← Back to Hub
Macro Commentary • March 2026

AI and the Investment Landscape:
Navigating the Noise

Artificial intelligence is the defining technological story of our era. But it is one story among many — and history suggests the distance between genuine transformation and speculative bubble is rarely obvious in the moment.

Prepared by Wealth & Security Planners • March 2026 • Click each section to explore

🤔
Transformative
or overstated?
10
Structural
themes interacting
📉
The productivity
paradox
$400B
Hyperscaler debt
est. 2026
60/40
Still a sensible
anchor
🔄
Advice is not
forever
Section 1

The Disruption Question

The honest answer is: probably both, at different time horizons. Choose a lens below.

  • +AI capability benchmarks are improving at measurable, documented rates — not just in marketing materials. METR (March 2025) confirmed a ~7-month doubling time for AI task performance.
  • +White-collar professional tasks — coding, legal drafting, medical diagnosis, data analysis — are being automated meaningfully, with commercial adoption already underway.
  • +The infrastructure buildout is real: Amazon alone plans ~$200B in capital expenditure in 2026, vs $53B in 2023. This is not speculative — it is committed spend.
  • +First-mover cost advantages are compounding. Companies that integrated AI tooling in 2023–24 now operate at structurally lower unit costs than those that did not.
  • +History shows that genuinely disruptive technologies do eventually restructure labour markets — just not on the timelines initially claimed. The direction is rarely wrong; the timing almost always is.
  • +AI is accelerating drug discovery, materials science, and energy research — areas with multi-decade productivity implications that markets may be systematically underpricing.
  • ?AI adoption requires institutional trust, regulatory frameworks, liability clarity, and procurement budgets. None of these move at exponential speed. The adoption curve is not the capability curve.
  • ?Economy-wide productivity gains remain elusive — the “Solow paradox” of the computer age may be repeating. Enormous investment has not yet translated into measurable macro productivity improvement.
  • ?Many “AI layoffs” appear to be over-hiring corrections from 2020–22 disguised with a convenient narrative. Analysts suggest AI is the “justification layer for a cost structure reset,” not the direct cause.
  • ?Current models achieve ~50% reliability on complex professional tasks. In medicine, law, and finance, that failure rate is not adoption-ready — the liability alone forecloses it.
  • ?Enormous capital investment does not guarantee proportionate returns. The dot-com era proved this thoroughly — the Nasdaq fell 78% while the internet continued to transform the economy.
  • ?AI models are demonstrably impressive pattern-matchers. Serious open questions remain around genuine reasoning, contextual judgment, and reliability in novel situations — questions that matter for high-stakes domains.

WSP’s Working Position: AI is a genuine structural force. Most investors will be better served by attending to how it interacts with their existing portfolio exposures than by making dramatic allocation shifts based on narrative momentum.


The pattern-recognition instinct to “buy the theme” is precisely the instinct that most reliably destroys wealth in the early phases of a new technological paradigm. The railway mania of the 1840s, the electrification boom of the 1890s, and the dot-com era of 1999–2000 all confirm this.


Preparedness to be wrong about the timing matters more than confidence about the direction. We hold our views accordingly.

Section 2

The Bigger Picture: Ten Structural Themes

AI is the most-discussed investment theme of 2025–26. It is not the only one. Click any theme to explore it.

🧠
Artificial Intelligence
🌎
Geopolitics and Trade
⚡
Energy Transition
🛡
Defence Spiral
👨‍👩
Demographics
🔴
Polarisation
⚖
Income Disparity
🏛
Fiscal and Tax Policy
📈
Debt and Capital Markets
🌡
Climate and Physical Risk

⚠
On emphasis and humility: This list reflects our current reading. We hold it with appropriate scepticism. The history of macro investing is littered with confident thematic bets that were directionally correct but catastrophically wrong on timing — or correct at the theme level but invested in the wrong vehicle. Monitoring all of these themes simultaneously, without over-weighting any single narrative, is precisely the work portfolio construction exists to do.
Section 3

The Productivity Paradox and the Hype Cycle

Transformative technologies go through predictable phases. Capability and sentiment diverge. Productivity lags — then suddenly doesn’t. History is instructive.

Illustrative representation of Gartner’s Hype Cycle framework applied to AI (2022–present). Not to scale.

1840s — Railways
The Railway Mania
Genuinely world-changing. Also one of history’s great speculative bubbles. Click to read more.
Capital destruction was enormous — thousands of companies were floated, most failed. The rails themselves became the foundation of the industrial economy. The technology was entirely correct; the valuations were entirely wrong. The infrastructure outlasted every speculator who funded it.
✓ Survivors: Infrastructure holders, not speculators
1880s–1920s — Electricity
The Electrification Lag
Electricity available from the 1880s. Productivity gains didn’t show up for 30 years. Click to read more.
Factory managers initially retrofitted electricity into buildings designed for steam. The productivity gains only appeared when factories were redesigned from scratch around electric motors — small, distributed, precise. Technology alone does not create productivity. Process redesign does. This is the critical lesson for AI in the workplace.
✓ Survivors: Those who redesigned the process, not just the tool
1999–2000 — Internet
The Dot-Com Overshoot
The internet did reshape everything. The Nasdaq fell 78% from peak. Most internet companies failed. Click to read more.
Amazon, Google, and eBay — the survivors — often looked unremarkable, or even struggling, in 2002. The majority of the capital raised in 1999–2000 was permanently destroyed. The theme was entirely correct; approximately 95% of the investments were entirely wrong. Diversification across the theme, rather than concentrated bets, captured the upside while limiting the downside.
✓ Survivors: Amazon, Google, eBay — not the 95% that failed
2022–Present — AI
The Current Moment
Real capability gains. Significant valuation premia. The question is not whether AI matters — it does. Click to read more.
The question is whether current stock prices reflect transformation or expectation of transformation. And critically: which of today’s AI winners will still be winners in 2035? The VHS/Beta dynamic — where a technically inferior product dominates through distribution advantages — may well apply. Microsoft in 2002 looked like a mature, boring company. It was not.
⚠ Unknown: VHS or Beta? Microsoft or Netscape?

The Solow Paradox — Repeating?

Robert Solow observed in 1987 that “you can see the computer age everywhere except in the productivity statistics.” Despite enormous IT investment through the 1970s and 80s, measured productivity barely moved — until the late 1990s, when it suddenly did.

The lag between technology deployment and measurable productivity gain can be a decade or more. We may be early in that lag right now. The absence of visible gains is not evidence of failure — it may be evidence of immaturity.

The Geometric Underneath the Noise

What appears to “sneak up” was happening all along. The doubling time for AI capability benchmarks has been measured at approximately seven months (METR, March 2025). That compounding is invisible quarter-to-quarter and overwhelming over five years.

The investment implication: timing risk operates in both directions. Being too early is costly. Being too late is also costly. Managed, diversified exposure — rather than concentrated thematic bets — is the risk-adjusted response.

Section 4

Capital and Jobs: Two Sides of the Same Coin

Tens of thousands of jobs are being cut while hundreds of billions in new capital is being deployed. Understanding this is essential to reading the investment signal correctly.

Company / Entity Capital Raised / Committed Jobs Cut Notes
Amazon~$50B bond + $200B capex30,000+Largest non-acquisition bond raise on record
Alphabet (Google)$32B—Includes debut 100-year bond, Feb 2026
Oracle$25B—US dollar notes, Feb 2026
Meta$30B shelf—Debt shelf prospectus, Oct 2025
Block (Jack Dorsey)—4,000 (~40%)AI cited; analysts note over-hiring correction
WiseTech Global—2,000 (~29%)“Era of manually writing code is over”
Atlassian—1,600Includes CTO; “AI-first company” reframe
Workday—1,700 (~8.5%)Redirecting resources to AI investment
Commonwealth Bank$90M AI program300Simultaneous upskilling and cuts
Morgan Stanley Estimate$400B total hyperscaler 2026~9,200 of 45,000~20% of 2026 tech cuts tied to AI/restructure
The Paradox Explained: These two phenomena are not contradictory — they are the same process viewed from different angles. Large capital deployment funds the infrastructure that enables smaller workforces. Restructuring frees cash flow to service the debt raised for that infrastructure. Neither story, read in isolation, gives you the full picture.
A critical nuance: Independent analysts suggest that for many companies, AI is functioning as the “justification layer for a cost structure reset” rather than the direct cause of job losses. Companies that expanded aggressively in 2020–22 needed to correct their cost base regardless of AI. Discerning investors should hold this distinction clearly.
Section 5

Portfolio Resilience: What Actually Changes?

Given all of the above — the genuine transformation, the legitimate scepticism, the ten intersecting themes — how should a well-constructed portfolio respond? Probably with more discipline than drama.

60/40
The traditional 60% growth / 40% defensive construction is still a sensible starting point and benchmark — not because it is perfect, but because it embodies a genuine insight: that uncertainty across multiple simultaneous structural themes is best managed through diversification rather than conviction. Departing materially from this benchmark requires specific, evidence-based reasons — not narrative momentum.
📈

Market Returns vs. Selection Risk

The Wealth PyramidTM puts selection risk front and centre.
Selection risk — the risk of choosing the wrong individual security, manager, or theme — is the primary source of underperformance relative to market averages. AI theme enthusiasm creates intense selection risk pressure: the urge to pick the “winners” is strong, and the evidence that most individual pickers underperform broad index exposure is overwhelming.
Tap to expand ▼
📹

Portfolio Drift Monitoring

Thematic surges cause silent drift — without a single active decision.
A portfolio that was 10% technology in 2022 may now be 20%+ through simple price appreciation — without a single active decision having been made. This is not a narrative; it is arithmetic. Drift monitoring is not exciting. It is also not optional. An undisciplined portfolio that drifts toward the current hot theme is, in practice, making an active concentrated bet — whether or not it looks that way from the outside.
Tap to expand ▼
⚖

Rebalancing as Discipline

Systematic rebalancing is a mechanical “buy low, sell high.”
Selling what has run ahead of target weight and buying what has lagged is a mechanical embodiment of the oldest principle in investing. It works directly against the emotional pull of narrative momentum. It means selling some of what has been working (uncomfortable) and buying more of what has lagged (also uncomfortable). That discomfort is precisely why it works — most participants cannot bring themselves to do it consistently.
Tap to expand ▼

⚠ The AI Selection Risk Warning

In 2000, the Nasdaq lost 78%. Most internet stocks lost 90–100%. The theme was correct; the valuations were not. AI may be the most important technology of the next decade. That does not mean that current AI-related stock prices are rational.

🔎 What We Watch For

Meaningful economy-wide productivity gains in macro data. Regulatory clarity in financial services, healthcare, and law. Evidence that capex is translating into earnings, not just revenue. And crucially: which of today’s AI leaders are still leaders in 2030.

🏮 The Benefit of Breadth

A diversified global equities portfolio captures AI’s benefits regardless of which companies “win.” Microsoft, Amazon, Google, and Nvidia are all inside a broad index already. You do not need to pick the winner — that is the whole point of market-weight diversification.

🕑 Time Horizon Matters Enormously

Electrification took 30 years. The internet took 15. Investors with 5-year horizons are in a materially different position to those with 25-year horizons. Portfolio design must match your actual time horizon — not the narrative’s implied urgency.

Section 6

The Service Cube: Capacity, Capability and the Limits of Advice

No single piece of advice should be treated as permanent. The environment changes. So does our understanding of it. Click each dimension to explore.

👤
What You Can Do Yourself
Understanding themes, holding the framework, maintaining discipline
▼

The Wealth PyramidTM gives you the conceptual tools to evaluate claims — including claims made by AI tools themselves — without outsourcing your judgment entirely. Understanding whether a dramatic headline represents a genuine structural change or a narrative inflection is a learnable skill. The frameworks exist precisely to develop that capability over time.

🤝
Where Professional Input Adds Value
Portfolio construction, tax structuring, estate planning, insurance design
▼

These areas require professional expertise not because clients are incapable, but because the cost of errors is high and the complexity is genuine. Knowing when to delegate is itself a skill — and one that is increasingly valuable as AI creates a flood of plausible-sounding but unreliable financial commentary. The ability to distinguish genuine professional insight from AI-generated noise is an underrated competency.

🔄
Ongoing Review Is Not Optional
The Service CubeTM model exists because advice is not a one-time event
▼

Circumstances change. Markets change. Legislation changes. The AI landscape in 2028 will look materially different from 2026. The technology that is restructuring workplaces today may have restructured entire industries by the time your current plan matures. Your portfolio needs to reflect your changing reality, not a fixed plan from a fixed moment.

Advice is not forever. This is not a disclaimer — it is a design principle. The WSP approach is built around the recognition that the most dangerous financial position is one where a plan was made at a specific moment, for specific circumstances, and then left undisturbed as the world moved around it. AI, geopolitics, interest rates, your own life stage — all of these change. The portfolio should too. Not reactively, and not based on noise. But deliberately, and with clear eyes.

WSP Proprietary Frameworks Applied in This Analysis

The Wealth PyramidTM

Our framework for understanding the relationship between investment risk, market returns, and the selection risk that sits above them. The AI theme is primarily a selection risk question. Registered June 2002, renewed to 2032.

The Service CubeTM

Our model for how client engagement, professional input, and personal capability interact across the full financial planning lifecycle. AI is already beginning to reshape what clients can do themselves — and therefore what professional advice must do to add genuine value. Registered June 2002, renewed to 2032.

Important Information and Disclaimers

General Advice Warning: This document has been prepared for information and educational purposes only. It contains general financial information and does not take into account your personal financial situation, objectives, or needs. Before acting on any information, you should consider its appropriateness to your circumstances and, if necessary, seek professional advice.

No Recommendation: Nothing in this document constitutes a recommendation to buy, sell, or hold any financial product, security, or investment strategy. Past performance is not a reliable indicator of future performance.

Data Sources: Statistical data is drawn from publicly available sources including company announcements, financial media reporting, Morgan Stanley research estimates, and the METR research paper “Measuring AI Ability to Complete Long Tasks” (March 2025). Data current at March 2026.

Authorisation: WSP Pty Ltd as trustee for SFPS Unit Trust (ABN 14 135 004 947) is a Corporate Authorised Representative (CAR 276624) of AFD Pty Ltd (AFSL 344971). This document does not constitute independent financial advice as defined under the Corporations Act 2001 (Cth).