Goldman expects U.S. core PCE to still be 3.2% at end-2026, but to fall to 2.2% by end-2027

Overview

This Goldman Sachs U.S. economics research report updates its midyear inflation outlook. It notes that core PCE inflation has reaccelerated since the start of 2026, driven mainly by energy price pass-through, AI-related memory price increases and the resulting measurement distortions in PCE, as well as financial-services inflation tied to higher equity prices. In the baseline case, the lower oil price forecast eases part of the upside pressure, while slower rent and wage growth continue to provide disinflationary forces.

Core views

The core view is: first, a U.S.-Iran agreement and lower oil prices have reduced the risk of energy inflation, but goods prices could still add about 0.4 percentage points to year-over-year core PCE in 2026. Second, the AI infrastructure boom is pushing up memory prices, which then lift core PCE through computer software and accessories, phones, and computers, with PCE more affected because of measurement issues. Third, slower housing rent and wage growth should help reduce underlying inflation, but medical services still face catch-up-style cost inflation. Fourth, core PCE will remain elevated in 2026, then ease in 2027 as the AI and energy effects fade; overall risks are net tilted to the upside.

Analysis framework

The report uses a disaggregated inflation framework, separately modeling energy price pass-through, AI-related hardware costs, housing rent, wage-sensitive services, medical services, financial services, and supply-chain input costs, and combines oil price scenarios, high-frequency memory price data, PPI pipeline pressure indicators, persistent inflation risk indicators, and a bottom-up core PCE model to form the forecast.

Methodology notes

• Commodity pass-throughFramework for pass-through from goods prices to core PCE

  Indirect inflation effects from oil prices, transportation costs, and refined-product spreads

  Goldman incorporates the latest oil price forecast, freight costs, and refined-product spreads into the pass-through framework, assuming refined-product spreads normalize gradually over the coming months and transportation costs return to a normal relationship with oil prices.

• AI inflation shockMemory-price-driven model of software and accessories inflation

  AI infrastructure demand lifts memory prices and passes through to PCE components

  Using high-frequency memory prices, memory vendor data, and equity analyst forecasts, the model estimates that computer software and accessories inflation will slow from around 4%-5% month-over-month in recent months to about 0.6% by year-end.

• Services inflationFramework for rent- and wage-sensitive services inflation

  Downward impact on core services from slower new-tenant rent and nominal wage growth

  The report argues that weak new-lease rents will keep housing inflation cooling, while slower wage growth will also pull down labor-intensive PCE service prices in the coming months.

• Inflation risk monitoringPersistent inflation risk indicator

  Composite of slack, wage pressure, inflation expectations, and inflation breadth

  The indicator shows persistent inflation risk has edged up recently, but remains broadly near 2000s levels and well below post-pandemic peaks.

• Forecast modelGS bottom-up core PCE model

  Forecasts split across core goods and core services by PCE weights

  The model indicates year-over-year core PCE falls from around 3.4% in May 2026 to 3.2% in December 2026, and then to 2.2% in December 2027.

Asset mapping & comparison

Structured mapping from thesis to named assets (strengths, weaknesses, peers, risks).

• U.S. core PCE

  Primary forecast target

  Strengths

  The disaggregated framework shows it could fall to 2.2% in 2027 as AI and energy effects fade.

  Weaknesses

  In 2026 it remains supported by AI-, energy-, and equity-related financial-services prices, with year-end forecast at 3.2%.

  Comparison

  Compared with core CPI, core PCE is more affected by AI measurement issues and financial-services prices.

  Risks

  A deterioration in Middle East tensions, rising inflation expectations, or firms raising prices more than input-cost increases would create upside risks.

• U.S. core CPI

  Comparison inflation metric

  Strengths

  Less affected by AI-related PCE measurement distortions and equity-driven financial-services pricing.

  Weaknesses

  Still subject to indirect pass-through from some goods, services, and energy prices.

  Comparison

  Expected to be 2.6% in December 2026, below core PCE's 3.2% over the same period.

  Risks

  If housing, services, or goods prices cool more slowly than expected, core CPI could also exceed the baseline forecast.

• Brent crude and energy prices

  Source of upside or downside inflation shocks

  Strengths

  The baseline forecast is lowered to an average of $80 in 2026Q4 and $75 in 2027, reducing upside inflation pressure.

  Weaknesses

  Middle East supply disruptions could still quickly push up oil prices and refined-product spreads.

  Comparison

  Relative to baseline, the upside scenario could add another 1.1 percentage points to headline PCE and 0.2 percentage points to core PCE; the downside scenario would subtract 0.4 and 0.1 percentage points, respectively.

  Risks

  Hormuz Strait disruptions, renewed widening in refined-product spreads, and transportation costs that are higher than normal relative to oil prices.

• AI-related hardware, memory, software, and accessories

  New inflation driver

  Strengths

  Memory-price gains have already slowed, and software and accessories inflation is expected to fall to about 0.6% by year-end.

  Weaknesses

  Memory prices rose 10-15x over several months from 4Q25 to 1Q26, and will still push related prices higher through cost pass-through.

  Comparison

  The impact on core PCE is far larger than on core CPI, mainly because of measurement distortions in the PCE software and accessories item.

  Risks

  An AI-related memory shortage lasting longer than expected, with stronger pass-through to phone and computer costs.

• Housing rent

  Source of disinflation

  Strengths

  New-tenant rent growth is weak, and housing inflation is expected to continue cooling to about 1-1.5 percentage points below the pre-pandemic pace.

  Weaknesses

  Monthly housing inflation has been volatile recently, and some regions may show statistical noise.

  Comparison

  The decline in housing inflation is expected to roughly offset the fact that medical-services inflation remains above the pre-pandemic pace.

  Risks

  If rent deceleration is slower than expected, core services inflation could remain stickier.

• Medical services

  Upside item within services inflation

  Strengths

  Slower wage growth and smaller Medicare payment updates help limit the upside.

  Weaknesses

  Costs have risen faster than consumer prices since 1Q20, so catch-up pressure remains.

  Comparison

  Medical inflation is expected to stay around 3%, about 1-1.5 percentage points above pre-pandemic levels.

  Risks

  Stronger-than-expected pass-through of medical costs to insurers and consumer prices.

Key data

• Core PCE forecast for December 20263.2% YoYAI-, energy-, and equity-related pressures keep core PCE elevated.
• Core PCE forecast for December 20272.2% YoYFalls as energy and AI-related effects fade.
• Core CPI forecast for December 20262.6% YoYCore CPI is less affected by AI measurement issues and financial-services pressure.
• Core CPI forecast for December 20272.2% YoYThe report expects both PCE and CPI to be close to 2% by end-2027.
• Brent baseline oil price forecast$80 in 2026Q4; $75 in 2027Down from the prior $90 and $80.
• Oil price upside scenario$130 in 2026; $105 average in 2027Assumes Hormuz Strait disruptions persist into 2027.
• Goods prices' contribution to core PCEabout 0.4ppExpected to lift year-over-year core PCE in 2026, with the effect fading in 2027.
• AI-related price pressure's contribution to core PCEabout 0.4pp in December 2026; peak about 0.6pp in 2H26About 0.25pp already realized; about 0.1pp contribution to core CPI.
• Preliminary June headline CPI and PCE estimates-0.13% CPI; 0.07% PCEFalling gasoline prices are expected to produce softer headline inflation prints.
• Medical services cost and price changesCosts up 26% since 1Q20; consumer prices up 18%Medical services remain a major catch-up-style inflation item.

Impact & implications

For macro assets, the report sends a mixed signal: inflation remains sticky in the near term but should ease over the medium term. The lower oil forecast helps relieve headline inflation and energy pass-through pressure, while slower rent and wage growth support a move toward 2% inflation in 2027. However, mismatches in AI hardware supply and demand, financial-services pricing, and Middle East geopolitical risk mean core PCE in 2026 could still exceed target, and policy rates as well as long-end yields remain sensitive to upside inflation surprises.

Risks

• A deterioration in Middle East tensions or Hormuz Strait disruptions pushing up oil and refined-product prices.
• Rising inflation expectations leading firms to raise prices more than input-cost increases.
• AI-related memory shortages taking longer than expected to resolve, keeping software, accessories, phone, and computer prices rising.
• Housing inflation slowing more slowly than assumed in the baseline.
• Persistent catch-up pressure in medical-services costs offsetting disinflation in other service components.
• Equity-market gains continuing to lift core PCE through financial-services prices.

What to watch

• Whether Brent oil, gasoline prices, refined-product spreads, and freight costs normalize as in the baseline scenario.
• Whether spot and contract memory prices continue to cool, and whether the computer software and accessories CPI/PCE components decline.
• New-tenant rent indicators and monthly housing inflation readings.
• Wage growth and labor-intensive non-housing services prices.
• Medical-services prices, Medicare payment updates, and insurance negotiation outcomes.
• PPI pipeline pressure indicators, inflation breadth, and medium- to long-term inflation expectations.
• Whether the gap between core PCE and core CPI continues to reflect PCE measurement effects and financial-services influences.