Pulse24
CD claude design prioritizes code integ ▲FAA faa advances ai air traffic system ▲AN abacus noir zero-copy wasm gpu infer ■MEITY meity establishes ai policy committe ■MEE malaysian experts highlight writing ■BOC governor flags ai cyber risks ▼ANTHR anthropic withholds mythos ai releas ▲FAA faa develops air traffic ai system ■SB advocates ai integration in hollywoo ■ANTHR anthropic valuation hits $800 billio ▲CP secures $18.5b out-licensing deal ▲ANTHR white house discusses mythos ai mode ▼
inPulse24 Tuesday Briefing
Edition #30 · 17–23 February 2026 · Read time ~5 min
Live · 23 Feb 2026
Tuesday Briefing/2 stories/4 signals

Capital Pledges and Safety Boundaries

India's AI summit week saw $210 billion in infrastructure pledges restated, but US defence procurement pressure may now test vendors' self-imposed safety limits.

Published23 Feb 2026
Coverage17 Feb 2026 – 23 Feb 2026
Stories tracked118
Featured2
AuthorPulse24 Desk
Last updated23 Feb 2026
This week’s pulse

In the India AI Impact Summit week, Reliance reiterated a ~$110 billion infrastructure plan and Adani restated a $100 billion data centre pledge (targeting 5GW by 2035). Across these two separate announcements, stated intent totals roughly $210 billion. Separately, 89 countries and international organisations endorsed the non-binding Delhi Declaration on equitable AI governance. The same week, the Pentagon threatened to label Anthropic a "supply chain risk" over its military use restrictions. The tension: sovereign compute ambitions demand government partnerships, but US defence procurement may require compromises on the safety commitments that built trust.

01

India's $210B AI Pledges Could Create a Regional Option for Specific Workloads

What happened

Reliance reiterated ~$110 billion in AI infrastructure plans and Adani restated a $100 billion data centre pledge (5GW by 2035). OpenAI announced a 100,000-student training partnership. Separately, 89 countries and international organisations endorsed the non-binding Delhi Declaration. Source

So what

If a meaningful share converts to operating capacity (Adani targets 2035), India could become a workable regional option for some buyers whose workloads are sensitive to latency, residency, or cost — but only where compliance requirements can be met by Indian providers.

The counter-case

Pledges are not capacity. India's data centre stock, grid reliability, and regulatory framework may delay execution by years. The $210 billion represents intent, not binding obligation.

Related signals

Heads of Infrastructure, Procurement leads, CFOs evaluating multi-region cloud strategy.

Action

If you run infrastructure procurement, ask your cloud partner for an India-region roadmap, certification posture, and power-resilience plan before shortlisting.

02

Pentagon's Anthropic Threat Puts Pressure on Safety-First Vendors

What happened

The Pentagon threatened to classify Anthropic as a "supply chain risk" over its restrictions on military use of Claude, according to Scientific American. Anthropic simultaneously launched Claude Code Security for automated vulnerability detection with strict human oversight. Source

So what

If enacted, a "supply chain risk" label would require prime contractors to certify their vendor stack — meaning Anthropic's restrictions could disqualify not just direct bids but any subcontractor relying on Claude.

The counter-case

The Pentagon has threatened vendor exclusion before without following through. Anthropic's enterprise revenue growth and expanding commercial partnerships (Infosys adopted Claude the same week) may reduce the impact of lost government revenue.

Related signals

CISOs, Procurement leads in government-adjacent organisations, Legal Ops teams managing AI vendor compliance.

Action

If you use Anthropic tools in government-adjacent work, confirm continuity clauses cover a vendor classification change and document a fallback provider.

So what

Sovereign AI ambitions and safety-first principles both claim to serve users, but one demands government partnerships while the other risks losing them. For vendors like Anthropic, the trade-off is direct: scale requires public capital, and public capital may demand compliance that erodes the safety commitments commercial buyers value.

---

📡 Signals

Worth tracking.

Markets
Vendor claim: Taalas launched a platform to create custom silicon for any AI model in two months, claiming 10-20x cost and performance gains over GPUs (not independently verified).Source ↗
Finance
Sequoia Capital is reportedly leading a $1 billion seed round in David Silver's Ineffable Intelligence at a reported $4 billion valuation.Source ↗
Risk
OpenAI flagged a school shooter's violent conversations before the attack but deemed them below the threshold for law enforcement referral under its policy at the time, according to Arkansas Online.Source ↗
Macro
FT analysis: US concentration of AI infrastructure risks creating a global compute divide, with emerging economies facing rising sovereignty risks.Source ↗
📊 Pulse check

The week by the numbers.

Stories tracked
100
Busiest category
13Product
OpenAI 5Anthropic 5Times 4Altman 3
🔭 The longer view

Trust and predictability are the new constraint.

Over the past 90 days, Pulse24 tracked four sovereign-linked compute funding and infrastructure commitments: UK £1 billion for sovereign compute (December), Blackstone $1.2 billion for Neysa's Indian GPUs (February 16), Saudi Humain $3 billion into xAI (February 18), and Reliance plus Adani's combined $210 billion in stated intent (summit week). Pledges are arriving faster than capacity can come online. The variable that determines whether buyers gain real alternatives is execution, not intent.

---

Pulse24’s view

This week's priority: protect optionality. Map vendor concentration, flag workloads portable to alternative providers by 2028, and verify that your AI vendor contracts include classification-change continuity clauses — because pledged capacity, if even partially delivered, could improve leverage in specific renewals.

👁 Forward watch

What we’re watching next.

4 March 2026
Apple product event — relevant to on-device AI and wearables strategy after Bloomberg reported Apple is developing a screenless AI hardware trio.Bloomberg, 22 February 2026
📚 References

Where this week’s evidence comes from.

India's $210B AI Pledges

Pentagon's Anthropic Threat

The Longer View