Does AI spending push up interest rates? Three channels, and what reaches the Philippines

Interest rates are usually explained by central banks and inflation. In 2026 a third explanation keeps appearing in bond market commentary: the enormous sums that technology companies are spending on artificial intelligence, much of it financed with debt. Here is the reasoning, the evidence on both sides, and what part of it reaches the Philippines.

Through which channels can AI raise rates?

Visual guide

Through which channels can AI raise rates?

1. Supply

   How it works

   Tech companies issue large amounts of bonds to fund data centres, competing for the same savings as governments

   What to look for

   Wider corporate spreads, yields rising around big deals

2. Investment demand

   How it works

   A wave of investment raises the return on capital, which lifts the neutral interest rate

   What to look for

   Real yields rising more than inflation expectations

3. Cost

   How it works

   AI absorbs memory chips and electricity, raising their prices for everyone

   What to look for

   Memory and power prices, and their pass-through to inflation

Channel	How it works	What to look for
Supply	Tech companies issue large amounts of bonds to fund data centres, competing for the same savings as governments	Wider corporate spreads, yields rising around big deals
Investment demand	A wave of investment raises the return on capital, which lifts the neutral interest rate	Real yields rising more than inflation expectations
Cost	AI absorbs memory chips and electricity, raising their prices for everyone	Memory and power prices, and their pass-through to inflation

The three can operate together, but they leave different marks in the data.

How much is Big Tech spending and borrowing?

The four largest spenders gave these 2026 capital expenditure guides in July:

Company	2026 capex guidance
Amazon	about US$220 billion
Alphabet	US$195-205 billion
Microsoft	about US$175 billion (calendar year)
Meta	US$130-145 billion
Total	about US$720-745 billion

That total is about 2.2-2.3% of US nominal GDP (Platformonomics tally). Goldman Sachs expects five companies including Oracle to reach US$1.2 trillion in 2027.

Much of it is borrowed. Morgan Stanley expects about US$570 billion of AI-related bond issuance worldwide in 2026, with about US$236 billion done by the end of May, four times the same period a year earlier. Amazon sold the equivalent of about US$53.8 billion in March, the largest corporate bond sale on record. Alphabet sold a 100-year sterling bond in February. Reported spreads on long hyperscaler bonds widened from a median of 108.5 to 118 basis points.

Does the bond market show a supply effect?

Not clearly. PIMCO examined six large AI bond deals in September 2026 and found no significant reaction in 10-year Treasury yields, the term premium or swap spreads around them. Capital Economics described the effect on yields as limited on 1 October. The cost shows up instead in the companies’ own borrowing: wider spreads and larger new-issue concessions.

Real yields rose much more than inflation expectations

US 10-year real yield vs 10-year inflation expectations, 2 January to early October 2026 (FRED DFII10, T10YIE)

US 10-year, 2026	2 January	Early October	Change
Nominal yield	4.19%	5.27% (6 Oct)	+108 bp
Real yield (inflation-indexed)	1.94%	2.91% (6 Oct)	+97 bp
Breakeven inflation	2.25%	2.36% (7 Oct)	+11 bp

Almost all of the rise in US yields this year is in the real component. The real 10-year reached 2.95% on 5 October, its highest since November 2008. That is what the investment-demand channel predicts: if the return on capital rises, so does the real rate. The New York Fed’s estimate of the neutral real rate rose from 0.89% in Q4 2025 to 1.01% in Q2 2026, and the FOMC’s long-run rate projection moved from 3.1% in June to 3.2% in September.

Two Federal Reserve governors have linked AI to the rate outlook. Governor Cook said on 28 September that “in the short term, AI appears to be adding inflationary pressures to the economy”. Governor Barr said in February that “the AI boom is unlikely to be a reason for lowering policy rates”.

But real yields also respond to Fed increases, government deficits and the term premium, all of which changed in 2026. The +97 basis points cannot be assigned to AI alone.

The cost channel: memory and electricity

Memory chips and power are the inputs most visibly squeezed by data centres. Conventional DRAM contract prices rose 93-98% in the first quarter of 2026 on the previous quarter, according to TrendForce. US consumer electricity prices were up 3.8% in the year to August (BLS). In the largest US power market, the PJM capacity auction for 2027/28 cleared at the cap of US$333.44 per MW-day, a third record in a row (Utility Dive). In the Philippines, the iPhone 17 256GB was repriced from ₱57,990 to ₱69,990 on 10 September 2026 (Technobaboy), which local tech sites linked to the memory shortage. The CPI group for information and communication rose only 1.0% in August, so the cost channel is visible in shop prices more than in the national index.

The case that AI lowers rates instead

PIMCO argued on 1 July 2026 that AI could lower neutral rates in the short and medium term. Its reasons: households may save more as a precaution if they fear for their jobs, a lower labour share of income can reduce spending, and fiscal expectations may change. PIMCO also found that on 43 days of major AI model releases, the US 5-year real forward rate five years ahead fell by about 100 basis points in total, while it rose by about 100 over the full period. On that reading, AI news has pushed long-term real rates down, not up.

What reaches the Philippines?

Mostly trade and prices, not bond yields.

Measure	Value	Period
Electronics exports	US$4.79 billion, 58.8% of total exports	July 2026
Increase in electronics exports	+US$869.72 million on a year earlier	July 2026
Data centre capacity	about 200 MW; aim of 1 GW by end-2026	DICT, February 2026
Data centre plan	1.5 GW committed by 2028; 18 GW goal within 10 years	DICT, February 2026
Philippine 10-year yield	6.05% to 7.99%	2 January to 7 October 2026

Exports. Electronics were 58.8% of Philippine exports in July 2026, and the largest year-on-year increase of any product group (PSA data). AI-driven chip demand is a direct source of export earnings.

Data centres. The Department of Information and Communications Technology said in February 2026 that Philippine capacity was about 200 MW and could reach 1 GW by year-end, with 1.5 GW committed by 2028 and a goal of 18 GW within ten years. These are plans and commitments, not completed capacity.

Rates. The Philippine 10-year rose 193 basis points this year, but the reasons given by analysts in the local press are mostly domestic or oil-related: inflation of 7.2% in September, three BSP increases to 5.00%, and a weak peso. To the extent AI has lifted global real yields, that is one more pressure on Philippine yields through US rates, but the data here cannot measure how much.

Open the past money calculator →