomous weapons and dual-use defense-technology startups raised more than $14.6 billion in the first five months of 2026 alone, already surpassing the previous full-year record, according to Tech Times, while total venture investment in defense and dual-use technology is on pace to exceed $18 billion for the year, and the U.S. Department of Defense requested $13.4 billion for autonomous weapons and systems in its fiscal 2026 budget. Additional reporting from OilPrice.com documents record participation by traditional defense contractors in this funding wave.
AI's healthcare applications represent some of its most unambiguously positive use cases, and investment reflects that: Grand View Research estimates the global AI-in-diagnostics market at $1.97 billion in 2025, projected to reach $9.68 billion by 2033, a compound annual growth rate of roughly 22%. The investment considerations here are narrower than in technology, but real: patient data privacy, informed consent for algorithmic diagnosis, and equitable access deserve continued attention as AI diagnostic tools scale.
AI adoption in banking, insurance, and asset management has moved from pilot to mainstream; roughly 65% of financial services firms report actively using AI as of early 2026, up from 45% a year earlier, concentrated heavily in fraud detection and credit-risk modeling, per the Cambridge Judge Business School. Because credit decisioning, insurance pricing, and investment suitability directly affect a customer's access to capital and fair treatment, regulators including the European Union have classified these AI applications as “high-risk,” requiring documented bias testing and human oversight. For the BRI investor, algorithmic fairness in lending and underwriting deserves the same scrutiny as any other fair-lending question.
As data center buildout accelerates, utilities, independent power producers, and data center-adjacent real estate have become AI-exposed in a way that has little to do with content or IP and everything to do with environmental and community stewardship considerations. This is a sector where positive screening - rewarding responsible energy sourcing, transparent community engagement, and water stewardship - is likely to matter more than negative screening in the years ahead.
Part One concluded that Inspire Insight would not create a new, standalone AI screening category. That conclusion holds, and the analysis above explains why: many of the AI-related concern raised in this paper already falls within Inspire's existing 14 negative screening categories and 26 positive materiality categories, from Sexually Explicit and Exploitation to State Owned Enterprise, Human Rights, Customer Privacy, Data Security, Ethical Labor Practices, Energy Management, and Water Conservation, among others, as detailed in Inspire's own Inside the Inspire Impact Score methodology.
What this means in practice is that Inspire Insight evaluates AI exposure the same way it evaluates every other business activity: by looking at what a specific company actually does, not by penalizing or excluding an entire sector because it participates in the AI economy. A semiconductor manufacturer that powers AI workloads is not screened negatively for existing in that supply chain; it is evaluated, like any company, on its labor practices, governance, environmental record, and the rest of Inspire's standard categories. An AI platform, by contrast, is evaluated specifically on what its tools are actually used to produce or enable, as SpaceX's Grok-driven Sexually Explicit and Exploitation violations demonstrate.
Inspire's research committee exists precisely to keep this evaluation current as the technology (and its risks) continue to move quickly. Investors can review any individual holding's current standing for free at inspireinsight.com. For the individual investor, we would offer three practical takeaways from this paper:
Part One of this series asked whether Christians should fear, fully embrace, or simply ignore artificial intelligence, and answered: none of the three. AI is a tool, under God's sovereignty, to be stewarded with discernment. This paper has attempted to make that discernment concrete: to show where AI actually lives inside a portfolio, what specific issues deserve an investor's attention, and how Inspire Insight's existing Biblically Responsible Investing framework already accounts for the great majority of them.
The picture that emerges is neither alarming nor complacent. AI is generating real economic value and, in areas like medicine and agriculture, real human flourishing. It is also concentrating market risk, unsettling labor markets, straining natural resources, and, in specific and documented cases, being used in ways that gravely violate the dignity of people made in God's image. A faithful steward does not need to resolve every open question about where AI is headed to act wisely today; what is required is the same posture Scripture has always commended: attentiveness to what has actually been entrusted to us.
As AI continues to reshape the companies, sectors, and indexes that make up the modern portfolio, Inspire Investing remains committed to that same faithfulness: knowing what our investors hold, evaluating it honestly against the whole counsel of Scripture, and stewarding it, in the fullest sense, for His glory rather than our own.
“Well done, good and faithful servant. You have been faithful over a little; I will set you over much. Enter into the joy of your master.” — Matthew 25:21
Bringing ten years of experience as a professor at a Christian college, Chris is deeply rooted in faith-based values. He excels as a faith-based writer, adept at crafting financial curriculum with spiritual depth. In his role, Chris is responsible for sourcing and evaluating reliable sources to accurately screen companies and assign them an Inspire Impact Score. Furthermore, he meticulously curates screening categories to maintain clarity and alignment with Inspire's faith-based mission.
This paper was prepared by Inspire Investing Research, August 2026. Scripture quotations are from the English Standard Version (ESV) unless otherwise noted.
References to specific companies, technologies, securities, or investment-related developments are provided to illustrate the subject matter discussed and should not be interpreted as a recommendation to buy, sell, or hold any security. All expressions of opinion are subject to change without notice. Investing involves risk, including the possible loss of principal.
Part One of this series, Artificial Intelligence: A Balanced, Biblically-Grounded Perspective for the Faith-Based Investor, asked a foundational question: how should a Christian think about artificial intelligence at all? We traced its history from Alan Turing's early theories to the release of ChatGPT and the AI-saturated economy of 2026, weighed its genuine benefits against its genuine dangers, and grounded our posture toward it in Scripture - namely, that AI is a human tool, under the sovereignty of God, to be stewarded neither with fear nor with uncritical enthusiasm, but with discernment.
That paper answered the question of posture. This paper answers a narrower, more practical one: now that AI is embedded in the companies we own, the funds we hold, and the indexes that anchor most retirement accounts, what does that mean for the faith-based investor's actual portfolio?
The question matters because AI is no longer a side bet. As we will show below, AI-linked companies now account for a plurality of the value of the U.S. stock market, and the infrastructure being built to power AI (chips, data centers, electricity) appears to be absorbing capital at a pace with few historical precedents. Whether an investor has ever knowingly bought a single “AI stock” is almost beside the point; if they own an S&P 500 index fund, a target-date retirement fund, or nearly any diversified mutual fund, they already own a substantial and growing position in the AI economy.
“Know well the condition of your flocks, and give attention to your herds, for riches do not last forever, and does a crown endure to all generations?” — Proverbs 27:23–24
A shepherd who could not describe the state of his flock would not be considered faithful, however large the flock had grown. In the same way, a Christian investor whose portfolio is quietly saturated with AI exposure has a responsibility to understand that exposure - not to divest from it reflexively - but to know it, evaluate it, and steward it with open eyes.
This piece proceeds in four parts. First, we examine what the AI investment boom actually looks like in 2026, and what it means for portfolio risk and return. Second, we map the different ways AI shows up in a portfolio, from direct holdings to embedded exposure inside “ordinary” companies. Third, we highlight the specific issues (moral and financial) that investors should watch for in AI-exposed companies. Fourth, we assess which sectors face the greatest AI-related screening considerations, and how Inspire Insight applies its existing Biblically Responsible Investing framework to this fast-moving landscape.
Begin with a simple fact: the performance of the broad U.S. stock market in 2026 is, to a remarkable degree, the performance of a small group of AI-linked companies. By the end of 2025, the ten largest companies in the S&P 500, most of them AI infrastructure providers, chipmakers, or hyperscale cloud platforms, represented nearly 41% of the index's total weight, more than double their share just a decade earlier, according to RBC Wealth Management. Separate analyses that define “AI-exposed” more broadly put the figure even higher, with AI-linked mega-caps accounting for an estimated 40–45% of total S&P 500 market capitalization by mid-2026.
The effect on returns has been just as concentrated. From May 2024 through June 2026, the S&P 500 gained approximately 42%. One third-party analysis estimated that the gain would have been roughly less than half if excluding a specified group of AI-related stocks, illustrating the outsized contribution of AI-linked companies to the index’s performance, according to data reported by Yahoo Finance. The latter figure is methodology-dependent and is not an official S&P 500 calculation. RBC's analysts note that the market-capitalization-weighted index has outperformed its equal-weighted counterpart by roughly 32% over the past three years, a gap that now exceeds the comparable outperformance seen in the late 1990s, in the years just before the dot-com crash.¹
None of this means a repeat of 2000 is necessarily inevitable; markets can remain concentrated for long stretches, and the AI companies driving this cycle have, unlike many historical dot-com-era darlings, real revenue and real earnings. But it does mean that “the market” and “the AI trade” have become difficult to separate. An investor who believes they are diversified simply because they hold a broad index fund should understand that a significant share of that fund's value now depends on a single, unproven technological and economic thesis playing out roughly as expected.
The scale of capital being committed to AI infrastructure helps explain the concentration above. Estimates of 2026 capital expenditure among the largest hyperscale cloud providers - Amazon, Microsoft, Alphabet, Meta, and Oracle - range from roughly $630 billion to more than $800 billion, depending on methodology and which categories of spending are included, per Futurum Group and AL Capital Advisory. Whatever the precise figure, nearly every analysis agrees on the trajectory: this represents an increase of roughly 60% or more over 2025's already-record spending of approximately $388 billion.
Individually, Amazon has guided toward roughly $200 billion in 2026 capital expenditure, Alphabet toward $175–185 billion, Meta toward $115–135 billion, and Microsoft toward $120 billion or more, with the overwhelming majority of each figure tied to AI compute and data center buildout. To fund this, some of the largest technology companies in the world are now devoting a majority of their operating cash flow to capital expenditure: CreditSights estimates Amazon Web Services alone is directing 57% of its revenue toward capex, with Meta near 52% and Microsoft near 48%, levels of capital intensity that are, by historical standards for technology companies, without precedent.
This capital buildout has also produced a financing structure that has drawn increasing scrutiny: so-called “circular” AI deals, in which chipmakers invest directly in the AI labs that are also their largest customers, who in turn use that capital, along with additional debt and vendor financing, to purchase more chips and compute from the very companies that funded them. Nvidia, for example, has invested billions of dollars directly into OpenAI and Anthropic, both major purchasers of Nvidia chips, and reports through mid-2026 indicate Nvidia was weighing a financing arrangement that could backstop tens of billions of dollars of OpenAI's data center buildout, according to Axios. Microsoft, meanwhile, books much of OpenAI's compute spending as Azure revenue, and cloud providers more broadly have been financing AI startups that turn around and spend that same capital on cloud services.
Prominent market commentators have drawn direct comparisons to the vendor-financing arrangements among telecom equipment makers in the late 1990s, deals that boosted reported growth for a time before unraveling when customers could no longer pay, a parallel Jim Cramer has argued explicitly, as reported by CNBC. Others describe the current arrangement as a “virtuous circle” in which suppliers, builders, and customers are simply coordinating to meet a genuine, exploding demand for computing capacity.
For this paper, we do not attempt to adjudicate which view is correct; reasonable, well-informed people disagree, and the truth likely will not be clear until well after the fact. What we would highlight instead is the posture Scripture commends toward this kind of uncertainty.
“The plans of the diligent lead surely to abundance, but everyone who is hasty comes only to poverty.” — Proverbs 21:5
Diligence here does not mean predicting whether AI valuations are justified; no one can do that with confidence. It means understanding how concentrated one's own exposure has become, and making that concentration a deliberate choice rather than an accidental one.
Given the scale described above, it is worth mapping concretely where AI exposure actually appears in a typical investment portfolio. We would suggest four layers.
The most visible layer is the companies built specifically to design, manufacture, or deploy AI: semiconductor and AI-accelerator companies, hyperscale cloud providers whose earnings increasingly hinge on AI compute demand, and specialized AI-native software companies. Many of the most consequential AI labs, OpenAI and Anthropic among them, remain private, which means most retail and even institutional BRI investors cannot buy them directly; their influence instead reaches portfolios indirectly, through the public companies that hold equity stakes in them or supply them with chips and compute. The first semi-foray into this type of exposure would be SpaceX (SPCX), which launched in mid-2026 and has deep developmental ties through its close-knit subsidiary, xAI, which is the developer of the prominent AI platform Grok.
A second, fast-growing layer is thematic AI exchange-traded funds, which package dozens of AI-related holdings into a single ticker. These range from broad developed-market AI funds to narrowly targeted vehicles: some track a patent-based index of AI research and development activity, others apply a strict revenue-purity screen requiring that at least half of a holding's revenue come from generative AI, and still others concentrate on a single sub-theme, such as the memory chips AI models depend on, per ETF.com and Kiplinger. The largest of these thematic funds now manage tens of billions of dollars and have posted extraordinary short-term returns, alongside correspondingly higher volatility and concentration risk; expense ratios for these funds typically run from 0.35% to 0.75%, several multiples of a broad index fund. Investors should understand that these funds are constructed around a theme, not a Biblical screen, and that the individual holdings inside them have not necessarily been evaluated for the concerns discussed below.
The third, and largest, layer is the least visible: AI capability being embedded inside companies that are not “AI companies” in any conventional sense at all. A regional bank using AI for fraud detection, a hospital system using AI for diagnostic imaging, a retailer using AI for logistics, an insurer using AI for underwriting - none of these show up in an “AI fund”; many of them have a high probability of now carrying AI-related operational dependencies, opportunities, and risks. As adoption becomes close to universal across large public companies, the practical question for an investor is less “do I own AI?” and more “which of my holdings' use of AI deserves a closer look?”
The fourth layer is the one most investors give the least thought to: the default allocation of retirement accounts. Because AI-linked companies now represent such a large share of total market capitalization, any market-capitalization-weighted index fund (and by extension any target-date fund or default 401(k) allocation built on one) now carries a meaningfully larger AI concentration than it did even three years ago, without the investor making any active decision to increase it. This is the exposure layer most likely to go completely unexamined, and precisely for that reason, the one most worth a faithful steward's attention.
Part One documented, at a general level, the moral hazards that accompany AI: deepfake exploitation, the erosion of human dignity, and the temptation toward idolatry of innovation. This section translates those concerns into specific issues a Biblically Responsible investor should watch for at the company level, alongside several financial and stewardship-related risks that carry their own moral weight. It is important to note that while many of the following considerations are covered by Inspire Insight’s screening categories (like the Sexually Explicit, Exploitation, or State Owned Enterprise categories), some instances referenced are included for investor awareness purposes as commentary on the current AI landscape in general, rather than a full explanation of Inspire’s screening methodology. Additionally, the following commentary presents awareness of these subjects through the lens of biblical worldview and is not intended to offer any specific investment/screening advice for any company referenced.
Part One documented in detail how SpaceX's AI subsidiary, xAI, and its Grok platform contributed to a –81 Inspire Impact Score for SpaceX, driven in significant part by Sexually Explicit and Exploitation violations tied to Grok's image and video generation features. While this might be a ‘newer’ issue in the publicly traded ecosystem, this is not an isolated case in the grander AI landscape. As generative AI tools compete for user engagement, several major platforms have introduced “uncensored” or NSFW modes as a deliberate product and retention strategy, and a 2025 systematic review published in ScienceDirect documented the same pattern of AI-enabled, non-consensual content spreading across both dedicated and mainstream platforms, with victims reporting profound experiences of violation and psychological trauma. Investors should treat a company's design choices around content moderation, age verification, and “uncensored” feature toggles as material, not peripheral, to its Biblical screening, precisely because these are business decisions, not accidents.
AI companies have also accumulated significant, and in some cases enormous, legal liability tied to how their models were trained. In 2026, Anthropic agreed to pay authors $1.5 billion, covering more than 482,000 books at an implied rate of roughly $3,000 per work, to settle claims that it used pirated copies of copyrighted books to train Claude, one of the largest copyright settlements in U.S. history, according to Fortune. Litigation against OpenAI and others remains ongoing, with courts in early 2026 compelling the production of millions of anonymized user conversation logs as part of discovery, per Norton Rose Fulbright.
“You shall not steal.” — Exodus 20:15
“The wages of a hired worker shall not remain with you all night until the morning.” — Leviticus 19:13
These commands long predate generative AI, but they apply directly to how a company obtains the raw material, whether text, image, or labor, that it monetizes. Investors should treat unresolved IP and training-data litigation as a material, and specifically moral, balance-sheet risk, not merely a legal footnote.
The clearest economic disruption from AI so far has been to labor markets. Employers cited AI in an estimated 101,743 U.S. job cuts through the first half of 2026 alone, nearly double the 54,836 cited for all of 2025, and in May 2026, AI was cited in 40% of all announced layoffs that month, the highest monthly share on record, according to tracking compiled by Founder Reports. Roughly 41% of employers surveyed say they plan to reduce headcount because of AI automation, and research suggests as much as 80% of the U.S. workforce could see at least some portion of their tasks affected by large language models, with the heaviest exposure concentrated among college-educated, white-collar workers.
Scripture does not condemn efficiency or productivity gains; the Parable of the Talents commends the servant who multiplies what he has been given (Matthew 25:14–30). But it does insist that workers are image-bearers, not disposable inputs, and it reserves particular warning for those who profit while treating labor carelessly:
“Behold, the wages of the laborers who mowed your fields, which you kept back by fraud, are crying out against you.” — James 5:4
For the investor, this does not mean avoiding companies that use AI to improve productivity. It means paying attention to how a company manages the resulting workforce transitions - through retraining, redeployment, and honest communication, or through abrupt, indifferent cuts - as a real signal of corporate character, not a soft ESG afterthought.
AI-powered surveillance, facial recognition, and predictive-policing tools have proliferated globally. Researchers have documented public facial-recognition surveillance systems deployed in at least 78 countries, and human rights organizations have raised alarms about AI surveillance vendors selling directly to authoritarian governments, as reported by WebProNews and the Business & Human Rights Resource Centre. This is not a hypothetical, abstract-future concern; it is an active, present-day one, and it maps directly onto categories Inspire Insight already screens for. A company whose AI hardware, software, or data infrastructure is majority-owned by, or contracted to, a government with a documented record of human rights abuses may already trigger Inspire's existing State Owned Enterprise considerations, independent of anything specific to AI.
The concentration and circular-financing dynamics described above are, themselves, an issue worth watching at the individual holding level, not just the market level. A company whose revenue growth depends heavily on a small number of related-party customers, or whose largest customers are also its largest investors, carries a different risk profile than one with diversified, arm's-length revenue. Investors reviewing individual AI holdings should look for customer concentration disclosures and vendor-financing arrangements with the same scrutiny they would apply to any company with related-party risk.
Part One noted AI's promising applications in clean energy and climate modeling; the other side of that ledger deserves equal attention. Global data center electricity consumption is projected to nearly triple by 2030 to roughly 945 terawatt-hours (more than the combined annual electricity use of Pakistan, Bangladesh, and Nigeria, three nations home to more than 650 million people), while U.S. data center electricity use alone is projected to more than double, from 183 terawatt-hours in 2024 to 426 terawatt-hours by 2030, according to the World Economic Forum. The associated water footprint is comparably significant; by 2030, AI data centers could draw between roughly 731 and 1,125 million cubic meters of water annually, comparable to the household water use of six to ten million Americans, per the Lincoln Institute of Land Policy. These costs do not seem to be evenly distributed: while a new data center's tax revenue typically benefits only its host community, the electricity infrastructure required to serve it is often paid for across the entire regional service area, at a time when roughly 30% of U.S. households already report some form of energy insecurity.
“Do not withhold good from those to whom it is due, when it is in your power to do it.” — Proverbs 3:27–28
Genesis 2:15 calls humanity to “work” and “keep” the garden - cultivation, not extraction without regard for cost. A company's environmental and community engagement record around data center siting, water use, and local electricity rates is not a peripheral “green” concern; it is a direct expression of whether that company deals justly with its neighbors. Encouragingly, this is already precisely the kind of behavior Inspire's positive materiality categories - Energy Management, Water Conservation, GHG Emissions, and Environmental Risk Mitigation, among them - are designed to reward.
Not every sector carries the same AI-related considerations. The following is not an exhaustive list, but it highlights where the issues raised above concentrate most heavily.
This is, unsurprisingly, where AI exposure and AI-related screening and BRI investment considerations are most concentrated. Semiconductor and hyperscale cloud companies carry the labor, concentration, and energy-related considerations described above; consumer-facing AI platforms and social media companies carry the heaviest content, consent, and human-dignity considerations - precisely the categories that produced SpaceX's –81 Inspire Impact Score. Investors should expect this sector to require the most active, ongoing monitoring of any in the stock universe.
AI-enabled defense technology has become one of the fastest-growing corners of venture and government investment. Autonomous weapons and dual-use defense-technology startups raised more than $14.6 billion in the first five months of 2026 alone, already surpassing the previous full-year record, according to Tech Times, while total venture investment in defense and dual-use technology is on pace to exceed $18 billion for the year, and the U.S. Department of Defense requested $13.4 billion for autonomous weapons and systems in its fiscal 2026 budget. Additional reporting from OilPrice.com documents record participation by traditional defense contractors in this funding wave.
AI's healthcare applications represent some of its most unambiguously positive use cases, and investment reflects that: Grand View Research estimates the global AI-in-diagnostics market at $1.97 billion in 2025, projected to reach $9.68 billion by 2033, a compound annual growth rate of roughly 22%. The investment considerations here are narrower than in technology, but real: patient data privacy, informed consent for algorithmic diagnosis, and equitable access deserve continued attention as AI diagnostic tools scale.
AI adoption in banking, insurance, and asset management has moved from pilot to mainstream; roughly 65% of financial services firms report actively using AI as of early 2026, up from 45% a year earlier, concentrated heavily in fraud detection and credit-risk modeling, per the Cambridge Judge Business School. Because credit decisioning, insurance pricing, and investment suitability directly affect a customer's access to capital and fair treatment, regulators including the European Union have classified these AI applications as “high-risk,” requiring documented bias testing and human oversight. For the BRI investor, algorithmic fairness in lending and underwriting deserves the same scrutiny as any other fair-lending question.
As data center buildout accelerates, utilities, independent power producers, and data center-adjacent real estate have become AI-exposed in a way that has little to do with content or IP and everything to do with environmental and community stewardship considerations. This is a sector where positive screening - rewarding responsible energy sourcing, transparent community engagement, and water stewardship - is likely to matter more than negative screening in the years ahead.
Part One concluded that Inspire Insight would not create a new, standalone AI screening category. That conclusion holds, and the analysis above explains why: many of the AI-related concern raised in this paper already falls within Inspire's existing 14 negative screening categories and 26 positive materiality categories, from Sexually Explicit and Exploitation to State Owned Enterprise, Human Rights, Customer Privacy, Data Security, Ethical Labor Practices, Energy Management, and Water Conservation, among others, as detailed in Inspire's own Inside the Inspire Impact Score methodology.
What this means in practice is that Inspire Insight evaluates AI exposure the same way it evaluates every other business activity: by looking at what a specific company actually does, not by penalizing or excluding an entire sector because it participates in the AI economy. A semiconductor manufacturer that powers AI workloads is not screened negatively for existing in that supply chain; it is evaluated, like any company, on its labor practices, governance, environmental record, and the rest of Inspire's standard categories. An AI platform, by contrast, is evaluated specifically on what its tools are actually used to produce or enable, as SpaceX's Grok-driven Sexually Explicit and Exploitation violations demonstrate.
Inspire's research committee exists precisely to keep this evaluation current as the technology (and its risks) continue to move quickly. Investors can review any individual holding's current standing for free at inspireinsight.com. For the individual investor, we would offer three practical takeaways from this paper:
Part One of this series asked whether Christians should fear, fully embrace, or simply ignore artificial intelligence, and answered: none of the three. AI is a tool, under God's sovereignty, to be stewarded with discernment. This paper has attempted to make that discernment concrete: to show where AI actually lives inside a portfolio, what specific issues deserve an investor's attention, and how Inspire Insight's existing Biblically Responsible Investing framework already accounts for the great majority of them.
The picture that emerges is neither alarming nor complacent. AI is generating real economic value and, in areas like medicine and agriculture, real human flourishing. It is also concentrating market risk, unsettling labor markets, straining natural resources, and, in specific and documented cases, being used in ways that gravely violate the dignity of people made in God's image. A faithful steward does not need to resolve every open question about where AI is headed to act wisely today; what is required is the same posture Scripture has always commended: attentiveness to what has actually been entrusted to us.
As AI continues to reshape the companies, sectors, and indexes that make up the modern portfolio, Inspire Investing remains committed to that same faithfulness: knowing what our investors hold, evaluating it honestly against the whole counsel of Scripture, and stewarding it, in the fullest sense, for His glory rather than our own.
“Well done, good and faithful servant. You have been faithful over a little; I will set you over much. Enter into the joy of your master.” — Matthew 25:21
