Doctrine, Dollars, and Deployment

If you only read one tech story this week, here's the line that sums it up:

“<em>Pope Leo XIV released his first encyclical — a 42,000-word document on safeguarding humanity in the AI age — and presented it alongside Anthropic co-founder Chris Olah. In the same seven-day window, Anthropic closed a roughly US$30 billion financing round at a US$900 billion+ valuation, surpassing OpenAI's private market value for the first time. Google I/O launched Gemini 3.5 Flash and a unified agent stack called Antigravity. And in Singapore, Nvidia opened its first lab on embodied AI, IMDA expanded SME support to 12,000 firms, and a refreshed Retail Industry Digital Plan rolled out for SME retailers.</em>”

Last week the story was that frontier labs were beginning to act like public companies and frontier models were beginning to act like research collaborators. The story this week is what happens when the rest of the world starts treating AI as a serious institutional question — the Vatican, the world's biggest LP cheques, and Southeast Asia's most active regulator all weighed in within the same seven days.

Here's what happened, why it matters, and what your business should do about it.

The Big Three

1. The Pope Wrote 42,000 Words on AI — and Brought Anthropic on Stage

On May 25, the Vatican formally released Magnifica Humanitas ("Magnificent Humanity"), Pope Leo XIV's first encyclical and the most sweeping papal statement ever issued on artificial intelligence. The text — signed on May 15, the 135th anniversary of Rerum Novarum — runs to roughly 42,300 words and urges governments, corporations, and individuals to slow the rate of AI development and keep these systems under meaningful ethical and political oversight.

The optics of the launch were unusual. The Vatican invited Christopher Olah, co-founder of Anthropic, to be one of the speakers at the presentation. A document that argues for restraint, presented alongside the co-founder of the lab that just became the world's most valuable private AI company. Both of those things are true at the same time, and the Vatican appears to be making that the point.

For SMEs, the encyclical itself is not a compliance document. But the signal is real. When the world's largest religious institution writes its most consequential text in years about AI, it confirms what the regulators in Singapore, Brussels, and Washington have already been saying: the assumption that AI policy is a niche concern for tech companies has expired. The conversation has moved into the same category as labour law, environmental policy, and financial regulation — and the businesses that handle it as a passing trend, rather than an emerging operating norm, will be on the wrong side of customer expectations within two to three years.

2. Anthropic's US$30B Round at a US$900B+ Valuation

The defining business story of the week. Bloomberg reported that Anthropic is expected to close a US$30 billion+ financing round at a pre-money valuation north of US$900 billion, with the round potentially wrapping the week of May 26. Sequoia Capital, Dragoneer, Altimeter, and Greenoaks are expected to co-lead, each writing checks of roughly US$2 billion. Existing investors including Founders Fund and General Catalyst are also in.

The milestone is competitive, not just numerical. At this valuation, Anthropic surpasses OpenAI's US$852 billion private market valuation from March — the first time it has done so. Coupled with the news from last week (Anthropic on track for its first quarterly operating profit on roughly US$10.9B in Q2 revenue, and openly discussing a year-end IPO), the takeaway is that the frontier AI market structure is rapidly resolving from "OpenAI vs everyone" into a genuine duopoly with a wide gap to third place.

Why this matters for an SME: the two vendors most likely to be embedded in the software your business runs on — through Microsoft 365, Google Workspace, QuickBooks, HubSpot, Canva, DocuSign, ChatGPT Enterprise, and Claude for Small Business — are now both on a trajectory to become public, profitable, audited mega-cap companies within 12–18 months. The argument that "the underlying vendor might disappear" is becoming much harder to defend. Vendor-risk is not zero, but it is no longer the right reason to delay an AI deployment.

3. Google I/O Made Agents the Default Product

Google I/O 2026 ran the week prior, with the news still rolling through enterprise channels into this week. Three releases matter for businesses:

• Gemini 3.5 Flash — Google's new flagship for agents and long-horizon coding tasks, generally available via the Gemini API, AI Studio, Android Studio, and Antigravity. Marketed as the first in a new series combining frontier intelligence with the ability to act.
• Antigravity — Google's unified agent development platform, now the single recommended environment for building Gemini agents, with direct integration into Google Cloud for enterprise customers.
• Gemini Spark — a 24/7 personal AI agent for Workspace and Gemini Enterprise customers that autonomously takes action on the user's behalf, under their direction.

Pair that with the earlier Anthropic launch of Claude for Small Business (Claude embedded directly inside QuickBooks, PayPal, HubSpot, Canva, DocuSign, Google Workspace, and Microsoft 365), and the pattern is unmistakable. Twelve months ago, "using AI at work" meant opening a chat window. By the end of this quarter, it will mean opening the software you already use and finding a competent agent already there. The shift from chatbots-as-a-product to agents-embedded-everywhere is not a future trend. It is the current shipping roadmap from both market leaders.

Closer to Home: Singapore Was Quietly Present in Every Storyline

Look at the same week through a Southeast Asia lens, and Singapore appears in almost every one of the global stories — by design, not coincidence.

Nvidia opened its first Singapore research lab. Announced at ATxSummit on May 20 and rolling through industry coverage this week, the lab focuses on embodied AI — robots and autonomous systems that perceive, reason, and act in the physical world, with application focus on intelligent inspection, autonomous assembly, and predictive maintenance. Nvidia Chief Scientist William Dally was present for the announcement. This is Nvidia's second research presence in Asia-Pacific.

A Center for Intelligent Robotics and a national robotics testbed are coming. Singapore announced a testbed later this year for companies to co-design, deploy, test, and validate commercially viable AI robotic technologies. Early users named include Certis, DHL, Grab, and QuikBot, with collaborations with Slamtec, Unitree, and QuikBot to trial use cases in food and parcel delivery, cleaning, and security patrolling.

IMDA expanded SME AI support to 12,000 firms. On May 21, IMDA announced new partnerships designed to support 12,000 SMEs in building digital and AI capabilities — a meaningful step up from the 10,000 SMEs targeted under the National AI Impact Programme. On May 26, Singapore also launched a refreshed Retail Industry Digital Plan (IDP) focused specifically on accelerating AI adoption among SME retailers.

A new National AI Strategy update. Singapore released an update to the National AI Strategy in May 2026 — 10 refreshed priorities, framed as a "double-click rather than a system reboot" on NAIS 2.0, and aligned with the National AI Council chaired by PM Lawrence Wong.

And across the Causeway, Malaysia's National AI Office (NAIO) is finalising its risk-based AI legislative framework for Cabinet submission in June — explicitly designed to reduce compliance costs and regulatory uncertainty for SMEs adopting AI. Malaysia's April electrical and electronics exports were up 46.4% year-on-year, primarily driven by AI-related demand. Singapore's electronic exports rose 66.7% in the same window.

If you stitch those threads together: the lab that builds the chips, the regulator that writes the SME framework, the testbed that validates the robots, and the export numbers that pay for it all — were all in the same news week, all within a two-hour flight of each other.

What This Means for Your Business

Three takeaways for SMEs in Singapore and Malaysia this week:

1. Take the Pope's framing seriously — even as a non-religious operator

The specific theology of Magnifica Humanitas is not the point. The structural point is that the expectation that you can articulate your AI choices to a non-technical audience — a customer, an employee, a regulator, a banker — is moving from "good to have" to "basic operating hygiene." Two practical steps:

1. Write a one-page AI Use Statement for your business — what you use AI for, what data goes in, what stays human-in-the-loop, who is accountable.
2. Make sure that one page is accurate, not aspirational. The most common failure mode in 2026 is an AI policy that describes a company that doesn't exist yet.

The IMDA Model AI Governance Framework for Agentic AI, updated earlier this month with 10+ real case studies from organisations like OCBC, GovTech, Ant International, and PwC, is a perfectly serviceable template to start from.

2. Plan as if Claude and Gemini will both be embedded inside everything you use by Q4

This week effectively confirms that the two market leaders are not slowing down on either funding or product velocity. Both are pushing agents directly into the SaaS stack — Anthropic via Claude for Small Business inside QuickBooks/HubSpot/Canva/DocuSign, Google via Gemini Spark inside Workspace and the new Antigravity stack for builders. The practical SME exercise this week:

• List the five software tools your business actually depends on.
• For each one, check the vendor's 2026 roadmap for native agent functionality.
• Identify one pre-built workflow per tool worth trialling against a real task in your business — payroll reconciliation, invoice chasing, lead nurture, contract follow-up, weekly social posts.
• Block a half-day this month to run the trial.

For most SG and MY SMEs, this is a Saturday-morning exercise, not a strategy project. The cost of waiting another quarter is no longer zero — your competitors are running the same trials.

3. If your business touches a physical operation, start watching embodied AI seriously

This is the genuinely new shift for the region. Nvidia's Singapore lab, the upcoming Center for Intelligent Robotics, and the named pilot partners (Certis for security, DHL for logistics, Grab for delivery, QuikBot for last-mile) tell you exactly which industries will see visible AI-driven operational change first: security, logistics, F&B delivery, cleaning, light manufacturing, and inspection.

If your SME operates in or sells into any of those sectors — directly or as a supplier — the right move this quarter is not to deploy a robot. It is to:

1. Identify the one repetitive physical or inspection task in your operation that costs you the most labour-hours per month.
2. Document it as a clean workflow (camera angle, inputs, decisions, exception handling).
3. Keep it ready. When the testbed partners start publishing pilot results in 6–12 months, you want a use case already scoped — not a brainstorming session.

The Practical Question

The headlines this week were a papal encyclical, a US$900B+ private valuation, a new flagship Google model, and Nvidia setting up a robotics lab two MRT stops away. The relevant question for most Singapore and Malaysian SMEs is, as always, smaller and more immediate:

"In the seven workflows that actually run my business, which one is a paying customer of mine willing to see automated — and which one would they prefer I never touch?"

That question doesn't require a model choice, a vendor selection, or a governance framework to answer. It requires a thirty-minute conversation with your three best customers. Have that conversation this month. The market structure for the rest of 2026 — embedded agents, embodied AI, billion-dollar vendor backing, and a regulator that has now told you what "good" looks like — is genuinely supportive of SMEs that move with that kind of clarity.

The Pope wants the industry to slow down. The investors want it to speed up. Singapore wants it to scale up — responsibly. None of those three is wrong. But the SME that wins this cycle is the one that has already picked the workflow worth handing over, and started.

At The Empyrean, we work with Singapore and Malaysia SMEs to find the practical, repeatable tasks where AI delivers value without disruption — and to set them up with the kind of governance that earns customer trust. If you're not sure where to start, we're happy to take a look at your operations and tell you honestly what would make sense.

Talk to us →